Project AC0115 Final Report -...

Project AC0115 Final Report

Appendix B: Chamber Calibrations

April 2014

Introduction This appendix provides details of the work carried out to calibrate the various chambers operated by the project partners. The first part of the appendix is an interim report that was delivered to Defra in April 2012, detailing the procedures followed. The second part of the appendix is a draft manuscript of the work carried out, including the results of measurements made, which is to be submitted for publication in an international refereed journal.

AC0115 Final Report Appendix B - Page 1

Interim Report on Chamber Validation Experiments Report prepared by Tom Gardiner and Marc Coleman, Environmental Measurements Group, NPL April 2012

Executive Summary • NPL are leading work within Project AC0115 to ensure comparability between the methane

emission chamber measurements being made across the project by validating the chamber measurements being made at each experimental site.

• Initial visits were carried out to each of the five groups carrying out chamber measurements to review current calibration activities and define the requirements for the validation work.

• A set of validation tools have been developed specifically for this work, including a suite of gravimetric reference standards for direct calibration of the methane sensors, and a calibrated dynamic flow mixing system to generate controlled methane emission rates.

• The successful use of these tools where demonstrated at the initial validation experiment at IBERS. This covered the main validation activity, in addition to assessment of the range of experimental characteristics such as time response and linearity, and the results are discussed in the report.

• The second validation experiment, at the University of Nottingham, has also been completed and the data are being analysed.

• The three remaining validation experiments are scheduled for completion by June 2012, and two peer-reviewed publications are planned based on the results.

Background The work on method standardisation (Task 2.2) is being led by NPL and aims to ensure that common calibration and validation approaches are being used by all the measurement groups to give consistency in the measurements and related uncertainty budgets. This report summarises the current status (as of March 2012) of work undertaken to achieve comparability of the methane flux chamber measurements being carried out by five different groups within the AC0115 project : IBERS, SAC, the University of Nottingham, the University of Reading, and AFBI. This comparability is being achieved through a set of validation experiments at each site using calibration tools developed specifically for this project by NPL.

The report covers the initial site visits, describes the design of the validation experiment, and provides an example of the outcomes from the measurements using the preliminary results from the first validation experiment carried out at IBERS. The report finishes with a discussion of the next phases of work in this area.


Initial Site Visits Visits were carried out by NPL to all sites to look at the metrology issues associated with the methane measurements, and to scope out the requirements for the validation experiments. These visits took place between April and July 2011. The general conclusion of the visits was that good QA/QC checks were already in place at each site, but they confirmed the requirement for some dedicated validation experiments to formally establish comparability between the sites.

Table 1 summarises the general properties of the chambers at each of the sites (information correct at the time of the visit). This information, together with that gathered during discussions with each measurement team, was used to define the requirements for calibration standards and the measurement procedures that would be used for the validation experiments. The methodology developed for the validation experiments is discussed in the following section.

The primary goal of these experiments is to validate the calculation of methane emission rate from livestock in each of the chambers. Although there are slight differences in the details of this calculation from site to site (depending on the experimental set-up), a typical example is given in the following equation :

𝐸𝐸𝐸𝐸 = 31.54 𝐴𝐴 �(𝑀𝑀𝐸𝐸𝑐𝑐 −𝑀𝑀𝐸𝐸𝑎𝑎)𝑅𝑅𝑀𝑀𝑀𝑀𝐶𝐶𝐶𝐶4

𝑅𝑅𝑅𝑅𝑃𝑃�

�

Where :

EF = methane emission factor (kg.y-1)

MEc = measured CH4 concentration in chamber (ppm)

MEa = measured CH4 concentration in ambient air (ppm)

RMMCH4 = relative molecular mass of CH4 (g.mol-1)

R = molar gas constant (J.K-1.mol-1)

A = flow through ducting (m3.s-1)

T = temperature of gas at point of flow measurement (K)

P = pressure of gas at point of flow measurement (kPa)

31.54 = Constant to convert from mg.s-1 to kg.y-1

The validation experiments will establish common traceability for the emission factor determinations carried out by each group and provide the relevant calibration factors together with uncertainties. The experiments will also look at a number of key experimental parameters, such as response time, to check the appropriateness of the experimental procedures being used.


Table 1. Summary of chamber characteristics for each site (N.B. A number of parameters are estimated)

Design of Validation Experiment Each validation experiment comprises two main phases. An initial direct calibration of the methane sensor, to derive a calibration function for the methane concentration measurements, followed by a series of experiments looking at emission rate determination using a controlled emission source. The details of the experiments depend upon the exact set up at each site – an example experimental configuration is shown in the example results section. However, the reference gases used for direct concentration calibration and emission rate determination are common for all the validation experiments and are described below.

Methane Sensor Calibration NPL have prepared a suite of seven methane standards with gravimetric traceability back to international standards (and the S.I.) specifically for this purpose to cover the range of concentrations observed by the different groups. These standards have an absolute concentration uncertainty of 0.45%-0.50% (k=2, 95% confidence) – this compares to a typical uncertainty of 2%-4% for the secondary standards available from the specialist gas suppliers that are typically used for regular span checks. In addition, the use of multiple concentrations, rather than a single span measurement, enables sensor linearity to be assessed, and the calibration function to be determined across the required measurement range.

An appropriate set of these standards, together with a high purity nitrogen zero gas (see below), is selected to match the requirements for each site. The standards are then decanted from their parent cylinders into lecture bottles for transportation to the site.

Site IBERS SAC Nottingham Reading AFBIChamberNumber 4 6 2 2 2Volume 4.6 m3 80 m3 80 m3 20 m3 27 m3Comment Sheep only New recovery roomsChamber FlowType Flow-through Recirculating Flow-through Recirculating RecirculatingRecirculating flow n/a 490 L/sec n/a 46 L/sec ~50 L/secExtracted flow 30 L/sec 50 L/sec 200 L/sec 33 L/sec 22 L/secAmbient ConditionsControl Uncontrolled Controlled Aircon inlet Controlled Controlledset T n/a 15 deg C 22 deg C 16 deg C 13 deg Cset RH n/a 70% n/a 60% 60%logging T,P and RH being installed T and RH T,P and RH T and RHAnalyserType ADC MGA3000 ADC MGA3000 Signal 9000 Custom ADC MGA3000Channels 8 8 4 4 3Calibration Span daily Zero and span daily Zero and span daily Zero and span daily Zero and span each exp.

