What happens when the laboratory is no longer the gold standard?
PEMS (Portable Real World Emissions Systems) is the route to bring the “real-world” back to testing. It works by simply measuring emissions outside of laboratory conditions and in the real world instead.
Why the fuss? Real world emissions are messy. Laboratory conditions exist solely to reduce variability and increase repeatability. In other words, laboratory conditions are just another type of tightly controlled model.
On the other hand, lab measurements of pollutants and fuel-efficiency on standard drive cycle will never match what is actually happening in those cars when consumers drive them. Ironically, a few recent scandals of bad-behavior on the part of OEMs have cemented the need for PEMs in the regulatory landscape.
Testing Scandals or Measurement Scandals?
You may have heard of a little thing called the “Volkswagen Testing Scandal”. If regulatory bodies test for pollutants on a very small set of boundary conditions, they will get entire fleets of vehicles optimized, or over optimized for those conditions. The New York Times recently reported that one of Volkswagen’s lawyers is partially justifying the carmaker’s actions saying that everyone is aware that all vehicles exceed emissions standards in street use. But the most interesting part of the NYT’s recent story was that the lack of PEMs available to the general public, watchdog groups and even research labs only ten years ago may have convinced some at the automaker that they had a small chance of getting caught.
Bring in the PEMS.
Yes, it is a fact that real driving emissions are substantially higher than in the certification tests. Transport and Environment, an EU environmental group, put out a report titled: Not a surprise in the industry, but outside of it, it’s a bit of a shock. So, with PEMs on the ascendant, are laboratory dynos and simulation out?
So what do we do with the laboratories (and simulation)?
Why would OEMs spend time modeling, or testing, in static boundary conditions when it’s the real world that will be tested in the end? The very nature of RDE is that it will be “real-world” and non-reproducible. After all you don’t need to be a rocket scientist (or mechanical engineer ) to see that Real ? Simulated.
At an recent SAE World Congress session on PEMS, many Tier 1s and OEMs expressed frustration with the PEMS process and the uncertainty in testing industry. Horiba outlined many of the issues with using PEMS. According to them, in Europe you have to use the worst case fuel because there is no allowance for fuel type in the testing. Two supposedly identical tests in real driving conditions will vary from one test to another in speed, ambient temperature, traffic and driver aggressiveness making hard for engineers to quantify individual variables. In addition, mounting the equipment is difficult and sometimes not street legal. Some of PEMs are large trailer towed equipment or mounted in various creatively mounted configurations that the local traffic police don’t always understand.
PEMs is a disruptive technology for testing and the testing industry. Panelist Charles Roberts from SWRI showed slides of their impressive facility. At SWRI alone there are 20 test cells and dynos. If everyone moves to real world testing, does this signal a sea-change for the industry? In our opinion, laboratory conditions will always be needed to run standard cycles and PEMs is another piece of the puzzle that the regulators and manufacturers will need to take into account.
Simulation and Laboratories – an increased need, not a decreased one
In the real-world, everything changes all the time. And with this, mathematical models have a larger role to play. One thing you can do in simulation is to change one variable at a time, keeping all other boundary conditions the same, and quantify the effect of that single variable on the emissions. There will always be argument and the nay-sayers that the emissions are impossible to calculate. But simulation can give you a pretty good indication of change and the general trend. Simulation will work in conjunction with PEMs,to identify and isolate every variable, providing guidance on the quantify of each variable’s effect. That quantify will give designers ideas which variables to change by how much and which ones will work together. Eg. Do they cancel or they multiply each other and go 50? So essentially, simulation will run your whole gamut of options, indicate which scenarios are totally out of the running, educate the designers on the interactions, and boil down the large matrix of tests into a manageable set of tests which show good promise. And PEMS takes it from there.