ISO/TC 146/SC 1/WG 36 N XX

ISO/DIS 16911-1:2025(en)

ISO TC 146/SC 1/WG 36

Version WD03-02 (N XX)

Previous version: WD03-01 (N xx)

Date: 2025-06-12

Stationary source emissions — Manual and automatic determination of velocity and volume flow rate in ducts —
Part 1: Manual reference method

© ISO 2025

All rights reserved. Unless otherwise specified, or required in the context of its implementation, no part of this publication may be reproduced or utilized otherwise in any form or by any means, electronic or mechanical, including photocopying, or posting on the internet or an intranet, without prior written permission. Permission can be requested from either ISO at the address below or ISO’s member body in the country of the requester.

ISO copyright office

CP 401 • Ch. de Blandonnet 8

CH-1214 Vernier, Geneva

Phone: +41 22 749 01 11

Email: copyright@iso.org

Website: www.iso.org

Published in Switzerland

Contents

Foreword viii

Introduction ix

1 Scope 1

2 Normative references 1

3 Terms and definitions 1

4 Symbols and abbreviated terms 3

4.1 Symbols 3

4.2 Abbreviated terms 6

5 Principle 7

5.1 General 7

5.2 Monitoring objectives 7

5.3 Principle of flow velocity determination at a point in the duct 8

5.4 Principle of measurement of volume flow rate 8

5.4.1 General 8

5.4.2 Principle of volume flow rate determination from point velocity measurements 8

5.4.3 Determination of volume flow rate using tracer dilution measurements 9

5.4.4 Determination of volume flow rate from plant thermal input 9

6 Selection of measurement technique 10

6.1 Monitoring objective 10

6.2 Choice of measurement technique to determine point flow velocity 11

6.3 Choice of the measurement technique for volume flow rate and average flow determination 11

7 Measuring equipment 12

7.1 General 12

7.2 Measurement of duct area 12

8 Performance characteristics and requirements for differential pressure devices and vane anemometers 13

8.1 General 13

8.2 Differential pressure devices 14

8.3 Vane anemometers 16

9 Measurement procedure 16

9.1 Measurement strategy 16

9.1.1 Site survey before measurement 16

9.1.2 Correction for time related flow variation for the characterisation of a velocity profile 17

9.1.3 Consideration of flow measurement assembly surface area in relation to measurement plane area 17

9.2 Determination of measurement plane and number of measurement points 18

9.3 Checks before sampling 18

9.3.1 General 18

9.3.2 Pre-test leak check 19

9.3.3 Check on stagnation and reference pressure taps (S-type Pitot tube) 19

9.3.4 Tests of repeatability at a single point 19

9.3.5 Swirl or cyclonic flow 19

9.4 Quality control 20

9.5 Measurement of flow at locations within the measurement plane 20

9.6 Post-measurement quality control 20

10 Calculation of results 21

10.1 General 21

10.2 Measurement of velocity 21

10.3 Determination of the mean velocity 21

10.4 Correction of average velocity for wall effects 21

10.5 Calculation of the volume flow rate from the average velocity 22

10.6 Conversion of results to standard conditions 22

10.6.1 General 22

10.6.2 Conversion of the volume flow rate to standard conditions 22

10.6.3 Dry volume flow rate in standard conditions 23

10.6.4 Conversion of the volume flow rate to a reference oxygen concentration 23

11 Establishment of the uncertainty of results 23

12 Evaluation of the method 24

Annex A (normative) Measurement of velocity using differential pressure based techniques 25

A.1 Principle of differential pressure based technique 25

A.2 Measuring equipment 25

A.2.1 Pitot tubes 25

A.2.1.1 L-type Pitot tube 25

A.2.1.2 S-type Pitot tube 26

A.2.1.3 3D Pitot tube 27

A.2.1.4 2D Pitot tube 28

A.2.1.5 Examples of Pitot tube designs 28

A.2.1.5.1 AMCA-type 28

A.2.1.5.2 NPL-type 29

A.2.1.5.3 CETIAT-type 30

A.2.2 Differential pressure flow measurement equipment 31

A.2.2.1 General 31

A.2.2.2 Pitot tube 32

A.2.2.3 Differential pressure measurement device 33

A.2.2.4 Measurements of stack gas conditions 33

A.3 Calculation 34

A.3.1 Determination of velocity using differential pressure devices 34

A.3.2 Density of the stack gas 34

A.3.3 Absolute pressure of gas 35

A.3.4 Molar mass of gas 35

Annex B (normative) Vane anemometer 37

B.1 Principle of vane anemometer 37

B.2 Calculation 38

B.2.1 Background 38

B.2.2 Lowest range limit 38

B.2.3 Highest range limit 39

B.3 Calculation of the uncertainty and calibrations 40

B.3.1 General 40

B.3.2 Example 40

Annex C (normative) Tracer gas dilution method determination of volume flow rate and average velocity 42

C.1 Tracer gas by dilution 42

C.1.1 Principle of the use of tracer gas injection 42

C.1.2 Tracer gas injection 42

C.1.3 Tracer gas concentration measurement 44

C.1.4 Tracer gas calibration equipment 44

C.1.4.1 Tracer gas injection 44

C.1.4.2 Tracer gas concentration measurement 44

C.1.4.3 Calculation of stack gas flow rate from tracer injection results 45

C.2 Uncertainty of the calibration result 45

C.2.1 General 45

C.2.2 Uncertainty of concentration measurement 45

C.2.3 Uncertainty of tracer gas mixing 46

C.2.4 Uncertainty of tracer injection rate 47

C.2.5 Uncertainty of stack flow rate 47

Annex D (normative) Transit time tracer gas method determination of average velocity 48

D.1 Existing standards 48

D.2 Transit time method 48

D.2.1 Principle of the method 48

D.2.2 Choice of the tracer 48

D.2.3 AMS flow calibration procedure 48

D.2.4 Calculation of the volumetric flow rate 49

D.2.5 Provisions for measurement site 49

D.2.5.1 Duct area 49

D.2.5.2 The length of the measurement section 49

D.2.6 Flow condition 49

D.3 Minimum requirements 49

D.3.1 Tracer 49

D.3.2 Mixing 49

D.3.3 Measurement section 49

D.3.4 Tracer concentration measurement 50

D.4 Performance requirements 50

D.4.1 Injection 50

D.4.2 Measurement of the tracer pulse 50

D.4.3 Calculation of the transit time 50

D.4.4 Calculation of the volumetric flow 51

D.5 Uncertainty of the calibration result 51

D.5.1 Calculation principle 51

D.5.2 Uncertainty of determination of the volume and measurement of time 52

D.6 Numerical example of uncertainty calculation in stack flow calibration 53

D.6.1 Uncertainty of q_V, ref 53

D.6.2 Expanded uncertainty 54

Annex E (normative) Calculation of flue gas volume flow rate from energy consumption 55

E.1 Principle 55

E.2 Fuel factor 55

E.2.1 Fixed factors for commercially traded fossil fuels 56

E.2.2 Factors corrected for specific energy 56

E.2.3 Factors derived from fuel composition 57

E.3 Energy consumption 58

E.4 Calculation of flue gas volume flow rate 58

E.5 Performance requirements 59

E.6 Example of uncertainty calculations 59

E.6.1 Example 1 — Coal-fired power plant 59

E.6.2 Example 2 — Biomass fired combined heat and power plant 60

E.6.3 Example 3 — Natural gas fired gas turbine plant 61

Annex F (informative) The use of time of flight measurement instruments based on modulated laser light 62

Annex G (informative) Example of uncertainty budget established for velocity and volume flow rate measurements by Pitot tube 63

G.1 Process of uncertainty estimation 63

G.1.1 General 63

G.1.2 Determination of model function 63

G.1.3 Quantification of uncertainty components 63

G.1.4 Calculation of the combined uncertainty 63

G.1.5 Other sources of errors 63

G.2 Example uncertainty calculation 64

G.2.1 Calculation of the physicochemical characteristics of the gas effluent 66

G.2.2 Calculation of uncertainty associated with the determination of local velocities 67

G.2.2.1 General 67

G.2.2.2 Standard uncertainty on the coefficient of the Pitot tube 67

G.2.2.3 Standard uncertainty associated with the mean local dynamic pressures 67

G.2.2.4 Standard uncertainty associated with the density of the gas effluent 70

G.2.2.4.1 General 70

G.2.2.4.2 Standard uncertainty associated with the molar mass of gas 70

G.2.2.4.3 Standard uncertainty associated with the temperature T_C 71

G.2.2.4.4 Standard uncertainty associated with the absolute pressure in the duct, p_c 71

G.2.2.4.5 Standard uncertainty associated with the density 73

G.2.2.5 Standard uncertainty associated with the local velocities 73

G.2.3 Calculation of uncertainty associated with the mean velocity 74

G.2.4 Calculation of uncertainty in reported values 75

G.2.4.1 Volume flow rate in the actual conditions of temperature, pressure, water vapour content and oxygen 75

Annex H (informative) Description of validation studies 76

H.1 Overview of validation studies 76

H.1.1 General 76

H.1.2 Municipal waste incinerator in Denmark 76

H.1.3 Coal-fired power plant in Germany 76

H.2 Results of laboratory validation 77

H.3 Results of field validation studies 78

H.3.1 Repeatability and uncertainty of manual methods in the first field validation study 78

H.3.2 Repeatability and uncertainty of manual methods in the second field validation study 80

Annex I (informative) Check of validity of the calibration of a Pitot tube 83

I.1 Flow generation facility 83

I.2 Calibration check procedure 83

Annex J (informative) Differential pressure measurement 85

J.1 General 85

J.2 Liquid manometers 85

J.3 Digital manometers and other electronic devices 86

J.3.1 General 86

J.3.2 Types of pressure sensors 86

J.3.2.1 Piezoresistive strain gauge 86

J.3.2.2 Capacitive pressure sensor 87

J.3.2.3 Magnetic pressure sensor 87

J.3.2.4 Piezoelectric pressure sensor 87

J.3.2.5 Optical pressure sensor 87

J.3.2.6 Potentiometric pressure sensor 87

J.3.2.7 Resonant pressure sensor 87

J.3.3 Differential pressure gauges 87

Annex K (informative) Degree of swirl determination example method 88

Foreword

ISO (the International Organization for Standardization) is a worldwide federation of national standards bodies (ISO member bodies). The work of preparing International Standards is normally carried out through ISO technical committees. Each member body interested in a subject for which a technical committee has been established has the right to be represented on that committee. International organizations, governmental and non-governmental, in liaison with ISO, also take part in the work. ISO collaborates closely with the International Electrotechnical Commission (IEC) on all matters of electrotechnical standardization.

The procedures used to develop this document and those intended for its further maintenance are described in the ISO/IEC Directives, Part 1. In particular, the different approval criteria needed for the different types of ISO documents should be noted. This document was drafted in accordance with the editorial rules of the ISO/IEC Directives, Part 2 (see www.iso.org/directives).

Attention is drawn to the possibility that some of the elements of this document may be the subject of patent rights. ISO shall not be held responsible for identifying any or all such patent rights. Details of any patent rights identified during the development of the document will be in the Introduction and/or on the ISO list of patent declarations received (see www.iso.org/patents).

Any trade name used in this document is information given for the convenience of users and does not constitute an endorsement.

For an explanation of the voluntary nature of standards, the meaning of ISO specific terms and expressions related to conformity assessment, as well as information about ISO's adherence to the World Trade Organization (WTO) principles in the Technical Barriers to Trade (TBT), see www.iso.org/iso/foreword.html.

This document was prepared by Technical Committee ISO/TC 146, Air quality, Subcommittee SC 1, Stationary source emissions, in collaboration with the Technical Committee CEN/TC 264, Air quality, of the European Committee for Standardization (CEN).

This second edition cancels and replaces the first edition (ISO 16911-1:2013), which has been technically revised.

The main changes are as follows:

— The monitoring objectives with different uncertainty requirements, ranging from very stringent (Emission Trading Schemes and calibration of automated flow measuring systems) to less demanding (support of isokinetic sampling) have been clarified.

— The level of quality control in relation to the uncertainty requirements of the monitoring objective have been clarified.

— Monitoring objectives have been grouped based on the required quality control.

— The measurement techniques and the associated requirements have been described in more detail.

— Performance characteristics and requirements for differential pressure devices and vane anemometers have been adapted to the state of the art.

— The example uncertainty calculations have been improved and corrected.

A list of all parts in the ISO 16911 series can be found on the ISO website.

Any feedback or questions on this document should be directed to the user’s national standards body. A complete listing of these bodies can be found at www.iso.org/members.html.

Introduction

This document describes a method for periodic determination of the axial velocity and volume flow rate of gas within emissions ducts and stacks and a method for the calibration of automated flow measuring systems permanently installed on a duct or stack.

This document provides a method which uses point measurements of the flow velocity to determine the flow profile and mean velocity and volume flow rates. It also provides for alternative methods based on tracer gas injection, which can also be used to provide routine calibration for automated flow measuring systems. A method based on calculation from energy consumption is also described. This document provides guidance on when these alternative methods can be used.

Stationary source emissions — Manual and automatic determination of velocity and volume flow rate in ducts —
Part 1: Manual reference method

This document specifies a method for periodic determination of the axial velocity and volume flow rate of gas within emissions ducts and stacks. It is applicable for use in circular or rectangular ducts with measurement locations meeting the requirements of ISO 15259. Minimum and maximum duct sizes are driven by practical considerations of the measurement devices described within this document.

NOTE ISO 15259 is identical to EN 15259^[12].

This document requires all flow measurements to have demonstrable metrological traceability to national or international primary standards.

This document applies to a range of monitoring objectives with different uncertainty requirements, ranging from very stringent (Emission Trading Schemes and calibration of automated flow measuring systems) to less demanding (support of isokinetic sampling). The level of quality control within this document is determined by the uncertainty requirements of the monitoring objective. Monitoring objectives are grouped based on the required quality control. The document specifies which requirements and performance characteristics apply to specified measurement objectives and application areas.

The methods specified in this document can be used as a standard reference method, if the user demonstrates that the performance characteristics of the methods are equal to or better than the performance criteria specified in this document and that the expanded uncertainty of the measurement results obtained by the methods, expressed with a level of confidence of 95 %, is determined and reported. The results for each method defined in this document have different uncertainties within a range of 1 % to 10 % at flow velocities of 20 m/s.

Other methods can be used provided that the user can demonstrate equivalence, e.g. based on the principles of EN 14793^[11].

The following documents are referred to in the text in such a way that some or all of their content constitutes requirements of this document. For dated references, only the edition cited applies. For undated references, the latest edition of the referenced document (including any amendments) applies.

ISO 15259, Air quality — Measurement of stationary source emissions — Requirements for measurement sections and sites and for the measurement objective, plan and report

ISO 20988, Air quality — Guidelines for estimating measurement uncertainty

ISO/IEC Guide 98‑3, Uncertainty of measurement — Part 3: Guide to the expression of uncertainty in measurement (GUM:1995)

For the purposes of this document, the following terms and definitions apply.

ISO and IEC maintain terminology databases for use in standardization at the following addresses:

— ISO Online browsing platform: available at https://www.iso.org/obp

— IEC Electropedia: available at https://www.electropedia.org/

3.1

Pitot tube

device to measure flow velocity at a point, operating on the principle of differential pressure measurement

Note 1 to entry: A number of designs of Pitot tube can be used, including standard L-type, S-type, 2D and 3D Pitot tubes. Annex A describes a number of Pitot designs currently in use in Europe.

3.2

measurement line

line across the stack, on a measurement plane, along which flow measurements are made to characterize the flow velocity profile or to determine the average flow

3.3

measurement plane

plane normal to the centreline of the duct at the measurement location at which the measurement of flow velocity or volume flow rate is required

3.4

measurement point

sampling point

position in the measurement plane at which the sample stream is extracted or the measurement data are obtained directly

3.5

volume flow rate

volume flow of gas axially along a duct

Note 1 to entry: If not specifically stated, the term may be taken to mean the mean volume flow passing through the measurement plane.

Note 2 to entry: Volume flow rate is expressed in cubic metres per second or cubic metres per hour.

3.6

point flow velocity

local gas velocity at a point in the duct

Note 1 to entry: Unless otherwise specified, the term may be taken to mean the axial velocity at the measurement location.

Note 2 to entry: Point flow velocity is expressed in metres per second.

3.7

average flow velocity

<1> velocity which, when multiplied by the area of the measurement plane of the duct, gives the volume flow rate in that duct

<2> quotient of the volume flow rate in the duct and the area of the measurement plane of the duct

3.8

standard conditions

conditions for reference values for pressure (101,3 kPa) and temperature (273,15 K)

3.9

uncertainty (of measurement)

parameter, associated with the result of a measurement, that characterizes the dispersion of the values that could reasonably be attributed to the measurand

3.10

uncertainty budget

statement of a measurement uncertainty, of the components of that measurement uncertainty, and of their calculation and combination

Note 1 to entry: For the purposes of this document, the sources of uncertainty are according to ISO 14956^[5] or ISO/IEC Guide 98‑3.

3.11

standard uncertainty

uncertainty of the result of a measurement expressed as a standard deviation

3.12

expanded uncertainty

quantity defining an interval about the result of a measurement that may be expected to encompass a large fraction of the distribution of values that could reasonably be attributed to the measurand

Note 1 to entry: In this document, the expanded uncertainty is calculated with a coverage factor of k = 2, and with a level of confidence of 95 %.