3 hr autozero 4 hr zero and span checkSpan gas 50 ppm BOC standard 500 ppm BOC standard 808.5 ppm BOC 193 ppm Cyroservices 509 ppm BOCSample flow 1.5 L/min 35-40 L/min 1 L/min 1-2 L/min 35-40 L/minTypical conc. 6-12 ppm (sheep) 20-40 ppm (2 sheep) 100 ppm (cow) 70-250 ppm (cow) 600 ppm (cow)

100-200 ppm (1 cow)Sample timing 3 mins per channel 45 s bypass / 45 s sample 2 mins per channel 95 s purge / 20 s sample 75 s per channel

24 min cycle 6 min cycle 8 min cycle 8 min cycle 3 min 45 sec cycle

Same base design


Validation of Emission Rate Measurements Methane emission rate validation is carried out using a customised dynamic mixing source that can produce time-varying emission rates with calibrated mass-flow control. The dynamic mixing process combined different flows of Ultra High Purity (UHP) methane, with a minimum purity of 99.9995%, and BIP grade nitrogen, with hydrocarbon contamination of less than 50 ppb (expressed in methane equivalents).

The mass flow control system was gravimetrically calibrated, using the appropriate gases, to determine the flows rates with an absolute uncertainty of within 0.5% (k=2, 95% confidence).

Example Results from First Chamber Validation

Testing Methodology The first chamber validation experiment took place at the IBERS facility in November 2011. The initial phase of the experiment was to introduce the series of gravimetric methane standards directly into the methane sensors, in this case an ADC MGA-3000 instrument. The configuration for this experiment is shown in Figure 1.

Figure 1. Schematic of experimental design for direct analyser calibration

The calibration gas is run on a by-pass line through a flow rotameter to ensure that there is excess flow available for the sensor to sample without over-pressuring the sample line.

The results from this initial test were used to test sensor linearity and derive the sensor calibration factor. Tests were also done to determine the sensor response time. This involved monitoring the

Lecture bottle

MGA-3000

Ambient air

3-way valve

Vent

Rotameter


response of the instrument as it switched from ambient / background methane to the calibration sample. The ‘T90’ response time was then calculated as the time to rise from the background level to 90% of the plateau value for the calibration gas. N.B. The methane sensor makes a measurement every second, however the recorded data is the reading at the end of each 3 minute sample period. Therefore, in order to obtain the rapid data required for the response time determination, readings were taken manually every 10 seconds.

The second phase of the experiment was to introduce a known level of continuous methane emission directly into the chamber extract. This direct emission measurement was to assess the performance of the extract and flow system without any influence from the chamber itself.

The third, and longest, phase was a series of experiments introducing known levels of continuous methane emission into the chambers themselves. The typical experimental configuration for the in-chamber measurements is shown schematically in Figure 2. A photo of the controlled methane source installed in a chamber is shown in Figure 3.These experiments assessed a number of aspects of chamber performance including validation of the overall chamber capture efficiency, the response time for emissions within the chamber, the variability of the measurement under stable emission conditions, the linearity of the chamber measurements over a range of emission levels. All of these characteristics were assessed for one of the four chambers. The final experiment looked at the comparative performance of the other three chambers at the IBERS facility.

The preliminary results of each of these experiments are discussed in the following sections.

Figure 2. Schematic of experimental design for chamber validation


Figure 3. Controlled methane source installed in chamber

Sensor Calibration The sensor (MGA-3000) was directly calibrated with gravimetric standards, and showed a linear response with small gradient difference (1.8%) and zero offset (0.2 ppm), as shown in Figure 4.

Figure 4. Sensor response to gravimetric calibration standards (concentration in ppmv).

Sensor Response Time The sensor reading was recorded manually every 10 seconds for a directly injected sample of calibration gas. The reading reaches 90% of the plateau value (T90) after 38.1 seconds, as shown in Figure 5. The mean plateau reading was 50.95 ppm with a stdev of 0.17 ppm, i.e. a stability of 0.34%. N.B. Instrumental standards typically specify taking reading after 3 x T90 i.e. 114.3 seconds in this case, which is consistent with the 3 minute sample time used for the logged readings.


Figure 5. Sensor time response to gravimetric standard from ambient background

Direct Emission Measurements For this experiment the flux source was positioned just inside the main sample pipe from the chamber, and 3 minute measurements were taken over a period of one hour. Manual 10 sec readings were recorded for four of the 3 minute samples. The manual readings showed a T90 response time of 60.4 seconds, and a plateau stability of ~0.5%.

The mean results of the hour-long experiment gave a measured methane emission rate of 2.715 L/hr (+/- 0.39%), compared to a controlled source emission rate of 2.278 L/hr (+/- 0.5%). This gives a direct emission calibration factor of 0.839 (+/-0.63%). This effectively acts as an adjustment factor to the flow reading in the flux equation.

Chamber Measurements

Measurements in single chamber The flux source was positioned near centre of one of the four chambers (identified as Chamber 2), and at approximate sheep mouth height. Readings were taken as in the previous direct emission experiment. The results of the manual readings taken during the first chamber measurement are shown in Figure 6, and show a similar T90 response time of 72.7 seconds, but with much more variability on the plateau reading (stability of ~11%).


Figure 6. Manual 10 second readings taken during four 3 minute measurements cycles

The measured emission rate, taking into account the previously determined adjustment factors, is 2.123 L/hr (with s.e. of 6.17%) compared to a controlled source emission rate of 2.278 L/hr (+/- 0.5%). Therefore, the fraction of source flux observed (capture efficiency) was 0.932 (+/-0.058).

In this first experiment the feedbox and water bucket not in place within the chamber. These are present for all animal measurements, and due to the chamber design could influence the capture efficiency. A series of further experiments were carried with and without the feedbox and water bucket in place, including a longer duration overnight test. The fractions of source flux determined for this series of tests is shown in Figure 7

Figure 7. Results from individual experiments in Chamber 2 with and without feedbox (and water bucket)


The results showed a lower fraction (i.e. bigger loss) observed without the feedbox and water bucket in place. The overnight (O/N) measurement gives the best individual estimate of the ‘with feedbox’ capture efficiency, of 0.959 (+/- 0.013).

Chamber Stability

Figure 8. Measurement of chamber stability over 20 minutes

The stability of the chamber readings was determined from a 20 minutes period of stable emissions and no sample switching, as shown in Figure 8. The results show 16.9% variability (after initial rise), and an average periodicity of ~122 seconds (N.B. The flow variability over same period is 2.2%).

