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
44
Embed
Project AC0115 Final Report - Defra, UKrandd.defra.gov.uk/Document.aspx?Document=13310_AC... · Project AC0115 Final Report ... gravimetric reference standards for direct calibration
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
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.
AC0115 Final Report Appendix B - Page 2
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.
AC0115 Final Report Appendix B - Page 3
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.
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
AC0115 Final Report Appendix B - Page 4
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
AC0115 Final Report Appendix B - Page 5
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
AC0115 Final Report Appendix B - Page 6
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.
AC0115 Final Report Appendix B - Page 7
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%).
AC0115 Final Report Appendix B - Page 8
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)
AC0115 Final Report Appendix B - Page 9
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
AC0115 Final Report Appendix B - Page 10
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
AC0115 Final Report Appendix B - Page 11
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 :
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.
AC0115 Final Report Appendix B - Page 12
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.