3.13

overall uncertainty

expanded uncertainty attached to the measurement result

Note 1 to entry: The overall uncertainty is calculated according to ISO/IEC Guide 98‑3.

3.14

swirl

cyclonic flow

tangential component of the flow vector providing a measure of the non-axial flow at the measurement plane

3.15

automated measuring system

AMS

measuring system permanently installed on site for continuous monitoring of flow

Note 1 to entry: See ISO 16911‑2.

3.16

metrological traceability

property of a measurement result whereby the result can be related to a reference through a documented unbroken chain of calibrations, each contributing to the measurement uncertainty

Note 1 to entry: The elements for confirming metrological traceability are an unbroken metrological traceability chain to an international measurement standard or a national measurement standard, a documented measurement uncertainty, documented measurement procedure, accredited technical competence, metrological traceability to the SI units, and calibration intervals.

This document provides a method for the determination of gas velocity and volume flow rate within an emissions duct. It describes a method to determine the velocity profile of the gas flow across a measurement plane in the duct, and a method to determine the total volume flow rate at a measurement plane in the duct based on a grid of point velocity measurements made across the measurement plane. In addition, alternative methods are described for the determination of volume flow rate based on the measurement of tracer dilution, tracer transit time, and by calculation from energy consumption.

Techniques for determining gas velocity at a point include a calibrated differential pressure device (Pitot tube, see Annex A) and a calibrated vane anemometer (see Annex B). Selection criteria for the use of different types of Pitot and the vane anemometer are given in Clause 6. However, it is up to the user to ensure the method selected for a given application meets the performance criteria defined by this document. The volume flow rate within a duct is determined by measuring the duct axial gas velocity at a series of points along measurement lines across the duct on a single measurement plane. The number of measurement lines and measurement points required depends on the duct shape and size. The spacing of the measurement points is based on the principle of equal areas as defined in ISO 15259. The volume flow rate is calculated from the average axial velocity and the duct area at the measurement plane. If required a correction is applied to account for wall effects (see 10.4).

Three alternative methods are also described to determine volume flow rate and average flow velocity:

— Annex C describes a method based on tracer dilution measurements. In this method, the volume flow rate is determined from the dilution of a known concentration of injected tracer.

— Annex D describes a method based on a tracer transit time measurement technique. The volume flow rate is determined from the time for a pulse of tracer gas to traverse between two measurement locations.

— Annex E describes a method to determine the volume flow rate using a calculation-based approach to derive the flow from the energy consumption of a combustion process.

The volume flow rate may be reported at stack conditions or may be expressed at standard conditions (273,15 K and 101,3 kPa) on either the wet or dry basis.

This document applies to a range of monitoring objectives (MO) with different uncertainty requirements, and it provides quality control checks to enable these to be met. The level and extent of quality control checks and the selection of performance characteristics and their criteria have been established and specified based on the monitoring objective uncertainty requirements.

Monitoring objectives are grouped based on the required quality control. The grouping of monitoring objectives is as follows:

For simplicity any reference throughout this document to MO1 or MO2 refers to the above list.

This document can be used for other monitoring objectives, but the user has to specify the required level of quality control based on the uncertainty requirement of the monitoring objective.

The axial flow velocity at a point in the duct is determined using one of two techniques described in this document:

— differential pressure based measurement using Pitot tubes, and

— vane anemometry.

Annex A provides details for the use of differential pressure based techniques. Annex B describes the vane anemometer in detail.

The flow velocity is determined as the duct axial velocity at each point determined according to ISO 15259.

The differential pressure based techniques are based on the principle of the Pitot tube as defined in ISO 3966.^[3] A probe with one or more pressure taps is inserted into the flow. The basic principle is that one pressure tap is impacted by the flowing gas, and one or more other pressure taps are exposed to the static pressure in the duct. The probe assembly allows the resultant pressure difference between these to be measured by an external differential pressure measuring device.

Different implementations of the differential pressure approach are available. These include standard L-type, S-type, and multi-axis Pitot tubes (3D and 2D Pitot tubes). Each has their own specific advantages and disadvantages, and these are described in this document. The methods used are based on those specified in ISO 10780,^[4] ISO 3966,^[3] and US EPA Method 2.^[16] Performance requirements and quality assurance procedures are applied to achieve the uncertainties defined in this document.

If 2D Pitot tubes are to be used, information on the necessary QA/QC can be found in US EPA Method 2G^[18].

Volume flow rate can be determined from a series of measurements of the point velocity in a duct made across the measurement plane or by alternative techniques including tracer dilution, tracer transit time or calculation from energy consumption. Annex C, Annex D and Annex E provide details of these alternative approaches.

Volume flow rate is determined from a number of point measurements of the axial flow velocity over a measurement plane. Sufficient point measurements are made to characterize non-uniformities in the flow profile. The measurement points across the measurement plane are selected to be representative of regions of equal area. The average velocity passing through the measurement plane is calculated with good approximation as equal to the average of the point flow measurements. The procedures in ISO 15259 are used to determine the measurement points for circular or rectangular ducts. The tangential methodology provided in ISO 15259 is used for circular ducts as described in this document.

The reason why for circular ducts, the tangential methodology is preferred to the general method for determining equal areas (as specified in ISO 15259), is that in the tangential method the points provide a measure of the average flow in each equal area. The centre point in the general method does not provide a measure of the average flow in the centre area, but rather the maximum flow value. This can be useful for characterising the flow profile, but is not recommended for determining the average flow in the duct.

The measurement plane is selected to be representative of the required duct volume flow rate, and also to be in a region where it is uniform and stable. If non-axial flow (swirl or cyclonic flow) is expected at the measurement plane due to geometry of the duct or other upstream conditions, then the degree of swirl is determined using S-type, 3D or 2D Pitot tube measurements and if it is significant, as defined in this document, then it is taken into account through the use of additional measurement procedures, or a different measurement plane is selected.

If required, improved uncertainty in the results is achieved by taking wall effects into account following 10.4.

The volume flow rate q_V is determined by multiplying the average velocity by the area of the measurement plane (i.e. the internal area of the duct at the measurement plane) according to Formula (1):

(1)

where

is the average of the point velocity measurements;

A is the area of the measurement plane.

It is also possible to determine an array of volume flow rates, determined from the point measurements at each equal area multiplied by the area represented by each measurement point. Each measurement point area is, by definition, equal to the area of the measurement plane divided by the number of measurement points. The volume flow rate is then calculated according to Formula (2) which is equivalent to Formula (1):

(2)

where

v_i is the velocity at measurement point i;

A is the area of the measurement plane;

n is the number of measurement points.

Tracer gas injection is used to measure the volume flow rate by determining the dilution of the injected tracer by the stack gas flow. A known traceable flow rate of a tracer gas with specified composition is injected into the stack. The concentration of this tracer gas is measured at a location downstream, representative of the measurement plane, after adequate mixing of the tracer with the stack gas has occurred. Adequate mixing can be achieved when sampling from a test location that meets the homogeneity requirements of ISO 15259 for the tracer gas concentration. Guidance for achieving good tracer mixing quality are provided in Annex C.

The dilution of the tracer gas by the stack gas provides a measurement of the volume flow rate, provided that:

— the tracer gas is well mixed in the stack gas;

— there is no tracer gas present in the stack gas prior to injection or the background concentration can be measured and subtracted accurately. Determination of volume flow rate using transit time tracer measurements

A small amount of tracer material is injected rapidly into the stack gas flow, to produce a short pulse of tracer. After the tracer pulse has moved over the cross-section of the flow, its transit time between two measurement points placed on a suitable straight duct section is measured. The volume flow rate is calculated by dividing the duct volume between the measurement points by the transit time. The flow determined using this technique is representative of a region of the duct defined by the pulse measurement locations, and these are chosen to be representative of the required measurement plane.

For most combustion sources the volume flow rate can be calculated from the stoichiometric flue gas volume, determined from the fuel composition and the thermal energy input rate. The possible calculation methods are described in EN 12952-15,^[7] which includes both direct and indirect methods. In a direct method the fuel flow is measured and the thermal input is calculated from the specific energy ("calorific value") of the fuel and the fuel flow. Use of an indirect method includes measurement of the energy produced and the thermal efficiency of the plant. Especially for heat generation, or combined heat and power plants, with a high net thermal efficiency of typically 90 %, the uncertainty of the indirect method to calculate the thermal input is very low.

To later determine the actual flue gas flow rate, the oxygen volume concentration at the measurement plane in many cases shall be used to take account of the excess air. The oxygen concentration can be determined using EN 14789.^[9] However, the calculation method can also provide results at reference oxygen volume concentration values without requiring the determination of the oxygen composition in the duct. The calculation approach determines the volume flow rate on a dry gas basis. It also can be used to determine the wet flue gas flow but the uncertainty in such cases increases.

This document provides methods that can be used for a number of different monitoring objectives. The user of these methods shall understand the monitoring objective and the associated maximum permissible expanded uncertainty before undertaking any measurements as required by ISO 15259. The selection of the method can depend on the monitoring objective. This document groups the monitoring objectives specified in 5.1 based on the required quality control. The monitoring objectives specified in 5.1 include specific measurement objectives that can require the use of one or more measurement techniques. Table 1 outlines measurement techniques which can be used to achieve some example key measurement objectives that fall under one or more monitoring objectives.

Table 1 — Selection of measurement technique

The point velocity measurement methods described in this document can be used to fulfil any of the above measurement objectives, subject to the performance requirements of the method to be met.

The alternative methods described in this document may be used for the determination of the volume flow rate and for the calibration of flow automated measuring systems (AMS), provided specific requirements are met. These are detailed in 6.3.

The objective of the flow measurement should be clearly defined before selecting the measurement technique. In particular, the required measurement and reference conditions, (e.g. stack gas or standard conditions, and wet or dry conditions) should be determined. These requirements can influence the selection of the measurement technique.

EXAMPLE If the flow measurements are used to calibrate an AMS which measures flow under stack conditions, then the flow is determined under these conditions to avoid additional uncertainties being introduced when converting between different conditions. Similarly, if mass emission rates are calculated using concentration data obtained on a dry basis, then flow values determined directly under dry conditions are preferred. It is not always possible to achieve this, and so this document provides procedures to convert the data to different reference conditions.

A measurement technique able to determine point velocity shall be employed in order to carry out the following measurement objectives:

a) velocity measurement at a point in the duct — this can be required as a part of another measurement method, e.g. for ensuring isokinetic sampling of particulates;

b) flow profile measurement across a plane in the duct;

c) determination of swirl.

These measurement techniques can also be employed to meet all other measurement objectives. This document allows the use of differential pressure devices or vane anemometer to determine point flow velocity.

The following provides some general advice on the selection of the point measurement technique. However, expert judgement and specific conditions inform the choice of the measurement technique on a case-by-case basis.

There are a number of different designs of Pitot tube which can be used. Annex A describes the use of these measurement techniques. These include the L-type, S-type and 2D and 3D Pitot tubes. Pitot tubes of different designs may be used provided that they meet the performance requirements given in this document, under the conditions of use. However, certain designs of Pitot tube are more appropriate to certain stack and measurement conditions.

A vane anemometer may be used provided the performance requirements given in this document are met. Annex B provides a procedure for the use of this measurement technique.

The objective is to determine the axial velocity at one or more measurement points in the stack. Most point measurement devices, if aligned to the axis of the stack, measure the magnitude of the flow velocity vector if the angle of the flow to the axis is small (<20° is typical as observed in the laboratory validation studies). This implies these devices could overestimate the axial velocity by a factor equal to the reciprocal cosine of the angle of the flow velocity to the axis.

NOTE 1 If there is significant swirl in the duct flow, the 3D-type Pitot can be the appropriate technique. A procedure is given in 9.3.5 to determine the swirl and provides criteria to determine if the swirl is significant.

NOTE 2 In ISO 10780^[4] and EN 13284-1,^[8] it is stated that the S-type Pitot tube is more sensitive than the L-type to alignment to the flow vector. However, the laboratory evaluation carried out for this document did not observe this. Both the L- and S-type Pitot tubes were observed to have similar response to being misaligned to the flow vector. In both cases, the Pitot tubes could be misaligned by up to 15° to 20° without significant (<1 %) change to the velocity reading.

For small measurement ports, and for use in conjunction with sample probes, the S-type Pitot tube can be more appropriate. The S-type Pitot tube can also be more appropriate for use in cases where there are droplets or significant dust loading in the stack. In conditions of high dust loading, a vane anemometer is not recommended as it can become fouled, potentially invalidating its calibration.

For velocities less than 5 m/s or differential pressures less than 5 Pa, the vane anemometer has the potential to give smaller uncertainties than the differential pressure-based techniques.

The average velocity can be determined from the average of a grid of point velocity measurements. In such cases, the measurement plane shall meet the requirements of ISO 15259. Annex C, Annex D and Annex E provide methods for the measurement of volume flow rate and average velocity. These methods are:

— tracer gas method by dilution;

— tracer method using transit time; and

— calculation from energy consumption for combustion processes.

The tracer gas methods can be used for the determination of volume flow rate and average velocity and for the calibration of an AMS used to determine these parameters.

NOTE 1 The use of the radioactive tracer time-of-flight can be restricted by national regulations on the use of radioactive tracers.

The section of duct where time of flight measurements is carried out shall be at least five hydraulic diameters of straight duct away from any disturbances/top of the stack.

Tracer gas methods require adequate mixing of the injected material and therefore the tracer gas methods require installation of an injection port at a suitable location and can require a port for the sensing element.

The calculation approach is suitable for the determination of volume flow rate from combustion sources or other processes, where the required process information is available as specified in Annex E. The methods shall not be used for the calibration of an AMS. This method determines the flow based on measurements or assigned values for input parameters from the combustion process, e.g. fuel composition and fuel amount. Where input data are measured, the measurement systems used shall be under appropriate quality control and shall be calibrated. For fuels with variable moisture content, a fuel sample per measurement period shall be taken and analysed.

NOTE 2 Fuel composition has only a small impact on the dry flow rate.

Methods which determine volume flow rate may also be used to determine average velocity at the measurement plane, through measurement of the stack diameter and the use of Formula (1), and vice versa.

Flow measurement can be carried out for a number of monitoring objectives, as described in Clause 6. The equipment used depends on the measurement technique adopted and is detailed in the relevant normative annex.

Measurements of the internal duct area shall be made using direct dimensional measurements. Care should be taken if external measurements are used to define the internal dimension, e.g. using external circumference for circular ducts or the measurement of sides of a rectangular duct. These approaches should only be used if the duct wall is constant in thickness, well defined and single skinned. The cladding thickness and wall thickness shall be subtracted from the internal diameter or the two sides of a rectangular duct, before the area is calculated.

Engineering drawings may only be used when accurate direct measurement of the internal duct area is challenging due to access restrictions, irregular dimensions, non-constant wall thickness and/or other complications. In such cases engineering drawings may be used for additional information in order to verify and increase confidence of the direct dimensional measurement approach. The method(s) and equipment used to determine the internal duct area shall be reported and the plant operator shall be notified about any improvements required to improve the measurement. If engineering drawings are used the maximum acceptance criterion shall be used as the associated expanded uncertainty.

The expanded uncertainty of the internal area of duct at measurement plane shall not exceed the acceptance criterion given in Table 2.

NOTE The dilution tracer method determines the stack flow rate directly and does not require an internal duct area measurement.

Table 2 — Acceptance criterion for the expanded uncertainty of the internal area of the duct at the measurement plane

The length of any ports and the thickness of the duct wall shall also be measured at each measurement port. The mean value of the length of all measurement ports (including wall thickness) can be used as the length value for each individual port as long as the difference of the individual port length including wall thickness with the mean of all measurements ports is less than 10 % of the mean, with a maximum difference allowed of 5 cm. If the length at each measurement port is more than 10 % of the mean or more than 5 cm on one or more of the measurement ports then the measurement points shall be marked differently for each measurement port to reflect the measurement port length including wall thickness.

NOTE If the measurement port protrudes beyond the inside wall, the extent of the protrusion needs to be measured e.g. with a U-shaped wire inserted through the measurement port and then retracted until the inside wall is found.

A laser measurement device (see Annex F) or a suitable rigid measuring rod can be used to directly determine the internal diameter on at least two axes. Care should be taken to ensure the measurements are perpendicular to the stack axis and that internal stack fittings and ports do not affect the measurement.

In high-temperature stacks, temperature effects can be considered and taken account of if using a measuring rod. These can be accounted for by using Formula (3):

(3)

where

dl is the change in length, in m;

L₀ is the measuring rod initial length, in m;

α is the linear expansion coefficient, in m/(m K);

T₀ is the initial temperature of the rod at the start of the measurement, in K;

T₁ is the final temperature of the rod at the end of the measurement, in K.

The linear expansion coefficient value depends on the measuring rod material. Regardless of whether the measuring rod length has been corrected for thermal expansion or not, temperature effects do not have to be accounted for within the uncertainty assessment. They are accounted for either through this correction or as a standard uncertainty which may be ignored as it accounts for less than 5 % of the maximum standard uncertainty.

This document specifies performance requirements for the manual determination of the point velocity across a measurement plane within an emissions duct for differential pressure devices and vane anemometers. Other measurement techniques may be used, in which case, it shall be demonstrated that they meet the performance criteria given in this document. In some cases, in order to meet the uncertainty requirements of certain measurement objectives under MO2, the performance criteria specified in Clause 8 should be more stringent.