If we take this variability as a measurement of the sample population variability and a 24 minute sampling rate, then the uncertainty on the mean for longer term averages can be estimated assuming normally distributed noise behaviour. Table 2 shows how the standard deviation of the mean - as would be used for a livestock experiment - will vary as a function of total measurement period.

Table 2. Estimated variation in measurement uncertainty with averaging time

Measurement period (hrs)

Number of samples

Stdev Mean (%)

8 20 3.9

24 60 2.2

48 120 1.5

72 180 1.3


Chamber Linearity The linearity of the chamber response was assessed by varying the source amount down from the initial value of ~2.3 L/hr, as shown in Figure 9. The results at lower fluxes were consistent with results at ~2.3 L/hr , and showed a linear response with a capture efficiency factor of ~0.96.

Figure 9. Chamber response to varying levels of delivered methane (red line shows the 1:1 response)

Comparison of Chambers Hour-long capture efficiency experiments were also carried out in the other three chambers. The results are consistent with the first chamber, i.e. within one standard error, as shown in Figure 10 which shows the capture efficiency (fraction of source observed) results for all four chambers. An uncertainty weighted average of all ‘with feedbox’ results gives an overall capture efficiency factor of 0.966, and this is the value used for the final validation parameters – see following section.

Figure 10. Summary of results for all four IBERS chambers


Conclusions and Next Steps This report has described the development of a set of validation tools developed specifically for the calibration of methane emission from livestock chambers. The appropriateness of these tools has been demonstrated and the first chamber validation experiment has been successfully completed. The results have enabled various aspects of chamber performance to be assessed and quantified. The experimental methodology was refined during the course of the measurements, giving a template for the experiments at the other sites.

The final output from the validation experiment is a revised version of the flux equation, together with a combined measurement uncertainty. The provisional result for the IBERS experiments is given below :

𝐸𝐸𝐸𝐸𝑣𝑣𝑎𝑎𝑣𝑣 = 31.54 𝛼𝛼𝐴𝐴 �(𝛽𝛽𝑀𝑀𝐸𝐸𝑐𝑐 − 𝛽𝛽𝑀𝑀𝐸𝐸𝑎𝑎)𝑅𝑅𝑀𝑀𝑀𝑀𝐶𝐶𝐶𝐶4

𝛾𝛾�𝑅𝑅𝑅𝑅 𝑃𝑃� ��

Where :

EFval = validated methane emission factor (kg.y-1)

MEc = measured CH4 concentration in chamber (ppm)

MEa = measured CH4 concentration in ambient air (ppm)

RMMCH4 = relative molecular mass of CH4 (g.mol-1)

R = molar gas constant (J.K-1.mol-1)

A = flow through ducting (m3.s-1)

T = temperature of gas at point of flow measurement (K)

P = pressure of gas at point of flow measurement (kPa)

31.54 = Constant to convert from mg.s-1 to kg.y-1

α = direct emission correction (0.839)

β = calibration function for sensor (1.0179x+0.2096)

γ = chamber capture efficiency (0.966)

The next phase of the work is to finalise the results and associated uncertainties for the IBERS validation experiment and carry out the experiments at the other chamber sites. The second of the experiments, at the University of Nottingham, has been completed and the results are being analysed. The remaining three visits are scheduled to take place over the next few months during at appropriate gaps in the livestock experiments at each site. The complete set of validation experiments should be completed by the end of June 2012. Two peer-reviewed papers are planned on the results of this work. The first covering the development and demonstration of the validation capability, and the second summarising the comparative results of the measurements at all of the sites.


Determination of the Absolute Accuracy of UK Chamber

Facilities used in Measuring Methane Emissions from Livestock

T.D. Gardiner1, M.D. Coleman1, F. Innocenti1, J. Tompkins1, A. Connor1, P.C. Garnsworthy2,

J. M. Moorby3, C.K. Reynolds4, A. Waterhouse5, D. Wills6

1National Physical Laboratory, Hampton Road, Teddington, Middlesex, TW11 0LW, UK

2The University of Nottingham, School of Biosciences, Sutton Bonington Campus,

Loughborough LE12 5RD, UK.

3Institute of Biological, Environmental and Rural Science, Aberystwyth University,

Gogerddan, Aberystwyth, SY23 4AD, UK

4Centre for Dairy Research, School of Agriculture, Policy, and Development, University of

Reading, PO Box 237, Earley Gate, Reading, RG6 6AR, UK

5Future Farming Systems, SRUC, West Mains Road, Edinburgh, EH9 3JG, UK

6Agri-Food and Biosciences Institute , AFBI Hillsborough, Large park, Hillsborough, Co.

Down, BT26 6DR, UK

*Corresponding author. E-mail: [email protected], Tel.: +44 (0)208 943 7143

Abstract

Respiration chambers are one of the primary sources of data on methane emissions from

livestock. This paper describes the results from a coordinated set of chamber validation

experiments which establishes the absolute accuracy of the methane emission rates measured

by the chambers, and for the first time provides traceability to international standards,

assesses the impact of both sensor and chamber response times on measurement uncertainty

and establishes direct comparability between measurements made across different facilities

with a wide range of chamber designs. As a result of the validation exercise the estimated


mailto:[email protected]

absolute uncertainty associated with the overall capability across all facilities reduced from

25.7% (k=2, 95% confidence) before the validation to 2.1% (k=2, 95% confidence)

afterwards.

keywords: respiration chambers, methane emissions, livestock emissions, calibration,

traceability

Introduction

Methane is a greenhouse gas (GHG) with a global warming potential 33 times that of carbon

dioxide (IPCC, 2013). Agriculture is a significant contributor to global methane emissions as

evidenced by the 2011 European Union (EU) inventory detailing that 50% of all methane

emissions were attributable to the agricultural sector (European Environment Agency, 2013).

Hence, it is clear that the agriculture sector has an important role to play if international

commitments to reduce emissions (e.g. European Climate Change Programme target of 20%

reduction by 2020 compared to 1990 levels) are to be met. Currently in the UK livestock

emissions (contributing ~85% of methane emissions from agriculture) are calculated using

the Tier 1 approach (IPCC, 2006) under the United Nations Framework Convention on

Climate Change (UNFCC). The Tier 1 approach is based on using emission factors (EFs) for

different livestock categories and associated manures, i.e. no account is made with respect to

farm activity or mitigation effort, e.g. breed, age, diet or seasonality (IPCC, 2006).

