PEMS data analysis at ICCT Vicente Franco Francisco

PEMS data analysis at ICCT Vicente Franco, Francisco Posada, Peter Mock Presentation RDE-LDV group Brussels, Sept. 16 2013

Contents § ICCT approach to PEMS data analysis § ICCT standard PEMS charts § A modest proposal

ICCT approach to PEMS data analysis § Problem § PEMS campaigns by contractors § Data and analysis commissioned together § Different contractors performs different analyses § No direct control over type of analysis, quality § Some inconsistencies, makes comparisons difficult § Solution § Commission PEMS testing + (methodological) campaign report excluding data analysis, develop semi-automatic, standard data analysis formats

ICCT approach to PEMS data analysis § ICCT’s Standard PEMS trip report charts § Goal: to have a standard set of charts that characterize individual PEMS trips and real-world emission behavior of vehicles § § Compare real-world to laboratory, emission limits Compare different trips Compare different vehicles Support scientific analyses, not (just) compliance checks § Analysis method draws from RDE-LDV interim methods

ICCT standard PEMS charts § Standard PEMS trip report charts § Workflow: MATLAB script that loads PEMS datasheets, applies MAW method (also simplified CLEAR w/o weighting) § Set of standard plots (from aggregate to detailed assessments) § Ready for publishing (eps format) § Adapts to different data sources, ‘batch’ mode § Not a standalone tool (requires MATLAB scripting)

ICCT standard PEMS charts: trip overview Generated for individual trips or collections of trips Quick qualitative assessment of driving conditions

ICCT standard PEMS charts: comparison to emission limits

Standard plots: pollutants by window Standard visualization of CO 2 windows, one trip

Standard plots: pollutants by window

Standard plots: distance-specific emissions One or several trips (time dimension preserved) Fuel-specific version possible

Standard plots: distance-specific emissions

Standard plots: pollutant scatterplots Investigation of correlation between emissions and window average gradient, power, etc. Marker size proportional to window distance

Distribution of driving modes

Standard plots: ‘emission map’ Preferred approach: not to derive ‘map’ plots from OBD data (GPS instead)

‘Map’ instantaneous emission plots Scatterplots of instantaneous emissions @ 3 -second bins • Not used to assess compliance • Illustration of emission events

A modest proposal Proposal: to make a smooth classification of windows based on their distance-specific CO 2 emissions Reasoning: aggressive driving, high-speeds, uphill driving, cold-start, cold ambient temperature, etc. all translate intro increased CO 2 emissions. Distance-specific window emissions should get multiplied by a factor <1 when distance-specific CO 2 over the window is higher than for the reference cycle:

A modest proposal

A modest proposal

A modest proposal Advantages of distance-specific CO 2 weighting: • Smooth weighting is fairer than discrete areas; distance-specific CO 2 is easily determined (no subcycle data required, no curve fitting, no extrapolation) • Compatible with the subsequent application of an NTE limit

PEMS data analysis at ICCT Vicente Franco, Francisco Posada, Peter Mock Presentation RDE-LDV group Brussels, Sept. 16 2013