Differential pressure devices shall meet the performance requirements and the frequency specified in Table 3.

Table 3 — Performance requirements of differential pressure readout devices

The calibration certificate of differential pressure readout devices can be provided by an external calibration laboratory or internally by the testing laboratory by carrying out a calibration check against a reference standard. The calibration check carried out internally shall meet the performance criteria in Table 3.

The resolution of an electronic pressure readout device shall have of at least 1 Pa (MO1) and at least one decimal place per Pascal (MO2). However, at very low differential pressures or very strict uncertainty requirements the use of two decimal places can be more appropriate. Table 4 provides a list of resolutions and standard uncertainties in relation to differential pressure values and velocities.

Table 4 — Resolution and associated standard uncertainty of differential pressure readout devices

NOTE The standard uncertainty associated with the resolution is derived through the formula , where α is the instrument resolution.

Pitot tubes shall meet the performance requirements and the frequency specified in Table 5.

Table 5 — Performance requirements of Pitot tubes

The calibration certificate of Pitot tubes can be provided by an external calibration laboratory or internally by the testing laboratory by carrying out a calibration check against a reference standard. The calibration check carried out internally shall meet the performance criteria in Table 5.

Density measurements and temperature and pressure measurement devices shall meet the performance requirements and the frequency specified in Table 6.

Table 6 — Performance requirements for density measurements and temperature and pressure measurement devices

The calibration certificate of temperature and pressure measurement devices can be provided by an external calibration laboratory or internally by the testing laboratory by carrying out a calibration check against a reference standard. The calibration check carried out internally shall meet the performance criteria in Table 6.

The lowest measurable flow by a differential pressure device is the minimum differential pressure measured or the limit of quantification, whichever is higher. Below 5 Pa differential pressure readout devices shall not be used.

NOTE 1 During the validation studies no sensitivity was detected for S-type misalignment effect from –0,1° up to 15° and for L-Type from 0° up to 15°.

NOTE 2 During the validation studies no sensitivity was detected for temperatures in the range of 20 °C to 290 °C for both L-Type and S-Type Pitot.

NOTE 3 Atmospheric pressure is unlikely to change significantly over a measurement. Therefore, the sensitivity of the system to atmospheric pressure is not significant.

NOTE 4 The acceptance criterion for the calibration uncertainty relates to an expanded uncertainty at a 95 % confidence level with k = 2.

NOTE 5 If the Pitot tube is used for isokinetic sampling and is attached on a sampling probe, it is preferable that the calibration is carried out under the same configuration (i.e. Pitot tube attached to sampling probe) to avoid the effect the proximity of the sampling probe can have on the Pitot coefficient. However, as the majority of test laboratories use interchangeable parts it is recognized that this is not always possible.

The uncertainty requirements specified in the individual standards (e.g. water vapour, oxygen, carbon dioxide) used for the determination of the density shall be applied.

The vane anemometer shall be calibrated at flow rates representative of the stack flow conditions, and the calibration points used to calibrate the vane anemometer shall encompass its range of operation (e.g. a maximum of twice the maximum flow plus fixed percentages of the maximum flow).

The calibration shall be a multiple point calibration spanning the velocity range of application. The calibration shall have metrological traceability. This may be achieved, for example, by the use of a flow facility with flow rates traceable to laser Doppler anemometry.

Vane anemometer devices shall meet the performance requirements and the frequency specified in Table 7.

Table 7 — Performance requirements for vane anemometers

The lowest measurable velocity by a vane anemometer is the minimum velocity measured or the limit of quantification, whichever is higher. Below the lowest range limit vane anemometers shall not be used. This value depends on flue gas characteristics and the design of the vane anemometer but is typically between 0,2 m/s to 0,5 m/s.

NOTE 1 Particulate concentrations (up to 50 mg/m³) have no influence on the performance of vane anemometers. Any fibre within the flue gas can affect the performance of the vane anemometer.

NOTE 2 Humidity has no influence on the performance of vane anemometers. However, any droplets in the stack gas can affect the performance of the vane anemometer.

NOTE 3 During the validation studies no sensitivity was detected for temperatures in the range of 20 °C to 290 °C.

The measurement site shall meet the requirements of ISO 15259. A survey of the plant and the measurement section before the measurement is necessary to gain information on access to the measurement plane, measurement port number and dimension, workplace conditions, weather protection, obstruction, and power supply at measurement site. From the information collected, select a fit for purpose measurement system and define a measurement plan. The measurement plan should include dates, starting times and duration of the measurement periods. Record plant operating conditions during measurement periods.

The measurement objective shall be discussed with the plant operator. The measurements shall be conducted under plant operating conditions that are suitable for the measurement objective. The plant flow should be allowed to stabilise before any measurements are conducted unless the measurement objective requires otherwise.

If the measurement objective is characterizing a flow profile two measurement devices (one at a fixed point and one that traverses the measurement plane) shall always be used. This is because, if in this case flow variations are not considered, there is a risk that a flow profile can seem to be asymmetric when in fact it’s not. This in turn can lead to incorrect decisions when it comes to confirming measurement plane suitability and/or selecting the type of flow AMS.

The fixed measurement device used can be a flow AMS as long as it has a valid QAL2 calibration function.

The correction factor to account for flow variations over time when using two devices can be determined by Formula (4):

(4)

where

f_v,cor,i is the velocity correction factor at measurement point i;

v_f,av is the average velocity of fixed device measurements, in m/s;

v_f,i is the velocity of fixed measurement device at measurement point i, in m/s.

The corresponding corrected velocity can be determined by Formula (5):

(5)

where

v_t,cor,i is the corrected velocity at measurement point i, in m/s;

v_f,i is the velocity of fixed measurement device at measurement point i, in m/s;

v_t,i is the velocity of traverse measurement device at measurement point i.

The area of the flow measurement assembly (sensing element and probe) shall not obstruct more than 5 % of the measurement plane area. For flow measurement assemblies that have integrated sampling devices (nozzle arm and nozzle, in-stack filter, etc.), mainly for the support of isokinetic sampling the area of the flow measurement assembly (sensing element and probe) shall not obstruct more than 10 % of the measurement plane area for stack or duct areas of less or equal to 1,5 m² to account for the larger measurement assembly area resulting from the integrated sampling device. For ducts or stacks with a very small measurement plane area where it is not possible to carry out isokinetic sampling and measure flow at the same time because the area of the sampling equipment can obstruct more than 10 % of the stack or duct measurement plane area, the flow measurement can be carried out prior to the isokinetic sampling and the values derived from the flow traverse used to control isokinetic conditions.

NOTE A typical point measurement survey, in a duct requiring 20 measurement points, can take up to 1 h. During this time, it is desirable to have flow conditions which are as stable as possible, unless the measurement objective requires otherwise.

The following requirements are applicable when a series of measurements are being made across the measurement plane, in order to determine the volume flow rate.

The measurement plane for the determination of volume flow rate shall conform to the requirements of ISO 15259. Any deviations from this document shall be reported.

If measurements are being carried out to characterize the flow conditions at the measurement plane, consideration shall be given to the need to characterize the flow under different process (load) conditions, as this can lead to different flow conditions. This is of particular concern when the measurements are being carried out to characterize the flow profile for the installation of a flow AMS or for the calibration of a flow AMS.

Depending on the measurement plan, the preliminary checks required by ISO 15259 shall be carried out before the start of sampling.

The following functional checks shall be carried out on the differential pressure readout device and/or the vane anemometer device before use:

— checking the zero value of the device,

— checking that the instrument responds to gas flow;

— ensuring if required that the correct input values are stored within the electronic differential pressure device when conversions from differential pressure to velocity are carried out internally;

— checking the internal damping time is fit for purpose in regard to the measurement task.

Measure the internal stack diameter (or stack width for rectangular/square stacks) and stack wall thickness at the measurement plane for each measurement line. The duct wall thickness should include the depth of the measurement port. Determine the required number of measurement lines and measurement points following the procedure given in ISO 15259, based on the equal area approach. For circular ducts, the tangential approach shall be used.

Calculate the position the probe should be inserted for each measurement point and mark the probe accordingly, taking account of the duct wall and port depth.

If required identify a suitable location for the fixed-point flow measurement system to check varying flow conditions with time. This may be a second flow measurement device, or a suitable flow AMS, provided it is located in a suitable position.

Before and after use Pitot tubes shall be checked for:

— deformities, burrs or damage to the tube;

— blockage of the pressure taps;

— leaks;

— cleanliness;

— straightness of the supporting tube.

Before and after use vane anemometers should be checked for:

— damage to the vane and housing;

— build-up of contamination on the vane blades;

— cleanliness;

— any sticking of the vane when gently blown on.

Any issues identified with the preceding two lists invalidate the calibration of the device. The Pitot tubes/vane anemometers affected shall not be used until the issues identified have been rectified (if possible) and the devices have been recalibrated.

Before undertaking a series of measurements and each time the system is disconnected, a leak check shall be performed for Pitot tubes. This can be achieved by pressurizing the tube at least as high a differential pressure as the differential pressure expected during the measurement or to 50 % of the range of the differential pressure readout device, whichever is higher. The pressure should remain stable to 2 % of the value or 10 Pa, whichever value is lower (for at least 15 s).

The S-type Pitot tube shall be positioned perpendicular to the direction of the flow. The static pressure is then measured using both taps. The difference in the measured static pressure shall be less shall be less than 1 % of the range of the differential pressure readout device or 10 Pa whichever is greater. If the difference in readings exceeds this tolerance this can indicate local swirl, blockage of pitot, etc. In this case the reason should be investigated and rectified or a new Pitot tube used.

If two measurement devices are being used, one fixed and one traverse, select one equal area location and make at least five paired readings of the velocity, using both measurement devices in the same equal area. Calculate the field repeatability from the standard deviation of the differences between these readings and compare to the field repeatability criterion in Table 7, as a percentage of the measured velocity.

The test of repeatability at a single point is only required under MO2.

If swirl is either known to exist from previous measurements or is expected due to the geometry or stack conditions, then non-axial flow (indicating swirl within the duct) shall be assessed at each measurement point using differential pressure devices. An example method is provided in the informative Annex K (only applicable to S-Type Pitot). Self-compensating anemometer models may also be used for flow angles up to 30°. If any of the tangential flow angles is greater than 15° to the axial direction at any measurement point in the plane, then swirl can be assumed to have a significant impact on the measurements. In such cases, measurement of the velocity at each point should be made using devices which can provide the flow velocity and angle of flow at each point. These include 3D, 2D, and S-type Pitot tubes.

If the swirl angle is greater than 15° but lower or up to 40°, the axis of the Pitot tube head shall be parallel to the local flow direction and the axial velocities at each point (where swirl angle is greater than 15° but lower or up to 40°) determined by Formula (6) using the cosines of the measured angles above 15° in order to account for the angle of swirl:

(6)

where

v_c is the axial velocity, in m/s;

v_m is the velocity measured with the Pitot tube head parallel to the local flow direction, in m/s;

cos θ_m is the cosine of the swirl angle measured.

When carrying out isokinetic testing and swirl is identified above 15° the number of measurement points shall be doubled (up to a maximum of 20 measurement points) and no sampling shall be carried out on the non-compliant points (i.e., the points that the angle of swirl has been identified to be above 15°). A distance of at least 5 cm or 3 % of the length of the measurement axis, whichever is greater, shall always be maintained between the inner wall and the first and last measurement points on the measurement axis. This deviation from standard sampling conditions according to ISO 15259 shall be documented and reported.

In instances where swirl above 15° is identified the operator should be informed and the first option should always be to seek an alternative sampling location. If another sampling location cannot be found, then depending on the measurement objective one of the above procedures may be followed.

A number of quality control and quality assurance procedures are required to ensure measurements made using differential pressure devices achieve the smallest measurement uncertainty. These are summarized in Table 8.

Table 8 — Performance requirements during field measurements

Specific requirements for vane anemometers are specified in Annex B.

The measurements of the velocity shall be carried out at each measurement point in the measurement plane.

The probe shall be inserted to the marked insertion depth. For differential pressure devices, at each measurement point, the average differential pressure shall be determined over at least 1 min, from the average of at least three instantaneous readings of the differential pressure. An electronic pressure readout device may be used to provide a direct reading of average differential pressure over at least a minute.

Sampling duration for the 3D Pitot tube can be significantly longer. It is important in such cases to ensure that temporal variations in the stack flow are considered.

If two flow measurement devices are used, the measurements from both devices shall be recorded in parallel.

The temperature at each measurement point shall be recorded, where appropriate. The measurements of oxygen, water vapour and carbon dioxide shall be recorded, if required, or the stack gas density determined by other means.

The atmospheric pressure shall be recorded.

The static pressure shall be recorded at least once for each measurement line.

For Pitot tubes a post-measurement leak check shall be carried out following the procedure in 9.3.2.

A blockage test is required for differential pressure devices.

This test is not required for an S-type Pitot tube. Its performance is required for each pressure tap of a 3D Pitot tube. It shall be carried out as a minimum after the measurement period. It is advisable to carry it out after each traverse as, if it is failed, all data since the previous test are invalidated.

Record the differential pressure at a location in the duct. When using 3D Pitot tube, purge it with pressurized air to clean the pressure taps. Repeat the measurement of the differential pressure at the same location. If the two readings are within 5 %, then the traverse data are acceptable. The fixed point readings shall be checked to ensure they do not vary by more than 2 % between these measurements. If they do, the blockage test readings may be corrected to account for the variation in flow.

In the calculation which follows, the gas properties are assumed to be the same across all points of the measurement plane.

Results may be reported under stack conditions or reference conditions. The choice of conditions under which to express the results should be informed by the monitoring objective, and by available data. If the measurements are being carried out in order to calibrate an AMS, then the flow data should be expressed in the same units. Where the volume flow rate data are being used to calculate mass emissions, then the flow data and the concentration data shall be expressed in the same conditions. To avoid additional uncertainty components, unnecessary conversions between conditions should be avoided. This requirement implies that, for example, if it is possible to measure simultaneously both concentration and flow data under wet (stack) conditions, then this is preferable as there is then no requirement to measure water vapour concentration in order to convert measurements to dry conditions, and the uncertainty component due to this measurement is avoided.

The velocity v_i at each measurement point i in the measurement plane is determined using the differential pressure device or vane anemometer (see Annex A and Annex B).

The mean axial velocity across the measurement plane is given by Formula (7):

(7)

where

mean axial velocity across the measurement plane, in m/s

i is the number of the measurement point, i = 1 ... n;

v_i is the local velocity at measurement point i, in m/s.

NOTE Formula (7) is only valid if the velocities are taken at equal area points.

The rapid curvature of the wall velocity profile is not fully captured by the relatively coarse measurement grid, defined according to ISO 15259, leading to a slight overestimate of flow rate. This document allows the operator to select a default wall adjustment factor (WAF), f_WA, which is used as a multiplier on the measured mean flow rate in a duct of circular cross-section according to Formula (8):

(8)

where

is the corrected mean velocity, in m/s;

is the mean velocity, in m/s;

f_WA is the wall adjustment factor.

The use of WAF in measurements in support of isokinetic sampling is not required. The default WAF is 0,995 for smooth walled ducts or 0,99 for rough walled ducts constructed from brick or mortar. Application of the WAF therefore results in a 0,5 % to 1,0 % reduction in volume flow rate. For rectangular ducts with 5 straight duct hydraulic diameters upstream and downstream of the measurement section, a default WAF of 0,97 can be assumed. For rectangular ducts with less than 5 straight duct hydraulic diameters upstream and downstream a default WAF of 0,95 can be assumed because of the increased influence of the walls and corner effects. It should be noted that rectangular ducts often contain side ribs and cross-bracing supports and, in solid fuel fired plant, dust build up in corners can alter the internal shape of the duct. Alternatively, factors for ducts of rectangular cross-section may be calculated using a procedure specified in US EPA CTM-041.^[15] This approach provides a more accurate factor but can need additional measurements in the duct and lengthy calculations.

NOTE US EPA Method 2H^[19] also specifies a method for calculating the WAF based on near wall velocity measurements in ducts of more than 1,0 m diameter, provided that the measured WAF is no lower than 0,97. US EPA CTM-041^[15] specifies a similar approach for ducts of rectangular cross-section, noting that the wall effect in these circumstances can be more significant since: the ratio of the perimeter to the cross-sectional area is greater; the test points are relatively further from the wall; and the corners of the duct are subject to increased wall effects from two adjoining walls.

The volume flow rate q_V,w under the conditions of temperature and pressure of the duct and on wet gas is given by Formula (9):

(9)

where

q_V,w is the volume flow rate under the conditions of temperature and pressure of the duct and on wet gas, in m³/s;

is the mean velocity, in m/s;

A_I is the internal area of the measurement plane, in m².

If required, the volume flow rate is converted to standard conditions according to the calculations in 10.6.2 to 10.6.4.

The volume flow rate shall be expressed at standard conditions, as a dry gas, where the standard conditions for temperature and pressure, i.e. 273,15 K and 101,3 kPa, respectively, shall be used.

If the pollutant is expressed to a reference oxygen concentration, the volume flow rate shall also be calculated under the same conditions.

The calculation of the uncertainty of the normalized flow rate shall take account of the uncertainties of related measurements to these conversions.