Consequently, the UK Government’s Department for the Environment and Rural Affairs

(DEFRA) have commissioned a programme of research to address such issues to facilitate

movement to a Tier 2 or 3 approach under UNFCC - the Agricultural Greenhouse Gas

Inventory Research Platform (http://www.ghgplatform.org.uk/). A key part of this research is

work to underpin national measurement infrastructure to ensure that facilities used for


http://www.ghgplatform.org.uk/

measuring livestock emissions are producing comparable data that is traceable to the

international system of units and has quantified uncertainties, and it is this work that we

report here.

A generally accepted method for determining emissions is the respiration chamber where the

animal is placed in the chamber for several days with a controlled throughput of ambient air.

Measuring the concentration difference between the outlet and inlet combined with the flow

rate gives the total emitted methane flux (Grainger et al., 2007; Yan et al., 2010). Historically

calorimetry chambers were used to estimate heat production for measurements of energy

metabolism (e.g. Armsby, 1908; Cammel et al., 1981; McClean and Tobin, 1987; Derno et

at., 2009), which required precise and accurate measurements of oxygen consumption and

carbon dioxide and methane production by animals housed in the chamber. However, due to

the reasons outlined above the focus has now shifted towards using chambers to determine

the impact of animal husbandry practises on methane emissions, often with simpler designs

(e.g. Murray et al, 1999, Klein and Wright, 2006; Pinares and Waghorn, 2012).

There have also been several reports comparing chambers to other measurement methods.

For example a number of groups have compared the sulphur hexafluoride (SF6) technique

(Johnson et al., 1994), which involves placing a permeation tube in the rumen to release SF6

tracer gas at a known rate, to respirator chamber measurements of lactating dairy cows.

McCourt et al. (2008) found the SF6 technique to measure 75% that of chambers, whilst

Grainger et al. (2007) found a relationship of 102%. Muñoz et al. (2012) initially found a

close correlation (similar to Grainger) between SF6 and chambers, although as testing

progressed they found the former began to measure significantly higher. In the later study,

removal of the tubes from the rumen revealed that the release rates had dropped on average to


66% of the pre-test rate, i.e. it was concluded that the increased measurements towards the

end of the testing were the product of the decrease in SF6 release rate.

There have also been reports of comparisons between respiration chambers and a method

involving measuring eructation during milking via sampling air directly from automatic

milking machine feed bins (Garnsworthy, 2012). However, rather than being a direct

comparison the purpose of their work was to firstly establish if a correlation existed between

eructation frequency and associated methane concentration with daily methane emissions.

Having found supporting evidence the authors were able to derive an expression to relate

measurements made during milking to daily emissions.

Whilst work such as that described above is of great value to the community as it shows the

relative comparability of different measurement methods it does not address the absolute

uncertainty. Often what is measured is the precision of a method and not the combined

uncertainty, which includes both the precision and any sources of bias. For example, two

identical methane sensors will have the same precision (noise) but if only one is calibrated

against a traceable reference material then they could provide very different readings of the

same chamber despite possessing the same precision. Hence, to truly understand the accuracy

of any method and to establish the comparability between different measurement systems

there must be comparison to an internationally accepted reference point. Historically, the

accuracy of chamber measurements has been based on calibration of flow meters and

analyser performance (McLean and Tobin, 1987) and measurement of emissions obtained

during a weighed release of the target gas into the chambers. Mclean and Tobin (1987) give

an extensive review of recommended procedures at that time and Cammel et al (1981)

summarise results for a number of published respiration chambers. More recently Hellwing et


al. (2012) report on the calibration of a simple respiration chamber for cattle. However, the

work reported here is, to our knowledge, the first to provide traceability to international

standards, assess the impact of both sensor and chamber response times on measurement

uncertainty, and establish direct comparability between measurements made across different

facilities with a wide range of chamber designs.

This paper describes the results from a coordinated set of chamber validation experiments

conducted at 6 chamber facilities at 5 leading agricultural research centres around the UK: the

Agri-Food and Biosciences Institute, Hillsborough; the Institute of Biological, Environmental

and Rural Sciences in Aberystwyth; Scotland’s Rural College, Edinburgh; the Division of

Animal Sciences in the School of Biosciences, University of Nottingham; and the School of

Agriculture, Policy and Development at the University of Reading.

Materials and methods

All of the test chambers across the six facilities were based on the same basic design principle

(Fig. 1), although there were marked differences in terms of size, flow conditions and age

across the different facilities. In all cases, ambient air is drawn into the chamber and mixes

with the emissions from the test subject before being vented to atmosphere via an extract

duct. An anemometer (hot wire or vane based) is positioned in the extract duct to determine

the chamber flow rate whilst an interfaced gas line is used to pump a sample of the extract

gas through an analyser to determine the methane concentration. Combining the flow rate and

concentration measurements allows the emitted flux to be calculated using in-house

methodologies. The details of the chamber designs and the differences between them are

beyond the scope of this paper and are only discussed if relevant to the reported observations.


Whilst incumbent facility staff were present in order to operate chambers and explain

configuration differences, all experiments were carried out by independent researchers from

the National Physical Laboratory.

A calibrated source of methane flux was produced by dynamically mixing ultra-high purity

methane (BOC Gases, ≥ 99.9995 % purity) and nitrogen (Air Products BIP grade, < 50 ppbv

methane equivalents of hydrocarbon contamination) using an bespoke blender based on Aera

FC-7000 series mass flow controllers (MFCs). The blender system consisted of two pairs of

MFCs. Each pair consisted of a MFC delivering methane and the other delivering nitrogen,

with one pair set up for chambers usually measuring sheep and the other for chambers usually

measuring cattle. The flows from the MFC pairs were set to provide an approximately

constant total flow of gas independent of the amount of methane being delivered. Rather than

relying on the manufacturers specifications, each MFC was directly calibrated for flow rate of

the relevant gas via weight loss using NPL’s gravimetric gas standard preparation facilities,

which are recognised by the International Committee for Weights and Measures

(http://kcdb.bipm.org/, accessed April 2014) as providing gaseous reference materials for

calibration of UK laboratories to internationally validated levels of uncertainty

(http://kcdb.bipm.org/appendixc/qm/GB/qm_gb_4.pdf, accessed April 2014). This enabled

mass emissions with an uncertainty of 1.0% (coverage factor of k = 2, 95% confidence level

– written as ‘k=2, 95% confidence’ hereafter) to be generated. The pair set up for sheep

chambers were typically used to deliver 0.4 mg/s (~0.035 l/min) of methane in a total flow of

~1 l/min, while the pair set up for cattle chambers were typically used to deliver 6.0 mg/s

(~0.5 l/min) of methane in a total flow of 3 l/min. The outputs from the MFCs were

combined using ¼” stainless steel tubing and Swagelok fittings. The blender system was leak


http://kcdb.bipm.org/

http://kcdb.bipm.org/appendixc/qm/GB/qm_gb_4.pdf

tested with soap solution prior to use and the line was isolated overnight and demonstrated to

maintain pressure over a 12 hour period with no significant losses.

