As vehicle manufacturers, regulators and governments around the world seek solutions to prevent another emissions cheating scandal similar to the Volkswagen case, a major vehicle emissions inspection company has compiled and analysed on-road emissions data indicating that emissions violations of vehicles under real-world driving conditions may well go far beyond VW diesels. Opus Inspection says a two-pronged approach that continuously monitors real-world emissions is the only effective remedy.
Lothar Geilen, CEO of Opus Inspection, says he is encouraged that officials in North America and Europe have acknowledged that greater independent testing of vehicles is necessary. In the US, the EPA says it will add on-road testing to its regimen; German Environment Minister Barbara Hendricks says regulations have to be demanding enough that diesel vehicles truly cause less pollution, with independent authorities' conducting controls to verify it; and UK Transport Secretary Patrick McLoughlin says the UK leads Europe in calling for real-world testing.
Such testing, Geilen says, should indeed focus on the road, where vehicles pollute and cannot easily cheat; however, he also indicates the true scope of the problem is not widely understood because few people have closely analysed large databases of on-road emissions gathered by roadside sensors. His company employs roadside sensors gathering millions of measurements each year under contract to various state governments. The engineers and scientists working for Opus have applied sophisticated, big data analysis techniques to this information and found some very troubling results. Not only does their analysis flag VW diesel engine emissions, it also flags many other makes and models, including the U.S. manufacturers of gasoline-powered vehicles. The data even suggest that emissions violations vary considerably for certain models from one model year to the next – giving even more credence to the view that continuous on-road monitoring is needed.
Geilen cautioned, "The data is not, by itself, an indictment of these vehicles. Confirmatory testing is needed under more controlled conditions. Nonetheless, the old cliché 'where there's smoke' should be taken seriously in this case."
Geilen says this doesn’t mean that most vehicle manufacturers are cheating. "The public needs to understand there is a gray scale of intent from outright cheating to innocent performance anomalies discoverable only through real-world performance measurement.
Unfortunately, our laws and regulations have focused too heavily on the laboratory emissions testing required for a vehicle to gain government-type approval and not enough on real-world performance. We are incentivising manufacturers to calibrate performance to the laboratory test. Unless a truly statistically significant sampling of real-world performance is made, manufacturers don't really know how their vehicles are performing once they're on the road. Our data suggest that sampling requires thousands of measurements on each model year – year in and year out throughout the model's life cycle."
Geilen goes on to say two complementary and distinctly different methods of testing vehicles on the road should be used: one attaches a Portable Emissions Measurement System (PEMS) to a vehicle, while the other incorporates roadside sensors commonly referred to as Remote Sensing Devices (RSD) or simply remote sensing.
"The two methods of on-road testing, PEMS and Remote Sensing, are complementary, and both need to be stepped up," Geilen said. "Because the first method is limited by its coverage of individual vehicles (one at a time) and associated vehicle recruitment challenges, it could benefit from an additional, real-world, in-use surveillance program that measures thousands of vehicles per day as they drive by, thereby identifying specific vehicle makes/models exhibiting abnormal emissions behaviour. These flagged makes/models would then be good candidates for extensive PEMS testing. The second method, remote sensing, is exactly the kind of surveillance program that can help the PEMS focus its detailed investigations."
For example, Geilen says that a remote sensing program in the Denver metropolitan area captures 7 million measurements per year. In 2014, over 4,000 distinct models identified by fuel, type, weight class, make and engine, representing 93 percent of all vehicles active in the region, were measured at least 100 times. Just over 1,000 more popular models representing 65 percent of active vehicles were measured over 1,000 times. These data provide striking statistical evidence of how the emissions control systems of all vehicle models are performing in the real world from their first year of sale until their retirement.
"The offending VW diesel vehicles could have been identified in the first year of their sale instead of five years later, had the data analysis goals of our contract included monitoring of specific make/model performance as opposed to focusing on individual vehicle performance. Even lacking that data analysis focus, our scientists and customer associates using remote sensing discovered the excess VW emissions via remote sensing data and presented it at an industry conference in spring of this year, well before its public unveiling," Geilen said.
"Extended to other metropolitan areas and including a mandate to search for possible noncompliant makes/models, remote sensing would provide valuable emissions data on vehicle models operating under many conditions."
By comparing the emissions of a model to those of similar models from other manufacturers and to emissions standards, Geilen says obvious patterns emerge. “Because a relatively low number of remote sensing devices can efficiently measure millions of vehicles under actual driving conditions, Remote sensing databases offer the advantage of identifying emissions behaviour of vehicle model groups over time, something PEMS and the current in-use compliance regulations cannot achieve. By adding a large-scale, continuous remote sensing monitoring program, we could track emissions levels of specific model groups in statistically significant numbers over years of in-use service, thereby illuminating the effects of wear as well as the effects of emissions-related calibration changes made by manufacturers to vehicle software. Feedback from such information would lead to earlier repairs and improved vehicle designs.”