The dry volume flow rate q_V,0d under standard conditions of temperature and pressure is given by Formula (10):

(10)

where

q_V,0d is the dry volume flow rate, under standard conditions of temperature and pressure, in m³/s;

q_V,w is the volume flow rate under the conditions of temperature and pressure of the duct, on wet gas, in m³/s;

p_c is the absolute pressure in the duct in the measurement plane, in kPa;

T_c is the temperature of gas in the measurement plane, in K;

is the water vapour volume concentration of gas in the duct, in %.

The dry volume flow rate under standard conditions and on reference oxygen volume concentration is given by Formula (11):

(11)

where

is the dry volume flow rate under standard conditions and on reference oxygen volume concentration, in m³/s;

is the dry volume flow rate, under standard conditions and on actual oxygen volume concentration, in m³/s;

is the reference oxygen volume concentration, in %;

is the oxygen volume concentration measured in the duct during the exploration of the duct on dry gas, in %.

An uncertainty budget shall be determined for the reported results in accordance with the principle of the calculation of the overall uncertainty as specified in ISO/IEC Guide 98‑3 and ISO 20988.

— Determine the standard uncertainties attached to the performance characteristics to be included in the calculation of the uncertainty budget by means of laboratory and field tests, according to ISO/IEC Guide 98‑3.

— Calculate the uncertainty budget by combining all the standard uncertainties according to ISO/IEC Guide 98‑3, including the uncertainties in the calibration of the measurement devices, and any uncertainties due to conversion to reported conditions. If required uncertainties due to wall effects and swirl shall be taken into account.

— Values of standard uncertainty that are less than 5 % of the maximum standard uncertainty may be neglected.

— Calculate the overall uncertainty at the measured value, at the reported conditions.

— Uncertainty contributions due to the determination of the stack area shall be considered. This should take account of uncertainties due to the measurement device and stack non-uniformities. An example is given in Annex D.

— Uncertainty contributions due to wall effects may be determined by considering the use of the fixed factor specified in 10.4.

— The uncertainty contributions due to swirl may be discounted if the procedures for aligning the flow measuring devices given in this document are followed.

NOTE The laboratory assessment of point measurement devices showed that the maximum effect due to 15° of flow was less than 1 % of the measured velocity.

An illustrative uncertainty budget for measurements made using differential pressure reading devices is given in Annex G. Examples for other volume flow rate determination methods are given in the respective annexes.

The techniques described in this document were assessed during three validation studies (Reference [23]).

The validation studies were subdivided into two parts:

— laboratory tests at a wind tunnel site with a series of test runs involving manual methods (SRM) and automatic measuring methods (AMS);

— field tests at two plants site with a series of test runs involving manual methods (SRM) and automatic measuring methods (AMS).

Full reports on the validation studies are available through the CEN/TC 264 Secretariat, and are described in overview in Annex H.

The assessment of the field validation data in accordance to ISO 20988 provides the following results. The standard uncertainty of the result measurement y from the application of a manual flow measurement techniques in the range 17,8 m/s to 21,2 m/s, is u(y) = 0,49 m/s. The expanded uncertainty of a result of measurement y using a manual flow measurement method in the range 17,8 m/s to 21,2 m/s at a level of confidence of 95 % is U_0,95(y) = 0,98 m/s.

The 95 % confidence interval [y_R − U_0,95(y): y_R + U_0,95(y)] is expected to encompass P = 95 % of the measurement points. It was found to encompass P = 97,5 % of the evaluated 62 measurement results y(k,j). Therefore, the expanded uncertainty U_0,95(y) = 0,98 m/s is considered to be a reasonable measure of the uncertainty.

The uncertainties determined are therefore applicable to the measurement of average flow for an emissions duct formed by taking a grid of samples of point flow measurements.

1. 
   (normative)
   
   Measurement of velocity using differential pressure based techniques
   1. Principle of differential pressure based technique

The principle of the determination of velocity in a gas using differential pressure measurement is described in ISO 3966.^[3] A Pitot tube provides a means to determine the differential pressure within a region of the measurement plane. The L-type Pitot tube is a basic Pitot tube, though the principles are the same for all Pitot tube designs.

At least two pressure taps are aligned in the flow stream, one directly impacted by the flow, measuring the stagnation point pressure p₂ and one or more measuring the static pressure p₃. The static pressure tap may consist of a ring of holes around the Pitot tube in an L-type design, or a single 'wake' pressure tap as in an S-type design. In 3D Pitot tubes additional separate pressure taps can be present to measure the flow vector in three dimensions. The pressure at these orifices is transmitted by the tubing of the Pitot probe to a differential pressure meter mounted outside the stack. The difference in pressure is measured at this point. The measurement of the differential pressure may be carried out by either a digital manometer or a manual inclined liquid manometer.

• 1. Measuring equipment
     1. Pitot tubes
        1. L-type Pitot tube

This type is a basic Pitot tube, which consists of a tube pointing directly into the fluid flow. As this tube contains fluid, a pressure can be measured. The moving fluid is brought to rest (stagnates) as there is no outlet to allow flow to continue. This pressure is the stagnation pressure of the fluid, also known as the total pressure or (particularly in aviation) the Pitot pressure.

Figure A.1 illustrates the measurement principle of the L-type Pitot tube. Figure A.2 is a schematic diagram of an L-type Pitot tube.

Key

a p3 static pressure

b p2 stagnation pressure point

c v flow

d Δp differential pressure measurement

Figure A.1 — Principle of differential pressure-based velocity determination

Key

1 p3 static pressure

2 p2 total or stagnation pressure

3 v flow direction

Figure A.2 — Schematic diagram of an L-type Pitot tube

• • • 1. S-type Pitot tube

This type is also a basic Pitot tube, which measures directly in the flow. The working principle is similar to that of the L-type. The velocity is calculated on the same way as for the L-type. The S-type has to be calibrated against a reference method because the measured “static” pressure is not the real static pressure.

Figure A.3 is a schematic diagram of an S-type Pitot tube.

Key

1 p₂ total or stagnation pressure

2 p₃ static pressure

3 v flow direction

Figure A.3 — Schematic diagram of an S-type Pitot tube

• • • 1. 3D Pitot tube

This type of probe consists of five pressure taps in a spherical (or prism-shaped, which was not used on laboratory tests) sensing head. The pressure taps are numbered 1 to 5, with the pressures measured at each hole referred to as p₁, p₂, p₃, p₄, and p₅, respectively.

The differential pressure p₂ − p₃ is used to yaw null the probe and determine the yaw angle; the differential pressure p₄ − p₅ is a function of pitch angle; and the differential pressure p₁ – p₂ is a function of total velocity. A typical spherical 3D Pitot tube is shown in Figure A.4.

Key

1 sphere of diameter 38,1 mm (1,5 inch)

2 type K thermocouple

3 tube of diameter 19,05 mm (0,75 inch)

a₁ angle between sensing holes of 44°

a₂ angle between sensing holes of 44°

L_p probe length of 1,067 m (42 inch)

P₁ … P₅ taps (sensing holes) at which pressures are p₁ ... p₅

Figure A.4 — 3D (spherical) Pitot tube

• • • 1. 2D Pitot tube

A 2D Pitot tube measures the velocity pressure and the yaw angle of the flow velocity vector in a stack or duct. Alternatively, these measurements may be made by operating one of the 3D Pitot tubes described in A.2.1.3, in the yaw determination mode only. From these measurements and a determination of the stack gas density, the average near-axial velocity of the stack gas is calculated. The near-axial velocity accounts for the yaw, but not the pitch, component of flow. The average gas volume flow rate in the stack or duct is then determined from the average near-axial velocity.

• • • 1. Examples of Pitot tube designs
         1. AMCA-type

An AMCA-type Pitot tube is shown in Figure A.5.

Key

1 inner tube diameter

2 outer tube diameter

3 eight holes of diameter 0,13d, not to exceed 1 mm diameter maximum, equally distributed and free of burrs

Figure A.5 — AMCA-type Pitot tube

• • • • 1. NPL-type

A NPL-type Pitot tube is shown in Figure A.6.

Key

1 head

2 total pressure hole

3 modified ellipsoidal nose

4 static pressure holes

5 spacer

6 alternative curved junction

7 mitred junction

8 stem

9 alignment arm

10 pressure tapping

d outer tube diameter

^a Total pressure.

^b Static pressure.

Figure A.6 — NPL-type Pitot tube

• • • • 1. CETIAT-type

A CETIAT-type Pitot tube is shown in Figure A.7.

Key

d outer diameter

^a The radius is only useful when the Pitot tube is used in liquids in order to avoid cavitation.

NOTE Static pressure taps may be limited to those indicated on section A–A, in which case section A–A shall be placed at 6d from the tube tip.

Figure A.7 — CETIAT-type Pitot tube

• • 1. Differential pressure flow measurement equipment
       1. General

The differential pressure at representative locations across the stack is measured by inserting a probe, with a Pitot tube at the end, into the stack. The probe transfers the differential pressure between the static and stagnation pressure taps, to the differential pressure device.

The measurement of gas velocity by differential pressure at a location in the stack therefore requires:

a) differential pressure sensing head — Pitot tube, generally of S-type (A.2.1.2) or L-type (A.2.1.1) design, or a 3D Pitot tube, e.g. of spherical (A.2.1.3) or US EPA Method 2G^[18] design;

b) probe — usually integral to the Pitot head, but extension tubes may be used, provided they are fully leak checked prior to use;

c) differential pressure measuring device — either an inclined manometer or digital manometer meeting the requirements of Table 5;

d) stack gas temperature sensor;

e) atmospheric pressure sensor.

For simultaneous measurements of the traverse and fixed point velocities, two sets of items a) to d) are required. The atmospheric pressure sensor can be a standalone one and the same atmospheric pressure readings can be used for both measurement systems. Weather station atmospheric pressure values are acceptable only for MO1 measurements and corrections for altitude should be applied where possible. A quality-controlled flow AMS may be used to provide the fixed point readings. In addition, an accurate measurement of the stack internal diameter and wall thickness are required.

NOTE In the case of a 3D Pitot sensor, the pressures at the pressure taps are transferred in a number of tubes to three differential pressure measurement devices.

• • • 1. Pitot tube

The Pitot tube shall meet the performance requirements specified in Table 5. It may be one of the designs described in this annex, or any other design meeting the performance requirements of this document. For many applications a 1D Pitot tube (e.g. S- or L-type design) can be used. Where there is significant non-axial flow or swirl, a 3D Pitot tube, e.g. spherical, can deliver smaller uncertainty values.

The probes and Pitot tube heads shall be made of material that is not affected mechanically by the temperatures within the stack.

The elements of the Pitot tube probe which remain external to the stack shall have a mechanism to identify the tube orientation within the stack and a methodology to mark the distances that the probe has been inserted into the stack. This enables the correct positioning of the sensing head within the required measurement points inside the duct.

The Pitot tube shall be calibrated by the manufacturer or an accredited calibration laboratory providing a corresponding calibration certificate (see Table 5). The calibration shall have metrological traceability.

The validity of the calibration can be checked periodically by comparison with a calibrated reference Pitot tube according to the procedure described in the informative Annex I.

If a Pitot tube (commonly an S-type) is used in a configuration with a closely coupled gas-sampling probe, then the device is calibrated preferably in this configuration. This document recognises this cannot always be possible or practical as test laboratories can have several measurement system components (differential pressure measurement and readout devices, Pitot tubes, filter holders, probes etc) which are interchangeable between different systems. They can also use different lengths of probes and lines to connect systems together. In this case the Pitot tube can be calibrated separately.

The use of standard Pitot calibration factors and performance based purely on Pitot tube design criteria is not allowed within this document. Devices shall be calibrated and have been subject to suitable performance evaluation.

Any other flow conditions affecting the differential pressure (e.g. turbulence in the case of S-type tubes) shall be properly taken into account in calibration.

The tubing or lines used to connect the Pitot tube to the differential pressure device should be as short in length and as thick as possible.

NOTE The calibration requirement can be fulfilled by a corresponding traceable certificate prepared by the manufacturer according to this document.

• • • 1. Differential pressure measurement device

The measurement of the differential pressure between the pressure taps of the Pitot sensor shall be made using a differential pressure meter; this can either be an inclined manometer or an electronic differential-pressure gauge.

For liquid manometers a correction for ambient temperature shall be carried out in order to account for liquid density changes with temperature. Formula (A.1) can be used to calculate the corrected height:

(A.1)

where

h_s is the corrected height of the indicating fluid to standard temperature;

h_t is the height of the indicating fluid at the temperature when read;

p_s is the density of the indicating fluid at standard temperature;

p_t is the density of the indicating fluid at the temperature when read.

The differential pressure gauge shall be calibrated in a range appropriate for the application. The calibration shall be metrologically traceable to SI units.

The range of the manometer associated to the Pitot tube shall be adapted to the values of static and differential pressures measured.

A damping device can be required if the stack conditions cause oscillations in the readout of the differential pressure which make it impossible to achieve the required uncertainty in the pressure reading. Damping is required if the magnitude of the fluctuations in the differential pressure reading (peak-to-peak) are >10 % of the mean reading at the measurement location. The damping device can consist of a damping pot or a capillary tube installed in a liquid manometer, or other device provided or recommended by the manufacturer. The damping device shall have been demonstrated not to cause a systematic effect on the differential pressure reading.

NOTE ISO 3966:2008,^[3] Annex D, describes a methodology for damping a liquid manometer system using a capillary tube.

• • • 1. Measurements of stack gas conditions

The calculation of stack gas velocity from the differential pressure measurements made with a Pitot tube requires knowledge of the stack gas density, which is determined from the temperature, static pressure and gas molar mass.

In order to refer the measured flow to standard conditions (see 10.6) measurements of the following parameters of the stack gas are required:

— temperature;

— water vapour;

— where necessary, measurements of the oxygen or carbon dioxid content.

These data may be obtained from calibrated automated measuring system installed on the stack.

A temperature sensor shall be used to measure the stack gas temperature at the Pitot measurement location. This sensor shall be calibrated and shall be able to meet the performance requirements given in Table A.1.

The measurement of oxygen concentration within the stack gas shall be made in accordance with the applicable standard (e.g. EN 14789^[9]).

The measurement of water vapour within the stack gas shall be made in accordance with the applicable standard (e.g. EN 14790^[10]).

In the case of dry air (<1 % humidity) the molar mass may be assumed to be 29 kg/kmol.

• 1. Calculation
     1. Determination of velocity using differential pressure devices

The principle of the determination the velocity of a gas at a measurement point of the measurement plane, by means of a Pitot tube, consists of measuring the dynamic pressure at this point with a Pitot tube associated with a manometer. The local velocity v_i at measurement point i is then calculated according to Formula (A.2):

(A.2)

where

v_i is the local velocity, in m/s;

is the average dynamic pressure measured at the measurement point i of the measurement plane, in Pa;

K is the coefficient of the Pitot tube which includes the Pitot calibration factor and constant values relating to Pitot design;

ρ is the density of the gas effluent under the conditions of temperature and pressure of wet gas, in kg/m³.

The average dynamic pressure is equal to the arithmetic mean of the n measurements of dynamic pressure carried out at each measurement point i as given by Formula (A.3):

(A.3)

Direct determination of is possible using a suitable digital manometer which provides an average reading of dynamic pressure over the required averaging period.

3D US EPA Method 2F^[17] can be used to calculate the velocity, the yaw, the pitch and the gas volume flow.

• • 1. Density of the stack gas

In the duct, the density ρ of the gas effluent under ambient conditions of temperature and pressure is given by Formula (A.4):

(A.4)

where

ρ is the density ρ of the gas effluent under ambient conditions of temperature and pressure, in kg/m³;

M is the molar mass of wet gas effluent, in kg/mol;

p_c is the absolute pressure in the duct in the measurement plane, in Pa;

R is the gas constant, i.e. 8,314 J/(K mol);

T_c is the gas temperature in the duct, in K.

• • 1. Absolute pressure of gas

The absolute pressure in the duct is given by the measurement of the atmospheric pressure on the site and the static pressure measured in the duct. The static pressure shall be measured in at least one point of each measurement line.

To improve quality of measurement, it is recommended that several instantaneous measurements of the static pressure be made to take into account the variability of the pressure and the repeatability of measuring.

The static pressure, which is used in the calculation of the absolute pressure, is the arithmetic mean of the average static pressures recorded on each measurement line (two diameters in the case of a circular duct) therefore the arithmetic mean of the whole of the recorded values of static pressure. The absolute pressure in the duct in the measurement plane is calculated according to Formula (A.5):

(A.5)

where

p_c is the absolute pressure in the duct in the measurement plane, in Pa;

p_atm is the atmospheric pressure, in Pa;

is the average static pressure in the measurement plane, in Pa.

The average static pressure is equal to the arithmetic mean of the average static pressures measured at each measurement point (at least one per measurement line) as given by Formula (A.6):

(A.6)

where

(A.7)

where

is the average static pressure, in Pa

is the i₂th measurement of static pressure at the measurement point i₁, in Pa; i₂ = 1 ... n₂

is the mean static pressure at the measurement point i₁, in Pa; n₁ is at least equal to the number of measurement lines explored in the measurement plane; i₁ = 1 ... n₁

• • 1. Molar mass of gas

The molar mass M of wet gas effluent under the conditions of pressure and temperature of the duct in kg/mol, is given by Formula (A.8):

(A.8)

where

M is the molar mass M of wet gas effluent under the conditions of pressure and temperature of the duct, in kg/mol;

M_B is the molar mass of component B, in kg/mol, for component number B = 1 ... q;

φ_B is the volume concentration of component B, for component number B = 1 ... q.