The source was delivered from the blender via a bespoke sample line and dispersion system

into the chambers. The sample line was a single continuous length of ¼” perfluoroalkoxy

(PFA) tubing with stainless steel fittings from Swagelok. The dispersion system consisted of

a series of ¼” Swagelok T-pieces which spread the emission over a volume of ~4300 cm3

through 18 separate outputs, without putting a restriction on the output flow.

Each facility consisted of several chambers, since in normal operation a group of test animals

is passed through all the chambers in order to determine an emission rate of statistical

significance. To facilitate evaluation, and help identify the sources of measurement

uncertainty, the chambers were considered as having three principal components: the

methane sensor; the ducting and flow system extracting gas out of the chamber; and the

chamber itself.

All six facilities used infrared gas filter correlation sensors to measure the methane

concentrations in the chambers. The sensor responses were tested by applying a series of

NPL prepared standards of methane in synthetic air (i.e. N2 and O2 only) designed to span

the concentration range typically seen by the particular sensor being tested. The standards

were introduced to one of the sensor sample ports using the by-pass flow arrangement shown

schematically in Figure 2. This arrangement ensured the sensor was able to take the required

sample volume without over-pressuring the input line, together with the ability to rapidly

switch between ambient air and calibration gas without disrupting the sample flow. A total of

eight calibration standards were prepared specifically for these tests to cover the complete

range seen by the different facilities, with methane concentrations from 10 ppmv to 500


ppmv These standards were prepared in NPL’s gravimetric gas standard facilities and

traceably certified to an uncertainty of 0.5 % (k = 2, 95% confidence). Each facility’s sensor

was tested for accuracy, linearity and response time. Carrying out a linear regression between

the sensor readings and reference values provided a linear calibration function for the sensor.

The response time was defined as the time taken to reach 90% of the final stable reading

(T90) when the sample was switched from ambient air to calibration gas, in accordance with

EN 15267-3. Note that each facility had a data logging system implemented to record the

sensor readings, and this was often on a relatively slow timescale compared to individual

sensor readings. Therefore manual readings of the sensor were taken every 10 seconds to

provide response time data.

The ducting efficiency was tested by directly releasing a known flux of methane inside the

ducting close to the interface with the chamber, i.e. the sample delivery pipe (without

diffuser) was inserted a few centimetres into the duct. The usual chamber emission flux

calculations could then be carried out and compared to the known emission rate from the

calibrated methane source, giving a calibration measurement that was not influenced by the

chamber itself. If this calculation is carried out after applying the sensor calibration function

to the methane concentration readings then the efficiency of the ducting can be determined in

isolation. Any deviation from unity could highlight issues with the accuracy of the flow

measurement combined with any losses or sampling issues in the duct itself. Time restrictions

meant that it was not possible to carry out ducting efficiency measurements for every

chamber at every facility, but ducting measurements were made for at least one chamber at

each facility to assess any issues with that particular facility’s design and methodology.


The final stage of each experiment was to release a known flux of methane through the

diffuser in the chambers themselves and compare this to the methane flux determined by the

facility. A flow diffuser system was used to spread the flow of methane emission, and this

diffuser was located at a typical animal head position in each chamber. This measurement

provided a direct calibration of the overall chamber emission measurement. However by

applying the previously calculated senor calibration function and ducting efficiency to the

readings it also allowed the chamber performance to be evaluated in isolation and the

function determined for chamber capture efficiency. Since the calibrated methane source

could be used to apply a step change to the methane levels inside a chamber, the results from

the chamber measurements were also used to determine the response times of the chambers

themselves.

Results and discussion

Calibration of Methane Sensors

Figure 3 shows the typical response time of one of the methane sensors used in a chamber

facility when a methane calibration standard is sampled. In this case the sensor reached 90%

of the stable plateau reading in 37.1 s. Table 1 summarises the results of the methane sensor

calibration experiments for all 6 (anonymised) facilities. The test span value gives the range

of concentrations over which the sensor was calibrated, while the calibration factor gives the

adjustment factor that has to be applied to the sensor reading at span together with the related

(k=1, 67% confidence) uncertainty. The plateau stability indicates the 1-σ noise level on the

stable plateau reading, and the sensor response times are given by the T90 values. Finally the

linearity of the sensor response over the measurement range is given by the R2 value of a

linear regression fit to the calibration data.


All sensors showed good linearity over the measurement range with R2 greater than 0.996 in

all cases, and high levels of accuracy with all responses equivalent to the reference value at

the 95% confidence uncertainty level. The plateau stabilities showed a general level of

instrument precision of 1% or better, with reasonably consistent response times varying

between 25 s and 39 s.

These results confirm the general suitability of the methane sensors in the ranges used for

each chamber facility. However, there were a number of specific issues that arose from the

sensor tests that could have implications for overall facility operations, and these are

discussed below:

• All groups performed regular span checks of the sensor using a reference gas to

ensure the long-term stability of the measurements, but it is important to ensure that

the actual reference gas used provides a span value close the measurement range.

• In some cases the data logging software used to record the instrument data did not use

the same output as the reading displayed on the sensor – many instruments have both

analogue and digital outputs. In this case the user should confirm that any quality

checks and calibration adjustments are relevant to the data that is recorded.

• The instrument response time provides the user with data on the appropriate time

between samples. In order to ensure that one reading is independent from the previous

one it is recommended to leave more than (3 x T90) between readings, at which point

one reading will have less than 0.1% influence on the next. If the time between

readings is too short then this can lead to significant biases in the data – as an

example, a typical sampling configuration is to alternate between ambient background

readings and chamber readings. If readings were taken at the T90 response time then,

for typical background and chamber levels of 2 ppmv and 200 ppmv respectively, the


chamber reading would be 180.2 ppmv and the background reading would be 21.8

ppmv. The resulting differential methane level would therefore be 158.4 ppmv rather

than the correct value of 198 ppmv, leading to a 20% underestimation of methane

emissions. This shows the importance of ensuring enough time between

measurements to ensure independent readings.