Geilen says expanding programs such as the effective remote sensing operation in Colorado is vital to ensuring low-emissions performance and avoidance of the premature deaths caused by motor vehicle air pollutants.
Lothar Geilen, CEO of Opus Inspection, says he is encouraged that officials in North America and Europe have acknowledged that greater independent testing of vehicles is necessary. In the US, the EPA says it will add on-road testing to its regimen; German Environment Minister Barbara Hendricks says regulations have to be demanding enough that diesel vehicles truly cause less pollution, with independent authorities' conducting controls to verify it; and UK Transport Secretary Patrick McLoughlin says the UK leads Europe in calling for real-world testing.
Such testing, Geilen says, should indeed focus on the road, where vehicles pollute and cannot easily cheat; however, he also indicates the true scope of the problem is not widely understood because few people have closely analysed large databases of on-road emissions gathered by roadside sensors. His company employs roadside sensors gathering millions of measurements each year under contract to various state governments. The engineers and scientists working for Opus have applied sophisticated, big data analysis techniques to this information and found some very troubling results. Not only does their analysis flag VW diesel engine emissions, it also flags many other makes and models, including the U.S. manufacturers of gasoline-powered vehicles. The data even suggest that emissions violations vary considerably for certain models from one model year to the next – giving even more credence to the view that continuous on-road monitoring is needed.
Geilen cautioned, "The data is not, by itself, an indictment of these vehicles. Confirmatory testing is needed under more controlled conditions. Nonetheless, the old cliché 'where there's smoke' should be taken seriously in this case."
Geilen says this doesn’t mean that most vehicle manufacturers are cheating. "The public needs to understand there is a gray scale of intent from outright cheating to innocent performance anomalies discoverable only through real-world performance measurement.
Unfortunately, our laws and regulations have focused too heavily on the laboratory emissions testing required for a vehicle to gain government-type approval and not enough on real-world performance. We are incentivising manufacturers to calibrate performance to the laboratory test. Unless a truly statistically significant sampling of real-world performance is made, manufacturers don't really know how their vehicles are performing once they're on the road. Our data suggest that sampling requires thousands of measurements on each model year – year in and year out throughout the model's life cycle."
Geilen goes on to say two complementary and distinctly different methods of testing vehicles on the road should be used: one attaches a Portable Emissions Measurement System (PEMS) to a vehicle, while the other incorporates roadside sensors commonly referred to as Remote Sensing Devices (RSD) or simply remote sensing.
"The two methods of on-road testing, PEMS and Remote Sensing, are complementary, and both need to be stepped up," Geilen said. "Because the first method is limited by its coverage of individual vehicles (one at a time) and associated vehicle recruitment challenges, it could benefit from an additional, real-world, in-use surveillance program that measures thousands of vehicles per day as they drive by, thereby identifying specific vehicle makes/models exhibiting abnormal emissions behaviour. These flagged makes/models would then be good candidates for extensive PEMS testing. The second method, remote sensing, is exactly the kind of surveillance program that can help the PEMS focus its detailed investigations."
For example, Geilen says that a remote sensing program in the Denver metropolitan area captures 7 million measurements per year. In 2014, over 4,000 distinct models identified by fuel, type, weight class, make and engine, representing 93 percent of all vehicles active in the region, were measured at least 100 times. Just over 1,000 more popular models representing 65 percent of active vehicles were measured over 1,000 times. These data provide striking statistical evidence of how the emissions control systems of all vehicle models are performing in the real world from their first year of sale until their retirement.
"The offending VW diesel vehicles could have been identified in the first year of their sale instead of five years later, had the data analysis goals of our contract included monitoring of specific make/model performance as opposed to focusing on individual vehicle performance. Even lacking that data analysis focus, our scientists and customer associates using remote sensing discovered the excess VW emissions via remote sensing data and presented it at an industry conference in spring of this year, well before its public unveiling," Geilen said.
"Extended to other metropolitan areas and including a mandate to search for possible noncompliant makes/models, remote sensing would provide valuable emissions data on vehicle models operating under many conditions."
By comparing the emissions of a model to those of similar models from other manufacturers and to emissions standards, Geilen says obvious patterns emerge. “Because a relatively low number of remote sensing devices can efficiently measure millions of vehicles under actual driving conditions, Remote sensing databases offer the advantage of identifying emissions behaviour of vehicle model groups over time, something PEMS and the current in-use compliance regulations cannot achieve. By adding a large-scale, continuous remote sensing monitoring program, we could track emissions levels of specific model groups in statistically significant numbers over years of in-use service, thereby illuminating the effects of wear as well as the effects of emissions-related calibration changes made by manufacturers to vehicle software. Feedback from such information would lead to earlier repairs and improved vehicle designs.”
Geilen says expanding programs such as the effective remote sensing operation in Colorado is vital to ensuring low-emissions performance and avoidance of the premature deaths caused by motor vehicle air pollutants.