The molar mass of gas is given, in general, with a good approximation starting from measurements of the contents of O₂, CO₂, and water vapour, by applying Formula (A.9):

(A.9)

where

M is the molar mass of wet gas stream, in kg/mol;

is the volume concentration of O₂ in the gas stream in wet gas, in %;

is the volume concentration of CO₂ in the gas stream in wet gas, in %;

is the volume concentration of water vapour in the gas stream, in wet gas, in %.

The value of corresponds to the nitrogen content, assuming that the components other than O₂, CO₂ and water vapour, expressed as volume concentration in wet gas, are insignificant.

The bias introduced by this approximation of the density is variable according to the nature and concentrations of the components not taken into account

This assumption holds true for most combustion sources.

1. 
   (normative)
   
   Vane anemometer
   1. Principle of vane anemometer

The measuring principle of the vane anemometer depends on the proportionality of the rotating velocity to the flow velocity  of the fluid into which it is inserted.

The simplified principle of a vane anemometer is shown in Figure B.1.

Key

Figure B.1 — Vane anemometer

Assuming no friction and a massless vane wheel, the resulting characteristic is purely geometric.

Given input variables are:

a) approach velocity in axial direction;

b) frequency, f;

c) geometry of the vane wheel or a vane wheel blade with its centre in spherical coordinates, r_Sp;

d) number of blades e.g. n_F = 4.

Pulsatance, ϖ = 2π(f/n_F), and peripheral velocity, v_t = ϖr_Sp, are also used for the calculation. See Figure B.2.

Key

flow velocity

α pitch of blade

f scan frequency

ω angular frequency

r_Sp geometry of the vane wheel

v_t peripheral velocity

Figure B.2 — Vane anemometer principle

The correct value for the r_Sp of the blade is of particular importance. Given that the acting compressive force can be determined as a constant acting force over the total area, this can be replaced by an acting single force in the centre of the blade, ruled area.

The vane wheel is activated via an inductive proximity switch. With the introduction of a second inductive switch, the rotational direction can be measured and the direction of flow determined. This type of angular frequency measurement has no braking effect on the vane wheel. Contamination has no impact on the pulse recognition. Due to the low mass of the vane wheel, depending on type in range of 0,1 g, it adjusts its angular frequency in the millisecond range to the velocity variations.

• 1. Calculation
     1. Background

The axial velocity in the frictionless case, , can be calculated using Formula (B.1):

(B.1)

where

v_t is the peripheral velocity, v_t = ϖr_Sp;

α is the pitch of blade;

r_Sp is the geometry of the vane wheel.

• • 1. Lowest range limit

However, in practice, friction and mass play a significant role. As a result, the lowest range limit is a minimum velocity, the so-called start-up velocity, v₀, required to overcome friction in the bearings and inertia of the vane wheel. Below this value vane anemometers shall not be used.

Furthermore, the basic Formula (B.1) can be expanded to Formula (B.2):

(B.2)

where is a non-linear calibration factor, which depends on the density, ρ₀, and the dynamic viscosity, η_dyn.

• • 1. Highest range limit

The upper limit for measuring flow velocity with a vane wheel is specified by the manufacturer and based on the demand for a defined lifetime of the sensor in respect of increasing mechanical stress caused by increasing velocity, the upper limit of the manufacturer’s calibration facilities and the signal output unit of the vane wheel sensor.

The rotation of a vane wheel is, according to B.2.2, only dependent on the density and viscosity of the flue gas. Ambient pressure, temperature, humidity or possible dust particles only have an influence on the axial velocity when flue gas density or viscosity vary significantly and only relevant in regards to the durability of the vane anemometer , i.e. susceptibility to wear and tear.

Relevant for the force acting on the vane wheel, F, and the resultant peripheral velocity v_t is the dynamic pressure on the vane wheel, p_dyn:

(B.3)

where

ρ is the density during calibration;

is the axial velocity during calibration;

c is a constant (device specific).

If, when measuring in the flue gas, the actual density ρ₁ deviates from the density during calibration ρ₀, then values displayed from the vane wheel when measuring in flue gases with density ρ₁ equate to an actual velocity

(B.4)

where

is the axial velocity;

is the measured value of axial velocity at vane;

Δv₀₋₁ is the adjustment of characteristic curve:

The adjustment of characteristic curve is given by

(B.5)

where

v_0,0 is the start-up velocity;

v_0,1 is the corrected start-up velocity for actual flue gas density;

ρ₀ is the density during calibration;

ρ₁ flue gas density.

This means that the axial velocity (B.2) is offset by the difference of the start-up velocity.

• 1. Calculation of the uncertainty and calibrations
     1. General

The uncertainty is affected by the precision of the measurement of the angular frequency (and thus of the scan frequency) and by the reproducibility of the output quantities.

For a vane anemometer with four blades, the precision is (ω/4) ±1 impulse, where ω is the angular frequency.

The uncertainty of the flow velocity u(v) of a vane anemometer results from the absolute error of measurement (reproducibility of output quantities) e_P and the uncertainty of the scan frequency u’(v), in m/s:

(B.6)

or

(B.7)

where

f is the scan frequency;

v_max is the highest velocity range limit of the vane type;

v_min is the minimum velocity range limit of the vane type (start-up velocity);

f_max is the maximum frequency of the vane type;

e_P is between 0,01 m/s and 0,02 m/s, for the vane anemometers tested during validation studies .

NOTE The value of e_P depends on the design of the vane. It can be estimated, for example, according to DIN 1319-3.^[22] The manufacturer is required to provide the value.

In sum, the uncertainty of flow measurement is given by Formula (B.8):

(B.8)

• • 1. Example

Minimum velocity, v_min = 0,5 m/s

Maximum velocity, v_max = 40 m/s

e_p = 0,02 m/s

Diameter of vane, 2r_Sp = 22,8 mm

Ratio (f/v) = 29,25

Maximum frequency of vane, f_max = 1 770 Hz

(B.9)

u(v) = u’(v) ± e_P = (0,04 ± 0,02) m/s (B.10)

1. 
   (normative)
   
   Tracer gas dilution method determination of volume flow rate and average velocity
   1. Tracer gas by dilution
      1. Principle of the use of tracer gas injection

Tracer gas injection is used to measure the volume flow rate in a duct. A tracer gas that is not usually present in the flue gas is injected at a known constant rate (Figure C.1, label 1). The tracer gas is mixed over the cross-section of duct, and mixing is enhanced by obstructions such as fans or bends, which create extra turbulence (Figure C.1, label 2). The diluted tracer gas is measured downstream of the injection point to give the volume flow rate (Figure C.1, label 3), provided that:

— the tracer gas is adequately mixed in the stack gas;

— there is no loss of tracer gas from adsorption on to duct walls or removal by the plant abatement system;

— there is no tracer gas present in the stack gas prior to injection or the background concentration can be measured accurately. Background level measurements shall meet the same performance requirements as the tracer gas concentration measurements (Table C.1). The measurement of any background should be included in the tracer uncertainty calculations.

Key

1 tracer injection lance

2 process fan

3 tracer sampling and analysis

Figure C.1 — Tracer gas dilution principle

• • 1. Tracer gas injection

A controlled injection of the tracer gas is required. The tracer injection flow rate is calibrated to traceable standards. The injection location should be as far upstream as possible and preferably upstream of plant components such as fans, compressors, ductwork bends and other plant items that promote turbulent mixing. Inert tracer gas can be injected into the inlet fuel or air upstream of a combustion process. Multiple point injection is preferred over single point injection, e.g., utilising an existing compressor water wash system or an existing emissions sampling grid in an exhaust duct. Alternatively, a temporary grid injection system can be used to achieve a multi-point injection across the combustion air inlets(s) for example.

Select an upstream location for introducing the tracer gas into the stack duct flow in order to promote mixing of the tracer with the duct flow. As a guideline, the injection point(s) should ideally be located at least 10d_s, where d_s is stack diameter, upstream of the tracer concentration measurement location. If possible, the tracer injection should be upstream of a source of mixing such as a fan or bends in the duct. The tracer injection shall be downstream of any process operation that may affect the tracer concentration.

For a given injection arrangement, it shall be demonstrated that there is no significant stratification of the tracer gas at the downstream location by means of a one off multi-point duct survey conducted according to ISO 15259. There should ideally be no additional flow inputs between the injection and the measurement locations, and no leaks from the duct. However, in-leakage that is fully mixed prior to the concentration measurement (as confirmed by the duct homogeneity testing according to ISO 15259), or out-leakage that takes place after the tracer is fully mixed, would not cause a measurement error.

Alternatively, grid sampling of the tracer concentration, at centres of equal area, according to ISO 15259, can be performed in order to ensure that a representative well mixed sample is obtained.

The duration of the injection shall be at least 4t_r, where t_r is the response time of the tracer gas analyser used for the measurement of the tracer gas, to ensure that a meaningful stable concentration response can be obtained. See Figure C.2 for example.

Key

Figure C.2 — Typical tracer gas response

Injection may be accomplished using a single lance or by multiple lances connected to a manifold off the tracer flow meter. Flow through each single or multi-point injection lance should be equal, maintaining the same injection line geometry where possible. Equal flow rates through multiple lances can be achieved by designing each lance to have the same flow resistance. and each lance can have either a single or multiple release points. The injection system shall be leak tight.

• • 1. Tracer gas concentration measurement

The tracer gas measurement equipment consists of a gas monitor (the analyser used for the measurement of the tracer gas) and suitable sampling equipment. The sampling system shall not react with or otherwise alter the concentration of the tracer gas. Any interference effects on the tracer analyser shall be known and quantified and included in the uncertainty assessment of the tracer concentration. The tracer gas analyser shall be calibrated at a suitable frequency to ensure drift across the test period is within the limits specified in Table C.1.

The tracer gas concentration shall have a traceable calibration. The choice of tracer gas should include consideration of the components of the stack gas, both in terms of the tracer gas itself and any interferents to the measurement of the tracer gas. The selection of the tracer gas is also influenced by the availability of measurement equipment able to measure it with a sufficient level of certainty. The tracer gas concentration measurement shall meet the performance requirements of Table C.1. The environmental impact of the release of the tracer gas shall be considered.

The tracer gas concentration may be measured either on a dry basis, following removal of water vapour from the sample to a dew point less than 5 °C, or on a wet basis using a heated sample and analysis train. This needs to be accounted for when calculating the stack flow rate. The stack flow rate is determined under the same conditions that the measurement is made, i.e. if the tracer is measured on a dry basis the flow determined is also on a dry basis.

Sampling should be as far downstream of the injection location as possible and preferably at a measurement location that has been demonstrated to be homogeneous according to ISO 15259 for the measurement of pollutant concentrations. The sample probe may be single or multi-hole, provided that there is no significant stratification of the tracer at the measurement location. Alternatively, grid sampling of the tracer concentration, at centres of equal area, according to ISO 15259, can be performed in order to ensure that a representative, well mixed sample is obtained. In this case, the sample gas flow rates through the individual sample probes should be equal which can be achieved by designing each sample line to have the same flow resistance.

• • 1. Tracer gas calibration equipment
       1. Tracer gas injection

A controlled injection of the tracer gas is required. The injection flow rate shall be determined using a mass or volume flow rate meter with a traceable calibration. The expanded uncertainty of the tracer flow rate measurement shall be ≤1 % of value based on the calibration certificate uncertainty and taking into account any relevant influencing parameters such as ambient temperature and pressure. The tracer injection line shall be leak tested in the field, from the flow meter inlet to the injection lance.

• • • 1. Tracer gas concentration measurement

The tracer gas measurement equipment consists of a gas concentration analyser and suitable sampling equipment. The sampling system shall not react with or otherwise alter the concentration of the tracer gas. The tracer gas analyser shall not be affected by other components in the stack gas. The tracer gas analyser shall be calibrated before each flow measurement.

Following the flow rate measurement, the tracer analyser shall be checked using calibration gas at the span value, in order to correct for allowable drift. Table C.1 gives the performance requirements of the tracer gas analyser.

Table C.1 — Tracer gas concentration measurement performance requirements

• • • 1. Calculation of stack gas flow rate from tracer injection results

The stack gas flow rate at sample oxygen content and moisture under standard conditions shall be calculated according to Formula (C.1):

(C.1)

where

is the stack gas flow rate at sample oxygen content and moisture under standard conditions, in m³/s;

q_m,t is the tracer mass flow rate, in kg/s;

ρ_t,0 is the tracer density under standard conditions, in kg/m³;

φ_t is the tracer volume concentration at sample oxygen content and moisture content.

The tracer gas volume flow rate q_m,t may instead be measured directly, replacing q_m,t/ρ_t,0 in the above formula.

Note that the background concentration is subtracted from the measured concentration to obtain φ_t.

• 1. Uncertainty of the calibration result
     1. General

The uncertainty of the derived stack flow rate is the combined uncertainty of the injection mass flow rate q_m,t and the tracer volume concentration φ_t. The uncertainty of the tracer volume fraction shall include the uncertainty related to the mixing quality of the tracer within the stack gas unless a multi-point flue gas sample is taken according to ISO 15259.

• • 1. Uncertainty of concentration measurement

The performance characteristics considered in the uncertainty analysis are dependent upon the technology employed for the tracer concentration measurement. The uncertainty analysis for the concentration measurement shall follow the approach defined in EN 15267-3 for any relevant uncertainty contributions.

For example a common contributions is the calibration gas uncertainty u_cal which is assigned a rectangular probability distribution to give the absolute standard uncertainty as follows:

where c_cal is the calibration gas value and x_cal is the range of the calibration value as a percentage value.

Other performance characteristics can be included in this analysis, e.g. linearity and cross-interference, if these are significant. Linearity errors can be minimized by selecting a calibration value that is close to the expected tracer concentration value.

The uncertainty of the background tracer concentration is evaluated in the same way and the absolute uncertainties of the tracer concentration, during the injection, and the average background tracer concentration (pre- and post-injection) can then be combined using the root sum of squares approach. If these uncertainties are identical then the uncertainty of the tracer concentration can be multiplied by 1,414.

• • 1. Uncertainty of tracer gas mixing

The standard uncertainty due to imperfect mixing of the tracer gas is calculated from an ISO 15259 tracer concentration survey conducted when injecting the tracer at a constant flow rate, unless a multi-point sample is extracted according to ISO 15259 in which case a representative sample is obtained, and the mixing uncertainty can be neglected.

The standard uncertainty u_mix is the standard deviation of the measured grid values s_grid:

A more accurate assessment can be obtained by correcting for temporal variations in the process (determined from the fixed reference measurement required under ISO 15259). Temporal variations can be accounted for by first normalizing each grid concentration measurement, multiplying by each normalization factor f_n,i:

where

φ_ref,i is the reference volume concentration (fixed location) corresponding to the sampling periods at each grid point;

is the mean reference volume concentration from all of the sampling periods.

A velocity measuring instrument can be used as the reference if a second tracer gas analyser is not available. In this case, since the tracer volume concentration is inversely proportional to the flow, the normalization factor is inverted:

Following normalization, the standard uncertainty is calculated as before using the normalized volume concentrations, s_n,grid:

The standard uncertainty due to mixing is then combined with other standard uncertainties for the volume concentration measurement to give the overall standard uncertainty of the volume fraction measurement:

where u_i is the uncertainty of influencing factor i.

• • 1. Uncertainty of tracer injection rate

The performance characteristics considered in the uncertainty analysis are dependent upon the technology employed for the flow measurement which shall have a traceable calibration. Common contributions are considered below.

The uncertainty contribution from the flow calibration of the instrument is based on the expanded uncertainty declared on the calibration certificate, which is usually expressed at 95 % confidence, assuming that the uncertainties are normally distributed. The standard uncertainty u_cal is therefore obtained by dividing the expanded uncertainty by a coverage factor of 2.

The uncertainty u_comb of the values of the combined sensor and transducer used to measure the tracer gas concentration shall include the uncertainty related to linearity and repeatability of the combined sensor and transducer:

where c_comb is the value of the tracer gas concentration and x_comb is the range of the values as a percentage value.

The combined mass flow rate uncertainty, , is the sum of the squares of the standard uncertainties.

where u_i is the uncertainty of influencing factor i.

• • 1. Uncertainty of stack flow rate

The standard uncertainty of the stack volume flow rate is based on the combined uncertainties of the measured tracer volume concentration and the tracer mass flow rate or the tracer volume flow rate u_qv,t. In general, a sensitivity coefficient C_sens is applied to each uncertainty, which can be obtained from the partial derivatives with respect to the parameter under consideration, e.g.

The overall standard uncertainty is then the sum of the squares of each uncertainty contribution multiplied by its sensitivity coefficient:

In this instance, since the stack flow is simply proportional to each of the parameters, it is acceptable to combine the relative standard uncertainties directly:

The standard uncertainty obtained in this way is then multiplied by a coverage factor of k = 2 to give the expanded uncertainty at 95 % confidence.

1. 
   (normative)
   
   Transit time tracer gas method determination of average velocity
   1. Existing standards

Even though heavily outdated as to measuring equipment, BS 5857-1.4:1980^[20] and BS 5857-2.4:1980^[21] give a thorough view of the physical background of the method. ASME MFC-13M-2006^[15] gives only a short qualitative description on the use of the transit time method.

• 1. Transit time method
     1. Principle of the method

A very small amount of tracer material is momentarily injected into the stack gas flow. After the tracer pulse has mixed over the cross-section of the flow, its transit time between two measurement points placed on a suitable straight duct section is measured. The measurement principle is shown schematically in Figure D.1.