Ducting Efficiency

Figure 4 summarises the results of injecting a known methane emission rate directly in the

sample duct for an example chamber in 5 of the 6 facilities (the design at facility D prevented

this type of measurement). Each column shows the ducting efficiency for an individual

chamber together with the associated (k=2, 95% confidence) uncertainty on the result, so a

ducting efficiency greater than one indicates a reading above the known emission rate and a

value below one indicates a reading below the known emission rate. Note that these

efficiency values have been calculated having allowed for the sensor calibration results

described above.

These results show a much wider spread than seen for the methane sensor calibration results

and, given that methane sensor effects have been allowed for, this must be due to a

combination of sampling issues/losses in the ducting and uncertainty in the chamber flow

measurements.

Chamber extracted air flow measurement presents a challenging issue, both in terms of

calibration and adjustment to the ambient conditions at the time of measurement. If we

assume the final methane emission rate will be reported as a mass flow (i.e. grams of methane

per hour), then the chamber flow measurement also needs to be determined as a mass flow.


Therefore, if the air flow measurements are actually a volume flow measurement (i.e. m/s),

then an air density correction will be needed to convert to a final methane mass emission rate,

and this will require calibrated temperature and pressure (and potentially humidity)

measurements to be made for the sample air at the point of the flow measurement. Whether

this correction is needed comes down to the nature of the flow measurement method. In

general terms, vane-based flow sensors measure the volume flow while hot-wire-based

sensors measure the mass flow, but the details depend upon the exact nature of the sensor

used. The location of the flow sensor within the duct can also influence the flow reading due

to flow variation across the pipe diameter, as can obstructions and bends in the pipework. All

of these issues make validation of the chamber flow reading particularly difficult, and indirect

validation through full system calibration using chamber recovery tests or more extensive

experiments such as described here probably provide the most viable way of assessing the

accuracy of the flow measurements.

Chamber Response

Figure 5 shows a typical response curve following the injection of a controlled source of

methane into the chamber, and Table 2 summarises the response results from all 6 facilities.

Since the controlled emission source is effectively constant during the measurements the

plateau stability reflects the flow variability within the chamber and sampling ducts, and

provides a measure of the precision of a single chamber measurement point. These results

show much greater variability in both plateau stability and response time, with the slower

response chambers showing better plateau stability. This is not unexpected as slower

response chambers will tend to smooth out any short-term variation in the flow, and there is

significant variation between chamber volume and air exchange flows which drives the

chamber response times.


Another aspect of the chamber response times that should be considered is that they give

information on the period a disturbance in the chamber, e.g. the chamber doors being opened,

will continue to influence the readings. If the operator wishes to exclude data affected by

known disturbances then results should only start to be included once a period of three times

the T90 response time has passed.

Although there are large differences in the precision of a single measurement, when we

consider the precision on a 24 hour average (4th column of Table 2) there is a much reduced

spread. This value is derived from the expected N1/2 reduction in the measurement noise that

results from averaging N independent measurements. The time between independent

measurements is taken as the (3 x T90) period for each chamber, i.e. there will be more

independent measurements from the faster response chambers within a given averaging time.

This result shows that, for a typical 24 hour measurement period during an animal

experiment, all the chambers have a measurement precision uncertainty of better than 1%.

Note that this precision value just shows the variability between one measurement (or

average) and the next, and it does not reflect the absolute accuracy of the measurement which

is discussed in the following sections.

The final column of Table 2 shows the linearity of the chamber readings to varying levels of

controlled methane emissions. These values are slightly lower than those seen for the sensors

themselves (see Table 1), but all show highly linear performance with R2 values of 0.99 or

higher for all chambers.


Chamber Capture Efficiency

As discussed above, the ducting efficiency was determined for an example chamber at all

facilities except one (facility D), where facility configuration made the test impossible. Figure

6 shows the ratio between a reference flux release in the chamber itself against the measured

value corrected using the respective facility sensor calibration function and ducting efficiency

factor. Such correction removes any bias in concentration determination and / or flow

measurement, isolating any differences from the reference value to the chamber itself. The

result for facility D is the combined efficiency of the ducting and chamber.

It can be seen that there are cases of both over- and under-estimation in the chamber capture

results, so it is not just a case of methane leaks out of the chamber. Given that all facilities

operate chambers at pressures slightly below atmospheric, leaks into the chamber are more

likely than losses out. An inward leak of ambient air into the chamber should not cause a

problem, as long as the air around the chamber has the same methane concentration as the

main external air inlet. However, this may not always be the case depending on farm activity

(e.g. nearby ruminants emitting methane) and the presence of any other local methane

sources.

Another effect which could cause the observed deviations is inhomogeneous mixing within

the chamber and ducting. Some limited testing of source location dependence was carried out

during the experiments. This showed that, in some chambers, the readings changed when the

emission source was moved between different locations to simulate animal movements, e.g.

feeding or sleeping positions (data not shown). This indicates that the intake air and emission

source gases are not well mixed at the point where a sample is extracted from the chamber for

measurement. The dependence of the determined emission on source location is a subject that


this work has highlighted would be an important topic for further investigation. This effect is

one of a number of cases where the presence of an animal in the chamber could potentially

influence the results. It was beyond the scope of this work to address this in more detail,

however the experiments described here establish the baseline chamber performance and the

underpinning measurement uncertainty.

Combined facility results

Figure 7 shows the combined validation factor for every chamber tested across the six

facilities, i.e. a single value combining any bias found in sensor calibration, ducting

efficiency and chamber capture efficiency, together with the associated (k=2, 95%

confidence level) uncertainty on each factor. As can be seen the spread in combined

validation factors across all the chambers is marked, with some chambers producing

measurements of less than half of that of others. An important question is whether this

variability is due to chamber-to-chamber differences within facilities or facility-to-facility

differences. Therefore, it is useful to determine the overall facility correction factor as is

shown in Table 3. These data demonstrate that the inter-facility variance is of a similar

magnitude to that between individual chambers. This result shows that the facility design and

operation is the largest source of absolute uncertainty rather than chamber-to-chamber

variability or instrumental noise. This result also confirms the suitability of each facility to

carry out relative measurements, e.g. to compare the effectiveness of different treatments, but

highlights the importance of this type of validation exercise in evaluating absolute

uncertainties and establishing comparability between different facilities.

