Andy Grieshop Assistant Professor Civil, Construction & Environmental Engineering North Carolina State University

Transcription

1 Aerosol emissions from cookstove biofuel combustion: measurement, characteristics, and uncertainties Andy Grieshop Assistant Professor Civil, Construction & Environmental Engineering North Carolina State University Assessing the Climate and Health Co-Benefits of Clean Cooking July, 2015 UN Foundation -- Washington D.C. 1

2 Burning questions I ll touch on What kind of aerosols/pm are emitted by biofuel (and other) combustion and how do we measure them? How/why do laboratory and field measurements of these differ so greatly, and (how) can we fix this? How do processes affecting emitted species contribute to our assessment of impacts/cobenefits? 2

3 A tiny fraction of fuel carbon ends up as PM, but this dominates PM mass emitted Elemental Carbon Figure: For improved ceramic stove, Smith et al, EPA, 2000 Organic Carbon (OC) 3

4 How does it get there? A fire is a complex reactor First: Pyrolysis and/or Volatilization O 2 The final mix after cooling Semi- or Non-volatile Organics Second: Soot (BC) Formation Images: FAO, Lund U, Johnson Matthey 4

5 Classifying carbonaceous PM components and their impacts Image: Andreae, M. O., and Gelencsér, A. (2006). Atm Chem Phys 5

6 BC is an optical measurement: methods measure light absorption (not mass!) Off-line methods Increasing complexity/cost Reflectometer Integrating plate Low-cost imaging (e.g. camera, scanner) Filter-based absorption Aethalometers Particulate Soot Absorption Photometer (PSAP) Photo-acoustic Photo-acoustic Extinctiometer (PAX) 6

7 Thermo-optical Organic/Elemental Carbon (OC/EC) analysis OC EC TC NIOSH 5040 / Sunset Labs / DRI Measures mass, but OC/EC split is operationally defined 7

8 Burning questions What kind of aerosols/pm are emitted by biofuel (and other) combustion and how do we measure them? How/why do laboratory and field measurements of these differ so greatly, and (how) can we fix this? What processes affecting emitted species contribute to our assessment of impacts/cobenefits? 8

9 What do emissions look like in a real intervention? Intervention trial in Karnataka, India Carbon-credit financed (first in India) stove change-out implemented by SAMUHA (Karnataka-based NGO) Coverage: 110 villages in rural Karnataka Intervention trial conducted in pilot village: ~190 Households Intervention stoves: Chulika 9

10 PM 2.5 Emission Factor (g kg -1 ) PM emission factors: a messy story from the field Before Intervention (Oct. - Dec.) After Intervention (May - July) Control Group N = 26 Chulika Group N = 21 Control Group N = 29 Chulika Group (partial) N = 8 Chulika Group (full) N = Range of lab tests (Roden et al., 2009) Chulika lab tests (Just et al., 2013) 10

11 Rocket stove: higher temperature higher EC emissions EC Emission Factor EC:TC Ratio 11

12 PM emissions in lab tests are substantially lower, EC less so Field Trad Field Imp. 12 Relative difference in mean EFs (Field to Lab): PM EF: Traditional = -45% Improved = -58% EC EF: Traditional = +15% Improved = -21%

13 Increasing Efficiency Patterns of Real-Time Emission Data (PaRTED) analysis helps link activity to emission characteristics 2-D frequency plot Type of particle During what type of combustion event MCE = CO 2 (CO+CO 2 ) SSA = Scattering Extinction PaRTED: Chen et al., ES&T, 2012 Light Absorbing Light Scattering 13

14 Patterns of real-time emissions are highly distinct as well Traditional Stove Chulika Field Lab (WBT) 14

15 So field and lab emission measurements don t agree Why do we care? Critical for predicting exposures Need to know for emission inventories Will we get expected benefits? How clean is clean enough? Why the differences? Fuel type/condition/moisture? Sampling approach? Cooking practices? Use patterns? 15

16 The idea: Histogram of (ΔCO+ΔCO 2 ) normalized by max value indicates fire power distribution during use Max (CO+CO 2 ) 16

17 Lab test (WBT) activity shows little resemblance to actual use. Field Traditional Stove Chulika N = 29 N = 14 Lab (WBT) Missing Activity N = 3 Missing Activity N = 6 17

18 New Testing Protocol (NTP): recreate firepower distributions from field using real-time feedback to tester Real-time histogram Target histogram Based on in-home observations More representative Less prescribed Less repeatable 18

19 New testing approach can better recreate inhome fire power distributions Traditional Stove N = 29 N = 14 Chulika Field N = 3 N = 4 New Lab Approach 19

20 Field vs NTP Emission Factors 20

21 Real time emission patterns are still quite distinct from field observations Traditional Chulika Field WBT New Testing Protocol 21

22 Take home points Burning biomass is complex and hard to do in an entirely clean way need to keep evolving carbon in contact with air and thermal energy Laboratory tests do not represent in-home performance or emission characteristics. Initial efforts to recreate field use in the lab show promise, though more work is needed. Fuel type, quality and other factors may have important influences and are difficult to explore in the lab (in the US). 22

23 Burning questions What kind of aerosols/pm are emitted by biofuel (and other) combustion and how do we measure them? How/why do laboratory and field measurements of these differ so greatly, and (how) can we fix this? How do processes affecting emitted species contribute to our assessment of impacts/cobenefits? 23

24 Organic aerosols are highly dynamic. Dilution Dilution and photochemical processing NO 3 OH h O3 Evaporation / repartitioning Photochemical aging deposition Near Source Regional (~ m) (1000+ km) Spatial Scale 24

25 Atmospheric chamber studies show that aging wood-smoke can rapidly generate large amounts of new organic aerosol Primary Organic Aerosol (POA) Grieshop et al., ACP

26 Dynamics of primary emissions may have important implications for local/regional and especially global impacts Concentration Nonvolatile Semivolatile + Aging Near source Village Regional Continental Spatial Scale 26

27 Evolution of biomass PM Increase of OC with aging will influence climate/health impacts of both baseline and alternative technologies Will affect amount and properties of PM Likely not an important issue for indoor exposure, but could have important impacts on regional/global climate/aq impacts. This behavior is being studied and starting to be incorporated in models, but is highly uncertain. 27

28 Acknowledgements (Field Trial) U British Columbia T Aung, M Brauer, M Kandlikar, H Zerriffi U Minnesota J Marshall, J Baumgartner, C Reynolds, M Bechle ROI (Bangalore, India) G Jain, Karthik S Research Team, Summer 2011 Samuha (Koppal, India) T Pradeep, Narayanswami S, field staff Village households Funding: Institute on the Environment (University of Minnesota); NCSU College of Engineering; The Liu Institute for Global Issues and Bridge Program (University of British Columbia) 28

29 Thank you. Students: Ryan Repoff Mutian Ma Supported by: 29