Key

a tracer pulse input

b data logging unit

c AMS signal

d mixing length

e measurement points

f measurement section

Figure D.1— Principle of tracer time of flight flow measurement technique

• • 1. Choice of the tracer

The tracer shall follow the stack gas flow in the duct. It may be gas or sufficiently small particles. It may be radioactive or inactive. The radioactive tracer has the advantage that it can be detected through the stack wall. The disadvantage is the possible difficulty or delay in getting a licence from the national radiation safety authorities for use, even when a radiotracer with a very short half-life is used.

Suitable inactive tracers need no safety licence, but are from a technical point of view more difficult to implement. The inactive tracer shall be such that the time constant in its detection is of the order of few milliseconds or less. This requirement comes from the fact that the transit times to be measured accurately are in the range from a couple of seconds to some fractions of a second. In the case of stacks and ducts ports can be installed through which the detector can be inserted into the duct or stack,. This enables the use of inactive tracers, detectable for instance by ultrasound or by optical methods.

• • 1. AMS flow calibration procedure

The reference flow value obtained by the transit time method is compared with the simultaneous AMS flow signal. In order to obtain the calibration result on a flow level, several measurement repetitions (normally 7 to 15) are made. The number of flow levels should preferably be 2 to 3 to obtain a good and representative calibration result.

• • 1. Calculation of the volumetric flow rate

The average stack gas velocity is derived according to Formula (D.1):

(D.1)

where

v is the average velocity, in m/s;

L is the length of the measuring section, i.e. the stack length between the two measurement levels, in m;

t is the transit time of the tracer pulse between the two measurement points, in s.

The volume flow rate, q_V, in m³/s, is derived according to Formula (D.2):

q_V = vA (D.2)

where A is the duct cross-section area, in m².

• • 1. Provisions for measurement site
       1. Duct area

The duct cross-section area shall be constant between the two measurement points.

• • • 1. The length of the measurement section

The duct between the two measurement points shall be straight to allow accurate determination of the volume. For transit time measurements with increased accuracy, the duct length between the detection points should be ≥10d_d, where d_d is the duct diameter.

• • 1. Flow condition

The flow shall be turbulent. This condition is by definition always fulfilled in stack gas ducts.

• 1. Minimum requirements
     1. Tracer

The tracer shall adequately mix over the cross-section of the stack gas flow, then having the same average velocity as the stack gas. No significant adsorption of tracer on the duct walls is admissible.

• • 1. Mixing

Tracer shall be mixed over the flow cross-section when it arrives at the measurement section. The degree of mixing required depends on the characteristics of the tracer concentration measurement used. If the measurement yields values that are well averaged over the flow cross-section, the mixing in a rough scale is sufficient. Concentration measurement in a point of the cross-section is the opposite case and requires more thorough mixing of the tracer over the cross-section. Insufficient mixing is normally indicated by a significant increase of the variation between the measurement repetitions.

• • 1. Measurement section

The measurement section where the tracer transit time is measured shall be a straight duct section with a constant cross-section and with sufficient total length. The minimum total length depends on the length of the tracer pulse on the measurement section, stack gas flow velocity, duct diameter, desired measurement uncertainty level, etc. For increased accuracy, the measurement section shall be ≥10d_d.

The volume of the tracer measurement section can be determined by using laser measurement for both the inner diameter and the length of the measurement section. The requirements regarding the measurement of duct area as specified in 7.2 shall be met.

• • 1. Tracer concentration measurement

The concentration of the tracer pulse is measured at the two measurement locations. The method of concentration measurement shall be shown to be linear (≤1,5 % relative to the calibration gas concentration).

The concentration measurement shall not be disturbed by possible rapid variations in stack gas characteristics or in the measurement environment. These are usually achieved by stable plant operating conditions.

Repeated measurements are required to get a representative result.

• 1. Performance requirements
     1. Injection

The tracer pulse injected into the stack gas flow should be so short that it does not dominate the pulse length on the measurement section. Injection times of 0,5 s to 1 s are normally sufficiently short.

• • 1. Measurement of the tracer pulse

The transit times to be measured are normally between some 0,5 s and a few seconds. The time constant of tracer concentration measurement and the data collection interval shall be of the order of some milliseconds to allow accurate reconstruction of the tracer pulse. A concentration measurement that gives an averaged value over the duct cross-section should be preferred because compared with point measurement, it is more tolerant to non-complete mixing and disturbed flow profiles.

• • 1. Calculation of the transit time

The transit time is calculated as the time difference of the concentration pulses measured in the beginning and the end of the measuring section. The time difference is calculated by using the pulse medians that divides the pulse in two parts having equal areas, see Figure D.2.

Key

Figure D.2 — The determination of the transit time

The physically correct way of calculating the time difference is by using the gravity points of the pulses. Pulse medians, however, are recommended because they are less sensitive to the correct background subtraction. In normal measurement circumstances, i.e. a straight measurement section and a duct with a constant cross-section, the difference between the velocities of the pulse gravity point and median have been shown to be insignificant.

• • 1. Calculation of the volumetric flow

The velocity value is obtained by dividing the length of the measuring section L by the transit time t. The value for the flow rate is obtained by multiplying the velocity value by the duct inner cross-section.

The length L is measured by using a suitable measuring rod or a calibrated laser distance meter. The inner cross-section is normally determined by measuring the inner diameter of a circular stack. The diameter is obtained with the most reliability by measuring it in two perpendicular directions by using a laser distance meter. Any measurement of stack dimensions shall be carried out in accordance with 7.2.

• 1. Uncertainty of the calibration result
     1. Calculation principle

The measurement uncertainty is calculated at a confidence level of 95 %. The method of calculation of uncertainty is in accordance with ISO/IEC Guide 98-3 .

The uncertainty of the calibration result is divided into two parts:

a) the uncertainty of the reference flow rate originating from the determination of the volume of the measurement section and the equipment’s measurement of time;

b) the uncertainty of the comparison of the reference flow rate and the indication of the AMS to be calibrated, including the random uncertainty.

• • 1. Uncertainty of determination of the volume and measurement of time

In stack emission measurements, the volume of the measurement section is normally determined by measuring L, the length of the measurement section and D, the inner diameter of the duct. In this case, the equation for the reference volume flow rate is:

(D.3)

The partial derivatives with respect to the different input quantities are:

(D.4)

(D.5)

(D.6)

The standard uncertainty of the reference volume flow rate is obtained by quadratic summing of the standard uncertainties u(x_i) of the different input quantities x_i multiplied by the corresponding sensitivity coefficients ci.

(D.7)

The standard uncertainties of the input quantities are obtained as in a) and b).

a) For a quantity which is obtained as an average of several independent observations with adequate resolution, standard uncertainty is obtained by dividing the experimental standard deviation with the square root of the number of observations (ISO/IEC Guide 98-3, type A).

b) For calibrated measurement equipment, the standard uncertainty is obtained from the calibration certificates by dividing the expressed uncertainty with the coverage factor k, which typically equals 2 (ISO/IEC Guide 98-3, type B).

In the case of transit time measurement, the standard uncertainties are obtained as follows.

— u(L): Calibrated measurement tapes or laser distance measurement device are used to the measurement of L. The uncertainty caused by field circumstances is however, significantly larger than the small uncertainties specified by equipment manufacturers and attainable under laboratory conditions.

— u(D): Under stack conditions, the preferred way to determine D is by using a laser distance measurement device. The measurement should be carried out, if possible, of perpendicular diameters to detect any possible ovality of the stack.

— u(t): Measurement of time by the calibration equipment in itself is easily and accurately calibrated and thus the hardware-based uncertainty is negligible (<0,01 %). The uncertainty of the calculated transit time is dominantly stochastic in nature. It is therefore included in the total statistical uncertainty of the calibration and discussed in the following in relation to the uncertainty of the comparison of reference flow rate and the AMS indication.

• 1. Numerical example of uncertainty calculation in stack flow calibration
     1. Uncertainty of q_V, ref

Length L (in this example 62,577 m):

a) specifications for measuring tape, class 2: U = (1 + 0,2L[m]) mm = >13,2 mm.

b) tape stretch by other than nominal 50 N force: U = 20 N × (5 µm/N)L[m] = >6,3 mm.

c) tape stretch by temperature; u = 10 °C × (12 µm/°C)L[m] = >7,5 mm.

d) transfer of the tape reading value to detector positions; u = 100 mm.

All these uncertainty components associated with L are given by upper and lower limits and probability distribution between the limits is assumed to be rectangular (ISO/IEC Guide 98-3, type B, case c). Standard uncertainty of each of these input estimate components is obtained by dividing the limit values by . The square root of the sum of the squares of these uncorrelated values gives standard uncertainty for L, u_L = 58,5 mm (transfer of tape reading is predominant).

Inside diameter D (= 3 580 mm).

No ovality of the stack was detected in this case.

The square root of the sum of the squares of these uncorrelated components 1) to 3) gives standard uncertainty for D, u_D = 2,9 mm.

Transit time t (average, 3,645 s).

— Detectors and the signal amplifiers have a small deviation of response time. Associated (expanded) uncertainty with them is u = 1 ms, (normal distribution, k = 2).

— Data-logging apparatus has time-measuring (expanded) uncertainty u = 100 µs or 0,01 % of the length of the time interval (normal distribution, k = 2).

These uncertainties associated with transit time measurement represent a type A evaluation according to ISO/IEC Guide 98-3. Standard uncertainty of each of these input estimate components is obtained by dividing the given u values by the coverage factor, k = 2. The square root of the sum of the squares of these uncorrelated values gives standard uncertainty for t, u_t = 0,53 ms.

Correlation of input quantities associated with q_V, ref: none of the input quantities for definition of q_V, ref are considered to correlate to any significant extent.

Combined uncertainty associated with q_V, ref: the effect of each component u_L, and u_t on q_V, ref is calculated by multiplying each u_i with the corresponding sensitivity coefficient c_i. The square root of the sum of the squares of these effects gives the standard uncertainty for q_V, ref,  = 0,32 m³/s.

• • 1. Expanded uncertainty

See Table D.1.

The reported expanded uncertainty of measurement is stated as the combined standard uncertainty of measurement multiplied by the coverage factor k = 2,37, which for a t-distribution with ν_eff = 8 effective degrees of freedom corresponds to a coverage probability of approximately 95 %. The standard uncertainties have been determined in accordance with ISO/IEC Guide 98-3.

Table D.1 — Uncertainty budget

1. 
   (normative)
   
   Calculation of flue gas volume flow rate from energy consumption
   1. Principle

This annex describes a procedure for the calculation of flue gas volume flow rate. The general method is to multiply the energy consumption by a fuel factor in order to obtain the dry stoichiometric stack gas flow rate at standard reference conditions (0 % oxygen volume concentration, 273,15 K and 101,3 kPa). The energy consumption may be determined directly, by measurement of fuel flow rate and specific energy, or indirectly from the plant output and thermal efficiency.

For mass emissions reporting, the dry stoichiometric flow rate rate is then corrected to a given reference oxygen content, for subsequent multiplication by emissions concentrations reported at the same reference conditions.

For dispersion modelling, and a number of other purposes, the dry stoichiometric flow rate is recalculated to the prevailing or expected stack conditions of oxygen, moisture content, temperature and pressure.

The required inputs, calculation steps and outputs are shown in Figure E.1 and are specified in later sections of this annex.

Key

1 inputs: fuel flow, q_m, F, in kg/s, with net specific energy, e_(N), in MJ/kg; input: gas release (fuel factor), S [m³/MJ]

2 calculation: process heat release, Φ_(N)F [MW] = q_m, F e_(N) = P/η

3 inputs (alternative): energy production, P, in MW, with thermal efficiency η

4 calculation: flue gas volume flow rate q_V,0d [m³/s] = S Φ_(N)F (at 273,15 K, 101,3 kPa, 0 % oxygen volume concentration, 0 % H₂O volume concentration)

Figure E.1 — Principle of calculation of stack gas flow from energy consumption

• 1. Fuel factor
     1. Fixed factors for commercially traded fossil fuels

The volume of flue gas generated by the combustion of a known quantity of fuel is required. On a thermal input basis, the fuel factor does not vary substantially for a given fuel type and a fixed value is often sufficient. The volume of dry, stoichiometric flue gas per MJ of net supplied energy, S, is given in Table E.1 for a range of fuels. An estimate of the uncertainty of the fuel constant is also given in the table based on a comparison of the correlated values with an exact mass balance approach for a wide range of samples. The fuel factors have been derived in accordance with EN 12952-15^[7].

The natural gas fuel factor is appropriate for Group H natural gas (EN 437^[6]) provided that this is also natural gas in which the methane content is higher than 80 %.

Table E.1 — Fossil fuel factors

NOTE Relative uncertainty, U_rel, is quoted at 95 % confidence assuming a normal data distribution and a coverage factor of 1,96 unless otherwise stated.

The gas oil factor is suitable for gas oil, diesel, light distillate and kerosene. The fuel oil factor is appropriate for other commercially available petroleum products from light to heavy fuel oil.

The hard coal factor is appropriate for commercially traded hard coal.

A lower uncertainty may be achieved for these fuels by applying a correction for the net specific energy (NSE) or by deriving a fuel factor from the fuel composition as described below.

Fuel factors for wet biomass are given in Table E.2, again derived in accordance with EN 12952-15.^[7] The uncertainty relates to a moisture content variation of ±10 %, e.g. a moisture content of 30 % mass fraction covering a range of moisture contents from 20 % to 40 %. The uncertainty increases non-linearly at high moisture contents.

Table E.2 — Biomass fuel factors

• • 1. Factors corrected for specific energy

A wider range of fuels may be considered, and a lower uncertainty can be achieved, by applying a correction for the NSE of the as-received fuel. "As-received" indicates that the fuel heating value is reported on the basis that all moisture and ash-forming minerals are present.

The NSE correction is derived from EN 12952-15:2003,^[7] Annex A:

(E.1)

where

e_(N) is the NSE of the as-received fuel, in MJ/kg;

a, b correction factors, see Table E.3.

Table E.3 — NSE correction factors

The natural gas fuel factor is appropriate for Group H natural gas (EN 437^[6]) provided that this is also natural gas with a methane content that is higher than 80 %.

For gaseous fuels, it may be more convenient to employ the volumetric NSE (MJ/m³ at 0 °C) in Formula (E.1) in which case a = 0,2 and b = 0,234. This approach is not suitable for low specific energy fuel gases for which the fuel factor shall be determined from the gas composition according to EN 12952-15,^[7] Clause 8.

For liquid fuels, this approach is suitable only for light petroleum fractions. Other liquid fuels should be assessed using the measured composition and heating value as described in E.2.3.

For solid fuels, the mass fractions of ash yield, w_ash, and moisture, , in the as-received fuel need to be taken into account using the dry, ash free, fuel mass fraction, w_f, where:

(E.2)

If the NSE of the fuel stream is variable, the uncertainty associated with the fuel variability shall be estimated from multiple fuel samples. This method is not suitable for fuels with an ash yield greater than 20 % mass fraction, in which case the assessment should be based upon the measured composition and heating value as described in E.2.3.

• • 1. Factors derived from fuel composition

For solid and liquid fuels, EN 12952-15:2003,^[7] 8.3.4.2 also defines a method for determining the mass specific fuel factor, q_V,0d, in m³/kg, from the as-received fuel composition based on an ultimate analysis:

q_V,0d = 8,893 0w_C + 20,972 4w_H + 3,319 0w_S – 2,642 4w_O + 0,799 7w_N (E.3)

where w is the mass fraction of an individual fuel component in the supplied fuel (as received) and C, H, S, O and N are the elements carbon, hydrogen, sulphur, oxygen, and nitrogen, respectively.

This is then divided by e_(N) to obtain the energy specific fuel factor S:

(E.4)

For inhomogeneous solid fuels, for which it is difficult to obtain representative samples, it is recommended that the measured specific energy be checked using a calculated value of e_(N), in MJ/kg, using the following formulas.

e_(N) = e_(G) – 21,22w_H – 0,08 × (w_O + w_N) – 2,442 5 (E.5)

Formula (E.5) is from ISO 1928^[2] and requires e_(G), the gross specific energy, either measured or calculated as follows.

e_(G) = 34,1w_C + 132,2w_H + 6,86w_S − 12w_O − 12w_N − 1,53w_ash (E.6)

Formula (E.6) is from Reference [24]. Note that methods for obtaining the fuel composition on an “as-received” basis are described in ISO 1170^[1].

If the composition of the fuel stream is variable, the uncertainty associated with the fuel variability shall be estimated from multiple fuel samples. In any case, it is recommended that multiple fuel samples be considered in order to minimize the uncertainty contribution associated with the analysis of composition.

• 1. Energy consumption

For gas and liquid fuels, the energy consumption, Φ_(N)F, in MW, may be determined directly from the metered fuel consumption, q_m, F, in kg/s, and the measured or supplied NSE, e_(N), in MJ/kg.

(E.7)

Quality ensured metering and specific energy determination has an estimated worst case expanded uncertainty of ±1,6 %, although lower values may be used if justified.