Table 4 shows a measure of overall capability for the facilities evaluated by taking the mean

of the individual facility validation factors and providing the 1-σ spread (i.e. the standard


deviation) in these values. Table 4 also shows the mean and spread values for the three main

components of a chamber facility – the sensor, the ducting and flow system, and the chamber

itself. These results highlight that it is the ducting and flow system that is the main source of

uncertainty in the combined chamber results.

The upper and lower panels of Table 4 shows these values with and without including data

from facility B, as this was a new facility which had undergone no quality assurance testing

prior to these measurements (and therefore had no prior influence in determining UK

livestock emissions). The validation tests on facility B revealed a significant issue with the

design which has since been rectified, and results from this facility are therefore excluded

from the following general discussion.

The final element of the work was an attempt to quantify the difference the validation

exercise had on the overall capability across all the facilities tested. The most appropriate

measure to do this is to consider how the absolute uncertainty has changed. The absolute

uncertainty is made up of both bias sources (e.g. ducting efficiency) and random sources (e.g.

noise associated with methane sensor). The significant bias sources for each facility are all

incorporated into the validation factor, i.e. the difference between the validation factor and

unity measures the impact of off-sets in the sensor calibration, ducting efficiency and

chamber capture. Random uncertainty sources such as methane sensor noise, anemometer

noise, etc. are incorporated into the 24 h precision values shown in Table 2.

Prior to the validation exercise the bias terms were unknown, and the overall uncertainty

would be dominated by these terms. The distribution of Combined Validation Factors can


therefore be used to give an estimate of the absolute uncertainty associated with the overall

capability across all facilities of 25.7% (k=2, 95% confidence).

Following the validation exercise, if each facility applies the provided validation factor to

future measurements then in principle the aforementioned bias uncertainty sources are

removed leaving only random sources (i.e. determined in the 24 h precision test) and the

uncertainties associated with the determination of the validation factors. Hence, the

uncertainty estimate decreases to 2.1% (k=2, 95% confidence). It should be noted that this

makes the critical assumption that all the facilities remain completely unchanged from the

time the validation exercise was carried out, which is unlikely. However, whilst effects such a

drift will result in an uncertainty increase from that above, it would require a very substantial

facility change before values of 25.7% are approached. This notwithstanding, if a regime was

put in place to repeat some of the measurements on a periodical basis this would ensure the

uncertainty remains close to the 2.1% estimate. Overall, the data evidence the importance of

validating national measurement infrastructure against traceable references and the potential

value that can be added to future measurements as a result, particularly when looking to

determine absolute emission values and when combining results from different facilities.

Acknowledgements

The authors would like to thank Elena Amico di Meane, Marta Doval Minarro and Ian

Uprichard from the Gas and Particle Metrology Group at NPL for the calibration of the mass

flow controllers and preparation of the reference methane standards. In addition, we would

like to acknowledge the invaluable support during the field campaigns from the researchers

and technical support teams at each chamber facility.


This work was funded by Department for Environment, Food and Rural Affairs, the Scottish

Government, Department of Agriculture and Rural Development (N.I.), and the Welsh

Government as part of the UK's Agricultural GHG Research Platform project

(www.ghgplatform.org.uk). The underpinning metrology was supported by the National

Measurement System’s Chem-Bio Knowledge Base programme under the Environmental

Technologies theme.

References

Armsby, H. P. 1908. The Principles of Animal Nutrition. John Wiley and Sons, New York,

NY, USA.

Cammell, S. B., D. E. Beever, K. V. Skelton, and M. C. Spooner, 1981. The construction of

open-circuit calorimeters for measuring the gaseous exchange and heat production in sheep

and young cattle. Laboratory Practice 30:115-121.

Derno, M., H. G. Elsner, E. A. Paetow, H. Scholze, and M. Schweigel, 2009. Technical Note

: A new facility for continuous respiration measurements in lactating cows. J. Dairy Sci.

92:2804-2808

EN 15267-3: 2007. European Standard : Air quality, Certification of automated measuring

systems. Performance criteria and test procedures for automated measuring systems for

monitoring emissions from stationary sources.

European Communities. 2006. The European Climate Change Programme (ECCP). Accessed

Feb 11, 2014. http://ec.europa.eu/clima/policies/eccp/index_en.htm


http://ec.europa.eu/clima/policies/eccp/index_en.htm

European Environment Agency. 2013. Trends and projections in Europe 2013 – Tracking

progress towards Europe's climate and energy targets until 2020. EEA Report No 10/2013

Garnsworthy, P. C., J. Craigon, J. H. Hernandez-Medrano, and N. Saunders. 2012. On-farm

methane measurements during milking correlate with total methane production by individual

dairy cows. J. Dairy Sci. 95:3166-3180.

Grainger, C., T. Clarke, S. M. McGinn, M. J. Auldist, K. A. Beauchemin, M. C. Hannah, G.

C. Waghorn, H. Clarke, and R. J. Eckard. 2007. Methane emissions from dairy cows

measured using the sulphur hexafluoride (SF6) tracer and chamber techniques. J. Dairy Sci.

90:2755-2766.

Hellwing, A. L. F., P. Lund, M. R. Weisberg, M. Brask, and T. Hvelplund, 2012. Technical

Note : Test of a low-cost and animal friendly system for measuring methane emissions from

dairy cows. J. Dairy Sci. 95:6077-6085

IPCC 2006, 2006 IPCC Guidelines for National Greenhouse Gas Inventories, Prepared by the

National Greenhouse Gas Inventories Programme, Eggleston H.S., Buendia L., Miwa K.,

Ngara T. and Tanabe K. (eds). Published: IGES, Japan.

IPCC, 2013: Climate Change 2013: The Physical Science Basis. Contribution of Working

Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change.

[Stocker, T. F., D. Qin, G.-K. Plattner, M. Tignor, S. K. Allen, J. Boschung, A. Nauels, Y.


Xia, V. Bex and P. M. Migley (eds.)]. Cambridge University Press, United Kingdom and

New York, NY, USA, 1355 pp.

Johnson, K., M. Huyler, H. Westberg, B. Lamb, and P. Zimmerman. 1994. Measurement of

methane emissions from ruminant livestock using a SF6 tracer technique. Environ. Sci.

