Estimating Full IM240 Emissions from Partial Test Results: Evidence from Arizona

Area of Science:

• Environmental Science
• Automotive Engineering
• Data Science

Background:

• Enhanced vehicle emissions testing, like the IM240 dynamometer test, is costly and time-consuming.
• Many states implement shortened tests or 'fast-passes' to reduce expenses, but this results in data loss and non-comparable results.
• Current partial-test results hinder accurate fleet emissions characterization and evaluation of emissions control programs.

Purpose of the Study:

• To develop a methodology for estimating full 240-second vehicle emissions (HC, CO, NO_x) from partial-test results.
• To enable states to convert all tests to consistent IM240 readings for improved emissions analysis.
• To evaluate the impact of inspection and maintenance programs on emissions over time.

Main Methods:

• Regression analysis was used on a sample of Arizona vehicles with full 240-second tests.
• Estimated the relationship between emissions at second 240 and earlier test seconds.
• Examined the influence of vehicle age, model-year group, and pollution level on emissions estimations.

Main Results:

• A methodology was presented to estimate full 240-second emissions from partial-test data for HC, CO, and NO_x.
• Probit analysis indicated that longer test progression increases the likelihood of a failing vehicle passing on retest.
• Forecasted fleet average emissions for light-duty vehicles closely matched actual averages, though accuracy decreased for trucks, especially for NO_x.

Conclusions:

• Estimating full test emissions from partial data allows for consistent fleet emissions characterization.
• The methodology supports better evaluation of emissions control programs and vehicle repair effectiveness.
• While effective for light-duty vehicles, further refinement is needed for accurate truck emissions forecasting.