For solid fuels with a directly measured fuel consumption rate using, for example, calibrated gravimetric feeders, and a stable fuel composition (specific energy), a similar uncertainty is achievable. However, in many situations, the instantaneous fuel consumption and specific energy are not available, in which case the energy consumption shall be derived from the plant energy production, P, in MW, and the fractional thermal efficiency, η:

(E.8)

The uncertainty of the thermal efficiency shall be justified for compliance purposes. For example, a coal-fired power plant, with an established heat accountancy and fuel management system, has an expanded uncertainty of the instantaneous thermal efficiency of circa ±5 %. This may be reduced to circa ±3 % with online correction of the efficiency calculation to account for changes in plant operation and ambient conditions. For heat-producing plants with typical thermal efficiencies of about 90 %, this uncertainty is below 1 %. A biomass steam-generating plant has a higher absolute thermal efficiency and can therefore achieve a lower thermal input expanded uncertainty of typically 2 % to 3 %, provided that the heat output measurement has a low uncertainty.

• 1. Calculation of flue gas volume flow rate

The stoichiometric dry flue gas volume flow rate at 273,15 K and 101,3 kPa, q_V,0d, in m³/s, is calculated from the fuel factor, S, and the thermal input Φ_(N)F:

q_V,0d = S Φ_(N)F (E.9)

The fuel factor is specified at stoichiometric conditions of 0 % oxygen. For mass emission reporting, this flue gas volume flow rate is corrected to the required standard reference oxygen content:

(E.10)

where is the dry oxygen reference condition for the plant as a dry volume concentration. For boilers, this is normally 0,03 for gas and liquid fuels and 0,06 for solid fuels and for gas turbines this is normally 0,15.

Additional corrections are required when calculating the flue gas flow rate at stack gas conditions:

(E.11)

where

is the flue gas oxygen content, dry volume concentration;

is the flue gas water content, wet volume concentration;

T is the flue gas temperature, in K;

p is the flue gas pressure, in kPa.

• 1. Performance requirements

The target overall performance requirement for the dry flue gas flow rate is given by fuel type in Table E.4 for which the method of determination is described in E.4 from the fuel factors listed in E.2 and the energy consumption calculation described in E.3. The performance requirements are given as expanded uncertainties at 95 % confidence level.

Table E.4 — Performance requirements of the calculation approach

The performance requirements for the main calculation inputs are given in Table E.5.

Table E.5 —Performance requirements of main input parameters

• 1. Example of uncertainty calculations
     1. Example 1 — Coal-fired power plant

The flue gas volume flow rate at a reference condition of 6 % oxygen volume concentration, dry, 273,15 K, 101,3 kPa is required at the base load operating condition for calculation of a mass release.

The oxygen correction is

The fuel factor, S = 0,256 m³/MJ, for hard coal is given in Table E.1 and the thermal input is determined from the plant electrical output, P = 500 MW electric, and thermal efficiency, η = 0,40, to give, at reference conditions, the volume flow rate:

Since the relationship between the parameters is linearly proportional, it is sufficient to combine the relative uncertainties using a simple root mean square approach and a coverage factor of 2,0.

Fuel factor S, relative uncertainty for hard coal (Table E.1): 1,0 %

Power output, P, relative uncertainty: 0,25 %

Thermal efficiency η, relative uncertainty: 2,5 %

Combined flue gas flow q_V, relative expanded uncertainty:

• • 1. Example 2 — Biomass fired combined heat and power plant

The flue gas volume flow rate at a reference condition of 8 % oxygen volume concentration, dry, 273,15 K, 101,3 kPa is required for the calculation of a mass release rate.

The oxygen correction is

The biomass fuel has a mean moisture content of 50 % mass fraction with a typical variation of ±10 %. Since the uncertainty associated with the moisture variation (Table E.2) is higher than desired, the deliveries are sampled and a moisture analysis is performed on each delivery to an expanded uncertainty of 5 % (reported by the laboratory). The ash yield is consistently low (2 % mass fraction on a dry basis) and the specific energy and the ultimate analysis of the dry fuel is invariable (a homogeneous biomass source). The standard uncertainty of the fuel factor for each measurement may be estimated as follows.

Note that the composition and/or specific energy of the samples may need to be determined for a less homogeneous fuel.

The thermal input is determined from the plant steam output (P = 20 MW thermal) and the thermal efficiency (η = 0,9) to give, under reference conditions, the volume flow rate:

Since the relationship between the parameters is linearly proportional, it is sufficient to combine the relative uncertainties using a simple root mean square approach and a coverage factor of 2,0.

Fuel factor S, relative uncertainty for biomass sample (see above): ~1,0 %

Thermal output P, relative uncertainty from flow meters and thermocouple calibrations: 2,5 %

Thermal efficiency η, relative uncertainty from boiler efficiency analysis: 2,0 %

Combined flue gas volume flow rate q_V, relative expanded uncertainty:

• • 1. Example 3 — Natural gas fired gas turbine plant

The flue gas volume flow rate at a reference condition of 15 % oxygen volume concentration, dry, 273,15 K, 101,3 kPa is required for the calculation of a mass release rate. The fuel factor is defined at stoichiometric conditions of 0 % oxygen.

The oxygen correction is

The fuel factor S = 0,240 m³/MJ for natural gas is given in Table E.1 and the thermal input is determined from the metered fuel input, q_m, F = 10 kg/s, and the measured NSE, e_(N) = 50 MJ/kg, to give, under reference conditions, the volume flow rate:

q_V = 3,521 × Sq_m, Fe_(N) = (3,50 × 0,240 × 10 × 50) m³/s = 420,0 m³/s

Since the relationship between the parameters is linearly proportional, it is sufficient to combine the relative uncertainties using a simple root mean square approach and a coverage factor of 2,0.

Fuel factor S, relative uncertainty for natural gas (Table E.1): 0,35 %

Thermal input P, relative uncertainty from fiscal flow meter and gas chromatograph: 0,8 %

Combined flue gas volume flow rate q_V, relative expanded uncertainty:

1. 
   (informative)
   
   The use of time of flight measurement instruments based on modulated laser light

This document requires a control of the physical dimensions of the duct, where the flow monitor is situated, and such a control can be performed by the use of a non-tactile optical instrument, using modulated laser light, beamed from the instrument to an opposing surface and the re-emission returned to the instrument. The emitted and the re-emitted (returned) signals are compared, and since laser light is modulated with a wavelength ranging from a few to several hundred metres, the distance can be calculated from the phase shift of the two signals.

The method offers a high accuracy, often in the range of a standard deviation below 1 mm, if the following is met:

a) The surface on which the measurement is performed is non-reflective, preferably matt, re-emitting the laser signal in “all” directions. If the laser hits a “reflective” surface, like polished stainless steel, the laser beam is reflected and hits another surface before it is received by the instrument, and thereby the distance measured is greater than that intended.

b) It is best to measure from one flange across the duct to another flange, where a piece of cardboard or wood can be held against the flange to secure a firm and well-defined surface from which to measure.

c) Although many light switches use reflective tape or reflectors to measure against, many distance measurements overload the receiver circuitry and introduce a considerable measurement error; a range of 10 % to 30 % has been experienced. An instrument with a specific signal overload alarm is preferred.

d) Since the measurement depends on the speed of light in air, and gas temperature and air pressure do have an influence, a correction can be necessary if the gas is very warm, the stack is very large and an accurate measurement is required. The influence of temperature is approximately 1 × 10⁻⁶/K, and that of pressure is about 0,3 × 10⁻⁶/hPa, and if the light runs faster than the instrument assumes, it measures too short. A measurement in 200 °C gas and 10 m diameter accordingly measures 200 × 10 000 × 1/1 000 000 = 2 mm too short.

1. 
   (informative)
   
   Example of uncertainty budget established for velocity and volume flow rate measurements by Pitot tube
   1. Process of uncertainty estimation
      1. General

The procedure for calculating measurement uncertainty is based on the law of propagation of uncertainty laid down in ISO 14956^[5] or ISO/IEC Guide 98‑3. The calculation procedure presents different steps (G.1.2 to G.1.5).

• • 1. Determination of model function

The measurand and all the parameters that influence the result of the measurement, called “input quantities”, are clearly defined.

All sources of uncertainty contributing to any of the input quantities or to the measurand directly are identified.

Then the model function, i.e. the relationship between the measurand and the influence quantities, is established, if possible in the form of a mathematical formula.

• • 1. Quantification of uncertainty components

Each uncertainty source is estimated to obtain its contribution to the overall uncertainty by using available performance characteristics of the measurement system, data from the dispersion of repeated measurements, data provided in calibration certificates.

All uncertainty components (e.g. performance characteristics) are converted to standard uncertainties of input and influence quantities.

• • 1. Calculation of the combined uncertainty

Then the combined uncertainty, u_c, is calculated by combining standard uncertainties, by applying the law of propagation of uncertainty.

In general, the uncertainty associated with a measurand is expressed in expanded uncertainty form. The expanded combined uncertainty U_c corresponds to the combined standard uncertainty, obtained by multiplying by a coverage factor, k: U_c = ku_c. The value of the coverage factor k is chosen on the basis of the level of confidence required. In most cases, k is taken to be equal to 2, for a level of confidence of approximately 95 %.

• • 1. Other sources of errors

The mathematical modelling of the measured local velocities and the determinations of mean velocity and volume flow rate that follow, are carried out starting from the basic formulae used to calculate these parameters.

In these formulae, all the parameters have an uncertainty associated with their value which contributes to the total uncertainty of the result of measuring.

However, a thorough analysis of the implementation of measurement could result in counting other sources of uncertainties that do not appear explicitly in the expression used to calculate velocity and the volume flow rate. These sources are, in particular, related to the operational limits of the method, and to the disturbances of the velocity to characterize by the realization of measuring itself:

— nature of the gas stream: the gas stream should be continuous in single phase or should behave as such;

— inhomogeneity of the physicochemical characteristics of gas across the measurement section;

— nature of the flow: the calculation formulae are rigorous only if the flow is stable and presents neither transverse gradient, nor turbulence — however, in practice, both coexist in closed ducts;

— dimension of the Pitot tube: the ratio of the diameter of the antenna of the Pitot tube to the diameter of the duct should be limited in order to minimize the error on the flow resulting from the gradient of velocity and the obstruction caused by the Pitot tube;

— influence of the turbulence: turbulence has an influence on the determination of the velocity and the measurement of the static pressure — the upward bias induced by turbulence on the determination of velocity is a function of the degree of turbulence;

— slow fluctuations of velocity: the error due to an insufficiently long time of measurement to allow a correct integration of the slow fluctuations of velocity decreases when the number and the duration of measurements in a given point increase;

— inclination of the tube of Pitot compared to the direction of the stream: the error increases with the angle of incline;

— pressure loss between total pressure port and static pressure ports: the static pressure ports being located at the downstream of the total pressure port, the dynamic pressure measured with an error equal to the pressure loss by friction in the duct at this distance — this error increases with the distance of the pressures ports and with the roughness of the duct;

— the position of the Pitot tube in the measurement section;

— the number of measurement points: if the curve distribution of velocity shows a distribution not sufficiently homogeneous, the number of measurement points usually prescribed in the standards may not be sufficient.

• 1. Example uncertainty calculation

Estimate of uncertainty velocity and uncertainty volume flow rate of a gas stream in a duct whose characteristics are as follows:

a) duct of a power plant with a diameter of 4,5 m, explored in 20 points by means of an L-type Pitot: uncertainty in the measurement of the diameter of the duct is calculated starting from the maximum permissible error equal to 2 % of the area;

b) temperature of gases on the measurement section: 110 °C = 383 K with an uncertainty of 1 % of the absolute temperature;

c) atmospheric pressure: 100 300 Pa — uncertainty in the atmospheric pressure is calculated starting from the uncertainty associated from its performance characteristics which for this example calculation is 300 Pa and the error due to the reading estimated at 25 Pa;

d) composition of gases:

1) oxygen content measured in the conduit: 6,9 % volume concentration on dry gas ±5 % relative (k = 2);

2) carbon dioxide content measured in the conduit: 12,5 % volume concentration on dry gas ±5 % relative (k = 2);

3) water vapour content: 10,9 % volume concentration on wet gas ±11,6 % relative (k = 2);

e) mean local dynamic pressures, in Pa, at each measurement point:

f) static pressures, in Pa, on each explored measurement line: it is carried out five measurements on each diameter:

The mean pressure on the measurement section is taken equal to the arithmetic mean of the mean static pressures on each diameter

• • 1. Calculation of the physicochemical characteristics of the gas effluent

a) Molar mass gases: M = 28,9 × 10⁻³ kg/mol;

b) density of gases: ρ = 0,909 kg/m³ in actual conditions of temperature and pressure, on wet gas;

c) absolute stack gas pressure: p_c = 100 069 Pa;

d) Pitot tube coefficient: K = 1,01

e) local velocities calculated using Formula (A.2):

e) volume flow rate:

1) q_v,w = 1 806 412 m³/h in actual conditions of temperature and pressure and on wet gas,

2) q_v,0d = 1 133 932 m³/h in standard conditions and on dry gas,

3) q_v,0d,O2ref = 1 065 669 m³/h in standard conditions, on dry gas and to a reference oxygen concentration of 6 %.

• • 1. Calculation of uncertainty associated with the determination of local velocities
       1. General

The uncertainty associated with the determination of local velocities is influenced by the uncertainty of the coefficient of the Pitot tube, the uncertainty associated with the mean local dynamic pressures and the uncertainty associated with the density of the gas effluent and is calculated through Formula (H.1).

(H.1)

• • • 1. Standard uncertainty on the coefficient of the Pitot tube

Characteristics of the Pitot tube: K = 1,01 ± 0,02 (coverage factor k = 2)

• • • 1. Standard uncertainty associated with the mean local dynamic pressures

(H.2)

where

The standard deviation is calculated as follows:

— If the number of measurements is lower or equal to 10:

where

— If the number of measurements is greater than 10:

where is the experimental standard deviation of the series of the dynamic pressure measurements.

The corrections to dynamic pressure measurements are related to:

— the resolution of the sensor used;

— its uncertainty of calibration;

— its drift;

— its linearity;

— hysteresis.

Characteristics of the pressure sensor used (in the example):

— range: 0 Pa to 1 000 Pa;

— resolution: 1 Pa;

— calibration uncertainty: ±2 Pa (with coverage factor k = 2);

— drift: 0,1 % of the range between two calibrations;

— Lack-of-fit: 0,1 % of the range.

NOTE sqrt(3) corresponds to a 95 % confidence interval of a rectangular distribution.

• • • 1. Standard uncertainty associated with the density of the gas effluent
         1. General

(H.3)

where

ρ is the density of the gas effluent under the conditions of temperature and pressure of gas, in kg/m³;

M is the molar mass of wet gas effluent, in kg/mol;

p_c is the absolute pressure in the duct in the measurement plane, in Pa;

T_c is the gas temperature in the duct, in K.

• • • • 1. Standard uncertainty associated with the molar mass of gas

(H.4)

Sensitivity coefficients:

NOTE Derivation after δϕ_O2,w, derivation after δϕ_CO2,w, derivation after δϕ_H2O,w

Standard uncertainty

where , , and are percentage volume concentrations on wet gas.

The contents on wet gas of oxygen and carbon dioxide are given by the following equations:

 % (volume concentration)

 % (volume concentration)

The uncertainty-types associated with the oxygen contents, carbon dioxide and water vapour on wet gas are calculated according to following equations:

(volume concentration) on wet gas

 % (volume concentration)

 % (volume concentration)

u(M) = 7,9 × 10⁻⁵ kg/mol

• • • • 1. Standard uncertainty associated with the temperature T_C

Uncertainty associated with the temperature measurement is dependent:

— with the resolution of the temperature sensor used;

— with the uncertainty of calibration of the measuring equipment: sensor and the measurement instrument;

— with the drift of the measuring equipment;

— with the linearity of the-measuring equipment;

— with the hysteresis-measuring equipment.

The expanded uncertainty associated with the temperature measurement is 2,5 K.

The standard uncertainty u(T_c) is thus equal to:

u(T_c) =  = 1,25 K

• • • • 1. Standard uncertainty associated with the absolute pressure in the duct, p_c

Uncertainty of the absolute pressure is given by:

(H.5)

Uncertainty associated with the atmospheric pressure measurement depends on:

— the resolution of the sensor used;

— the uncertainty of calibration of the sensor;

— the drift of the sensor;

— the linearity of the sensor;

— the hysteresis of the sensor.

In this example, the uncertainty from drift, lack of fit, and hysteresis, as well as the uncertainty due to the reading is known. Standard uncertainty is given by:

 Pa

NOTE sqrt(3) corresponds to a 95 % confidence interval of a rectangular distribution.

If, at each measurement point k, p measurements are carried out, the standard uncertainty associated with the mean static pressure in this point is given by Formula (F.6):

(H.6)

where

The standard deviation of static pressure measurements is calculated in the following way:

— If the number of measurements is lower or equal to 10:

where

— If the number of measurements is greater than 10:

σ_Pstat,k = s_Pstat,k or σ_Pstat,k = d_n × (p_stat,k,max – p_stat,k,min)

where is the experimental standard deviation of the series of the measurements.

The corrections to static pressure measurements are due to:

— the resolution of the sensor used;

— its uncertainty of calibration;

— its drift;

— its linearity;

— hysteresis.

Standard uncertainty associated with the mean static pressure is equal to:

(H.7)

Standard uncertainty associated with absolute pressure is equal to:

(H.8)

In the example, the static pressure is measured with the same pressure sensor as that used to measure the dynamic pressures. Uncertainties of corrections are thus the same.