Technol. 28:359-362.

Klein, L. and A. D. G. Wright, 2006. Construction and operation of open-circuit methane

chambers for small ruminants. Aust. J. Exp. Agric. 46:1257-1262

McLean, J. A. and G. Tobin. 1987. Animal and Human Calorimetry. Cambridge University

Press, Cambridge, UK.

McCourt, A. R., T. Yan, and S. Mayne. 2008. Validation of the sulphur hexafluoride (SF6)

tracer technique as a method of predicting methane output from dairy cows. Proc. Brit. Soc.

Anim. Sci. 120. Scarborough, UK.

Muñoz, C., T. Yan, D. A. Wills, S. Murray, and A. W. Gordon. 2012. Comparison of the

sulfur hexafluoride tracer and respiration chamber techniques for estimating methane

emissions and correction for rectum methane output from dairy cows. J Dairy Sci. 95:3139-

48.

Murray, P. J., A. Moss, D. R. Lockyer, and S. C. Jarvis. 1999. A comparison of systems for

measuring methane emissions from sheep. J. Agric. Sci. 133(Pt4):439-444.


Pinares, C., and G. Waghorn (Eds), 2012. Technical manual on respiration chamber designs.

Ministry of Agriculture and Forest, Wellington, New Zealand. Accessed Feb 11,

2014. http://globalresearchalliance.org/app/uploads/2012/03/GRA-MAN-Facility-BestPract-

2012-ch1.pdf and related links.

Yan, T., C. S. Mayne, F.G. Gordon, M. G. Porter, R. E. Agnew, D. C. Patterson, C. P. Ferris,

and D. J. Kilpatrick. 2010. Mitigation of enteric methane emissions through improving

efficiency of energy utilization and productivity in lactating dairy cows. J. Dairy Sci.

93:2630-2638.


http://globalresearchalliance.org/app/uploads/2012/03/GRA-MAN-Facility-BestPract-2012-ch1.pdf

http://globalresearchalliance.org/app/uploads/2012/03/GRA-MAN-Facility-BestPract-2012-ch1.pdf

Tables

Facility Test Span (ppmv)

Calibration Factor

Plateau Stability

T90 Response (sec)

Linearity (R2)

A 500 0.987 +/- 0.009 0.11% 37.1 0.9999

B 100 1.040 +/- 0.020 0.93% 33.7 0.9999

C 50 0.978 +/- 0.083 0.34% 38.1 0.9999

D 100 1.086 +/- 0.075 0.63% 35.4 0.9965

E 200 1.008 +/- 0.024 0.57% 28.9 0.9995

F 500 0.995 +/- 0.013 1.01% 25.6 0.9999

Table 1 Summary of Sensor Calibration Results, showing for each facility sensor assessed :

the span of concentrations covered, the calibration factor and its uncertainty at the span value,

the variability once a stable reading is reached, the T90 response time and the R2 linearity of

response across the calibrated range.


Facility Plateau Stability

T90 Response (min sec)

Precision on 24 hr

average Linearity

(R2)

A 1.33% 40' 39'' 0.39% 0.9996

B 1.63% 21' 58'' 0.35% 0.9995

C 11.00% 01' 13'' 0.55% 0.9981

D 4.27% 09' 00'' 0.58% 0.9900

E 2.07% 27' 42'' 0.50% 0.9971

F 2.70% 54' 05'' 0.91% 0.9999

Table 2 Summary of Chamber Response Tests, giving for an example chamber at each

facility : the variability on a stable reading, the T90 response time, the measurement precision

extrapolated to a 24 hr average, and the R2 linearity of response across the tested range.


Facility Combined Validation Factor

Uncertainty (k=2)

A 1.045 0.004 B 0.590 0.005 C 1.154 0.030 D 0.827 0.028 E 0.945 0.013 F 0.897 0.008

Table 3 Combined Facility Validation Factors, giving the combined validation factor across

all chambers at each facility, together with the (k=2, 95% confidence) uncertainty on the

derivation of each factor.


Combined Validation Factor Mean 1-σ Spread

Complete facility 0.9097 19.4% Methane Sensor 1.001 2.4% Ducting incl. flow 0.9308 19.8% Chamber only 0.9849 4.0%

Combined Validation Factor Mean 1-σ Spread

Complete facility 0.9736 12.8% Methane Sensor 0.992 1.3% Ducting incl. flow 0.9968 15.3% Chamber only 0.9971 3.4%

Table 4 Chamber Performance Summary, showing the mean validation factor across the

facilities and the 1-σ spread of values. The results are given for the combined facilities, and

separately for the three main system elements in each facility (sensor, ducting and flow,

chamber). The upper table includes all chamber facilities, the lower table excludes the new,

untested facility (facility B).


Figures

Figure 1 Schematic of operational principle of livestock respiration chambers across five UK

research farms. Mass flow controllers, sampling line and diffuser only included during

National Physical Laboratory testing of facilities.


Figure 2 Schematic of gas delivery system used to evaluate the response of the methane

sensors. The 3-way value allows switching between ambient and calibration gas, and the

rotameter is used to ensure a positive by-pass flow to the vent.


Figure 3 Response of a methane sensor to injection of calibration standard of methane. Black

diamonds show the sensor’s response with time measured every 10 s, the solid line shows the

plateau reading level, and the dotted line shows the 90% value of plateau reading.


Figure 4 Ducting efficiency results. Columns give the efficiency values for the individual

systems tested together with the (k=2, 95% confidence) uncertainties on the determination of

the efficiencies.

*Note that, due to facility design, it was not possible to determine the ducting efficiency

result for facility D.


Figure 5 Response of system to injection of controlled flow of methane into chamber. Black

diamonds show the measured response with time, the solid line shows the plateau reading

level, and the dotted line shows the 90% value of plateau reading.


Figure 6 Chamber Capture Efficiencies. Columns give the efficiency values for the

individual chambers tested together with the (k=2, 95% confidence) uncertainties on the

determination of the efficiencies.

*Note that for facility D the efficiency shown is the combined efficiency of the chamber

capture and the ducting.


Figure 7 Individual Compete Chamber Efficiencies. Columns give the efficiency values for

the individual systems tested together with the (k=2, 95% confidence) uncertainties on the

determination of the efficiencies.


Date post:	26-Oct-2020
Category:	Documents
Upload:	others
View:	2 times
Download:	0 times

Project AC0115 Final Report -...

Documents