NOTE Diameter 1 and 2 each 10 points/axis:

Uncertainty associated with the absolute pressure:

u²(p_c) = u²(p_atm) + u² = (173,3)² + (2,37)²

results

u(p_c) = 173,4 Pa

• • • • 1. Standard uncertainty associated with the density

u(ρ) = 4,18 × 10⁻³ kg/m³

• • • 1. Standard uncertainty associated with the local velocities

The standard uncertainty associated with the local velocities is given by:

(H.9)

The results are recapitulated in the table which follows:

• • 1. Calculation of uncertainty associated with the mean velocity

Uncertainty associated with the mean velocity is calculated as follow:

(H.10)

where

(H.11)

(H.12)

Result of the combined standard uncertainty of the mean velocity:

 = 0,32 m/s

Result of expanded uncertainty:

 = ±0,64 m/s (k = 2)

 = ±2,0 % (k = 2)

• • 1. Calculation of uncertainty in reported values
       1. Volume flow rate in the actual conditions of temperature, pressure, water vapour content and oxygen

Standard uncertainty associated with the volume flow rate in the actual conditions of temperature, pressure, water vapour content and oxygen is given by:

(H.13)

— where in the case of a circular duct of diameter D:

— in the case of a rectangular conduit on sides a and b:

Calculation of combined standard uncertainty:

u(q_V,w) = 45 079 m³/h

Calculation of the expanded combined standard uncertainty:

U_c(q_V,w) = ±90 158m³/h (k = 2)

U_c,rel (q_V,w) = ±4,99 % (k = 2)

1. 
   (informative)
   
   Description of validation studies
   1. Overview of validation studies
      1. General

The laboratory validation studies were carried out by Müller BBM, with assistance from E.ON, ABB, Hoentzsch, Sick at wind tunnels at Technische Universität Berlin, Institut für Luft- und Raumfahrt, (TUB). The fan of the wind tunnel was rented by TUB, the wind tunnel was manufactured and delivered by Müller-BBM (MBBM). Further testing was carried out on a heated wind tunnel at TUB. A summary of the results of the validation studies is presented in Reference [23].

The field trials were carried out at locations described in H.1.2 and H.1.3.

• • 1. Municipal waste incinerator in Denmark

The incinerator was operating with three combustion lines feeding a shared stack of 2,8 m internal diameter. The stack gas is typically at 130 °C at 10 % oxygen volume concentration_, dry and contains about 20 % volume concentration water vapour. The bulk velocity was about 20 m/s during the tests. The level of swirl was less than 15°.

Two measurement platforms were available at about 4d_s and 20d_s, where d_s is the stack diameter, from the stack inlet. Four test teams performed 20 point velocity traverses using L-type, S-type, and spherical (3D) Pitot tubes and a vane anemometer. Two tracer methods were also employed — a transit time method using a radioactive tracer and a dilution method using nitrous oxide and methane tracer gases.

The Pitot measurements at the lower level indicated very non-uniform velocity profiles when compared with the very uniform profiles obtained at the upper level which approached the fully developed condition. Despite this, representative bulk velocity averages were obtained in both cases.

There were four separate incineration lines firing mixed waste (mostly municipal), each fitted with NO_x abatement (SNCR), particulate abatement (bag filters) and individual continuous emission monitoring (CEM) systems. Only three lines were operational during the trials.

• • 1. Coal-fired power plant in Germany

The validation study was carried out at a 700 MW electric coal fired power plant in Germany. The flow from the boiler is split between two abatement trains, each with Electrostatic precipitators, NO_x removal (SCR) and wet flue gas desulfurization (FGD). These feed a shared stack of 7 m internal diameter. The stack gas is typically at 120 °C at 6 % oxygen volume concentration dry and contains about 12 % volume concentration water vapour. The bulk velocity ranged from 24 m/s to 31 m/s during the testing and the level of swirl in the flow was less than 15°.

One measurement platform was available at about 6,5d_s from the stack inlet at the 52,3 m level in the stack. Four test teams performed 20 point velocity traverses using paired trains of L-type, S-type, spherical (3D) and 2D Pitot tubes. The L-type Pitot tubes were strapped together and inserted through a single port. It was not possible to use tracer methods at this plant due to the difficulty in obtaining permission to use a radioactive tracer and the poor mixing quality obtained for dilution flow methods.

The Pitot measurements indicated non-uniform velocity profiles. Despite this, representative bulk velocity averages were obtained.

The L-, S-type and 3D Pitot tubes gave comparable results for average velocity with the 3D Pitot being about 1 % lower than the L-type and showing a greater difference between the two trains. The L-type showed the least variation between trains as might be expected since they were nominally measuring at the same point. These results agree with the plant flow rate calculated from the electricity generation and the plant efficiency.

• 1. Results of laboratory validation

The performance of the manual methods assessed during the laboratory test programme is summarized in Table H.1 which presents the linear regression of the methods, and Table H.2 which summarizes the uncertainty assessment of the methods from the laboratory study. Two results are provided for the 3D-Pitot tubes (ES and AP), which relate to two different calibration factors provided using two different suppliers and approaches. This is explained in more detail in the laboratory test report.

Table H.3 presents the lack of fit data which has been determined from the laboratory regression studies, in accordance with the procedure given in EN 15267-4.^[13] This quantifies lack of fit as the largest (absolute) deviation from the determined regression line of any single measurement data point. For illustrative purposes, the lack of fit has also been compared against the criterion for lack of fit given in EN 15267-4,^[13] which is 3,0 % of the testing range.

Table H.1 — Linear regression data for manual methods from laboratory test data

Table H.2 — Uncertainty analysis according to ISO 20988 for manual methods in laboratory assessment

Table H.3 — Lack of fit determined from laboratory test data for manual methods

• 1. Results of field validation studies
     1. Repeatability and uncertainty of manual methods in the first field validation study

In order to provide an assessment of the repeatability of the manual methods in the first field validations study, the paired sets of data for the 3D and S-type Pitot tubes were assessed in accordance with the procedure in EN 14793^[11] which provides a method of determining the pooled standard deviation of paired results (paired data were not available for the L-type Pitot and vane anemometers). This was done by determining the standard deviation of each pair of measurements and then combining those as variances (i.e. mean sum of squares). This assessment includes the effects of any systematic differences between the methods.

In order to assess the overall standard deviation of the manual methods, the standard deviation of each set of coincident 3D, S-type, vane and L-type results was determined and the pooled standard deviation for all these sets of measurements was calculated, again in accordance with the approach given in EN 14793.^[11] In addition the pooled standard deviation for the paired 3D and paired S-type Pitot tubes was determined. The results of these tests are given in Table H.4.

Table H.4 — Pooled standard deviations of manual methods

The standard deviations include both random and systematic variations. The measurements were also made at different sample locations (60 m and 20 m elevations) and so this analysis also includes any variability caused by the different sampling configurations. Care should therefore be taken in interpreting these results.

The pooled standard deviation for measurements made using the L-type Pitot and the manual vane anemometer was also calculated (Table H.5). This analysis used all 18 paired measurement periods made using these two methods.

Table H.5 — Pooled standard deviation for L-type and vane anemometer

The uncertainty of the manual methods was assessed using the techniques defined in ISO 20988. As it is proposed that any of the manual approaches may be used to calibrate the AMS techniques, in this analysis the set of manual methods are considered as implementations of a single method. In this way, the uncertainty of the ensemble of the methods is determined. The results may therefore be interpreted as the uncertainty for any of the manual methods. The set of parallel measurements may be considered as an experimental design consisting of parallel measurements with identical measuring systems, defined in ISO 20988 as experimental design A8. “Identical” in this context is taken to mean complying with the requirements of this document. This assumes the uncertainties of the different implementations of the method are similar (the assumption is that all the results from the techniques represent samples of an overall population of results representing "the method" as a whole, consistent with a normal probability distribution).

In the first assessment, the results from the six manual methods, the two 3D Pitot tubes, two S-type Pitot tubes, L-type Pitot and the vane anemometer, were assessed. This addressed the methods which are all considered as comparable implementations of the manual method which provide point velocity measurements.

The ISO 20988 analysis gave the following results. The standard uncertainty of the result measurement y from the application of a manual flow measurement techniques in the range 17,8 m/s to 21,2 m/s, is u(y) = 0,49 m/s. The expanded uncertainty of a result of measurement y using a manual flow measurement method in the range 17,8 m/s to 21,2 m/s for a confidence level of 95 % is U_0,95(y) = 0,98 m/s.

The 95 % confidence interval [y_R − U_0,95(y), y_R + U_0,95(y)] is expected to encompass P = 95 % of the measured points. It was found to encompass P = 97,5 % of the evaluated 62 measurement results y(k,j). Therefore, the expanded uncertainty U_0,95(y) = 0,98 m/s is considered to be a reasonable measure of the uncertainty.

The uncertainties determined are therefore applicable to the measurement of average flow for an emissions duct in m/s formed by taking a grid of samples of point flow measurements.

A similar uncertainty assessment was carried out to include all of the periodic measurement technique results reported in Table G.6, i.e. including the results of the tracer techniques. These assessments were carried out using the ISO 20988 assessment approach as described in the preceding. Table H.6 presents a summary of the set of ISO 20988 uncertainty assessments.

Table H.6 — Uncertainty evaluation of the manual flow methods

• • 1. Repeatability and uncertainty of manual methods in the second field validation study

During the second field validation study there were a number of different periods of measurement, and therefore it was not possible to form ensemble performance statistics for all the different methods deployed. However, paired sets of measurements were carried out for each Pitot type which was used. In addition, the 2D Pitot tube was used during the second validation study, which had not been available during the first study.

The paired data from each Pitot method were analysed using the same methodology as described for the first field validation study, to provide pooled standard deviations for the methods, reported in Table H.7. As can be seen, L-type Pitot tubes gave very good repeatability performance. Because these Pitot tubes were mechanically linked together, this uncertainty analysis is not affected by differences caused by any inhomogeneity in the flow profile or other parameters (e.g. gas density).

The variability, Var_f, for the paired methods was determined in accordance with the procedures given for R_f in EN 15267-4.^[13] These data are reported in Table H.8. Note Var_f has been determined from paired data using a calculation based on that for reproducibility as defined in EN 15267-4.^[13] However, as the validation study is not a performance test and did not use paired instruments, the calculation has been used to give an indication of the variability of the methods and is not a strict application of reproducibility as specified in EN 15267-4.^[13] The standard deviation of the differences obtained from the paired measurements is denoted s_D.

Before determining these results the paired data sets were assessed for outliers, by performing the Grubbs test. Two pairs of S-type Pitot results were identified as outliers. These data were excluded from the statistical analysis. This is acceptable so long as a similar exclusion of outliers is also carried out when the methods are used to determine flow for mass emissions calculations or calibrate AMS.

The uncertainties of the manual methods were also determined from an assessment of the paired data undertaken in accordance with ISO 20988and are reported in Table H.9. The experimental design can be considered to be paired measurements of two identical measuring systems as defined in ISO 20988 as experimental design A6. The analysis provides information on the uncertainty due to the bias between the two measurements. The uncertainty procedure then makes use of the relative size of this term u_B compared to the standard uncertainty u, to determine a the method to use to estimate the degrees of freedom, and hence the coverage factor to be used to determine the expanded uncertainty U_0,95. For the L-type Pitot, this assessment passed the criterion , with the number of degrees of freedom equal to the number of paired tests. The other techniques do not meet this criterion, and for these, ISO 20988:2007, 7.4 applies.

Table H.7 — Uncertainty in the manual paired manual methods used in second validation study, determined from pooled standard deviation

Table H.8 — Variability determined for paired manual methods for the second field validation study

In interpreting these uncertainty values, it should be recognized that the uncertainties determined for the S-type, 3D, and 2D Pitot tubes include the effect of determining the average flow across the duct over the same period, but with the grid of flow measurement being determined in a different order between the pair of methods (i.e. S₁ will have sampled the grid over the same period as S₂, but they will have sampled different parts of the gird at the same time, whereas L₁ and L₂ both sampled the same points across the grid at the same time). The difference between the L-type uncertainties and the other techniques implies there is an effect of the sampling process, and therefore the uncertainties for the S-type, 3D, and 2D Pitot tubes may be considered more representative of the uncertainty of a single calibration point made when using these methods, and the uncertainty of the L-type may be considered representative of the uncertainty of individual point flow measurements, made using these Pitot tubes.

Table H.9 — Uncertainty analysis of paired manual method results from second field validation study

1. 
   (informative)
   
   Check of validity of the calibration of a Pitot tube
   1. Flow generation facility

Clause 8 of this document requires a Pitot tube to be calibrated before the first use in the field and after any repair required. Periodically the validity of the calibration can be checked by comparison with an annually calibrated reference Pitot tube.

This can be carried out in a facility that can simulate stack flow conditions provided that the facility:

— has a system to regulate the flow (e.g. inverter);

— has a measurement section and measurement plane that follow the requirements of ISO 15259 to the best degree possible;

— is able to generate a stable velocity generated and representative of velocities typically encountered in stacks (e.g. 5-30 m/s).

• 1. Calibration check procedure

The calibration check can be made by comparing the output of both the reference Pitot tube and Pitot tube under test at a number of measurement points under the same velocity conditions following the below procedure:

a) Checks for damage (Pitot tubes) and functional pre-checks (differential pressure readout devices) are carried out following the requirements of 9.3.1. A pre-test leak check is carried out on the two systems following the requirements in 9.3.2.

b) The reference Pitot tube is inserted into the measurement section and located at the first measurement point. The location and minimum number of measurement points follow the requirements of ISO 15259. Measurements points close to the facility wall are avoided. The insertion port is sealed so there is no ingress of air that could disturb the flow at the measurement plane. The differential pressure ∆p_r is then recorded following the requirements of 9.5.

c) The reference Pitot tube is then removed and the Pitot tube under test inserted at the same measurement point as the reference Pitot tube with step b) repeated in order to obtain the differential pressure ∆p_x.

d) The process is repeated until three pairs of ∆p_r and ∆p_x have been obtained.

e) The calibration factor K_x of the Pitot tube for each pair is calculated according to Formula (I.1):

(I.1)

where K_r is the calibration factor of the reference Pitot tube.

f) If any of the values of K_x obtained differs by more than 0,02 from the K_x average value obtained over the three repeated measurements the reason needs to be investigated and the calibration check either repeated or the Pitot tube under test replaced.

If the Pitot tube under test is an S-Type, an additional check is carried out during one of the paired measurements:

The Pitot tube under test is rotated 180° in order to obtain the differential pressure of the second pressure tap.

If the calibration factors for each pressure tap of the Pitot tube under test calculated through Formula (I.1) differ by more than 0,01 the reason needs to be investigated and the calibration check either repeated or the Pitot tube under test replaced.

1. 
   (informative)
   
   Differential pressure measurement
   1. General

There are many ways in measuring differential pressure (DP) for determination of flow using a Pitot tubes which include:

— liquid manometers;

— digital manometers;

— differential pressure gauges.

The rangeability of the flow system is an important point.

The basic flow formula is given by Formula (A.2). Formula (J.1) shows the proportionality between the flow and the differential pressure:

(J.1)

where

q_V is the flow;

Δp is the differential pressure.

From Formula (K.1), at 30 %, flow would give a 9 % differential pressure. This would mean a flow meter has a 3:1 rangeability.

Typically, liquid filled manometers have an accuracy of ±1 % of reading whereas hand-held digital manometer accuracy is based on the full scale reading typically ±0,5 % F.S.

It is very important when specifying and purchasing a digital hand-held manometer that the correct range is chosen. Typical ranges when using Pitot tubes are 0 kPa to 2,5 kPa.

Based on this it can be seen that when monitoring at the lower flow point, i.e. 30 %, then the accuracy of reading on a liquid manometer would be 0,225 Pa and 1,25 Pa on the digital manometer which equate to an inaccuracy of 5,5 % of reading.

However, there are very high resolution precision manometers available that can read down to 0,001 Pa with very high accuracy.

• 1. Liquid manometers

Manometers measure a pressure difference by balancing the weight of a fluid column between the two pressures of interest. Large pressure differences are measured with heavy fluids, such as mercury (e.g. 760 mmHg = 1 atmosphere). Small pressure differences, such as those experienced with Pitot tubes are measured by lighter fluids such as water. See Figure K.1.

Key

Figure J.1 — Principle of liquid manometer

• 1. Digital manometers and other electronic devices
     1. General

Digital manometers are available from many companies using a variety of pressure sensor technologies.

• • 1. Types of pressure sensors
       1. Piezoresistive strain gauge

This device uses the piezoresistive effect of bonded or formed strain gauges to detect strain due to applied pressure. The piezoresistive effect describes the changing resistivity of a semiconductor due to applied mechanical stress. The piezoresistive effect differs from the piezoelectric effect. In contrast to the piezoelectric effect, the piezoresistive effect only causes a change in electrical resistance; it does not produce an electric potential.

• • • 1. Capacitive pressure sensor

This device uses a diaphragm and pressure cavity to create a variable capacitor to detect strain due to applied pressure.

• • • 1. Magnetic pressure sensor

This device measures the displacement of a diaphragm by means of changes in inductance (reluctance), linear variable differential transformer, Hall effect, or by eddy current principal.

• • • 1. Piezoelectric pressure sensor

This device uses the piezoelectric effect in certain materials such as quartz to measure the strain upon the sensing mechanism due to pressure.

• • • 1. Optical pressure sensor

This device uses the physical change of an optical fibre to detect strain due applied pressure.

• • • 1. Potentiometric pressure sensor