Wednesday, September 23, 2015

MPGe fraud

This issue has been bothering for a long time, even before I got an EV. In fact, I tried to establish true out of pocket cost MPGe pretty early in my blog. When gas cars list MPG, the consumer typically expects it to correlate to amount of money they pay to drive it with current $ per gallon at local gas station. With MPGe, it has very little to do with the amount of money one pays to drive the EV. When you use similar terminology that most people are familiar that relate to money and change the meaning completely to something else, that is fraud. It's like VW claiming their emission numbers were only meant for testing, but not for actual driving. Actual MPGe based on cost is described in the table in my previous blog.

I had long winded discussion in SparkEV forum to get to the bottom of this matter. What we found, with help from those in the forum, was that EPA uses "well to wheel" for MPG of gas cars while "tank to wheel" for EV. Simply put, MPGe assumes (somewhat correct under ideal conditions) that a gallon of gas has 33.7kWh of energy, and using that as a metric. For example, if a car uses 33.7 kWh of energy to drive 100 miles, that would be 100 MPGe. More popular example might be 16.85 kWh to drive 60 miles to yield 120 MPGe. Since SparkEV gets 119 MPGe EPA rating, this would be close to what SparkEV is doing.

But this is problematic. How does one relate EPA MPGe to out of pocket cost to drive an EV? You can't. It is related to efficiency of the car, but that means little to most people; if it costs $1,000,000 to drive 1 mile while it's 99.99% efficient is meaningless. But costing $1 to drive 1,000,000 miles while it's 1% or 0.000001% efficient (maybe Mr. Fusion?) is meaningful.

Far better number would've been miles/kWh (or kWh per 100 miles). But EPA being a bureaucratic nightmare government agency, they conducted a poll that showed miles/kWh "confused" the consumers when such metric was discussed. Instead, poll indicated MPG was better understood. Instead of applying MPGe in the sense that consumers understand (ie, $ to drive X miles) and can test for themselves, they applied it as completely nonsensical fashion. But since it relates to energy efficiency, EV drivers should be able to replicate EPA MPGe numbers at home, right? WRONG!

I analyze various EV to see what I can find. Only fast charge capable cars are considered as I consider non fast charge EV as toys. Given the battery capacity and the range, the formula to compute EPA MPGe should be

EPA MPGe = R / B * 33.7 * e

where R is range in miles, B is battery capacity in kWh, 33.7 is kWh per gallon of gasoline, e is EV efficiency (must be less than 100%). As you can see from the table, it gives some radical numbers for efficiency, far more than 90% in many cases. Knowing that motors and controllers are 90% or less efficient, these are probably not true.

An interesting observation is VW eGolf. The number is suspiciously close to EPA number. Did they actually test the car or did they simply take the range and battery capacity and do the math? It's VW, guys who faked EPA testing with the "clean" diesel, so questioning such conincidental number is warranted.

Keen observer will note that modern EV do not use 100% of their battery capacity for stated range. Then how much do they use? This information is not readily available; it seems to be a "secret sauce" for many EV. For Volt, it was speculated to be about 50%, for SparkEV 80%, but the exact number is hard to come by. Without knowing this number, we can only guess, and I use 85% as guesstimate.

We seem to get more reasonable numbers, but there is a problem. SparkEV (119 MPGe) is only 71% efficient while VW eGolf (116 MPGe) is 85% efficient. Given that SparkEV has one of the highest head room, even more than Tesla, one would expect more air resistance by being taller (more frontal area) and may be lower efficiency than others. But to show VW eGolf to be more efficient than any other EV is questionable. Once again, remember the fraud VW committed with their "clean" diesel.

Basically, what it shows is that without knowing the actual battery used for given range, one cannot replicate EPA MPGe number. Since actual battery capacity for range is not published, EPA MPGe is a meaningless number (it's a religion, not science). One may argue that this number can be used to compare one EV to another, but if cannot be replicated by the consumer, it's a dubious comparison at best.

mi/kWh to MPGe

Here's another angle one can take. Take the EPA MPGe number and convert to mi/kWh.

mi/kWh = MPGe / 33.7kWh/gal

Then multiply that number by full battery capacity to figure out the range. Since not all battery is used, resulting range should be more than EPA range estimate. But that's not the case. If one uses EPA MPGe number, one gets far less range than the EPA estimated range (except eGolf). Inconsistency? You betcha!

Again, far better metric would be mi/kWh. But that number varies with different charging methods as shown in my previous blog posts (80% efficient with L1, 90%+ efficient for DCFC). But taking some conservative figure as described in my SparkEV efficiency blog post, SparkEV gets 4 mi/kWh. This is taking into consideration L1 charging loss (80% efficient), which is total energy from outlet to wheels not just energy from battery to wheels.

This corresponds to

4 mi/kWh * 33.7kWh/gal = 134.8 MPGe

If one considers battery to wheels per EPA, then the figure is 20% higher, or 161.76 MPGe.

The real-world MPGe for SparkEV is far more than EPA MPGe even from "well to wheels". Then how the heck did EPA get 119 MPGe for SparkEV? Is it time to say "Who's John Galt?"

Honest MPGe

To solve this confusion, better would be to qualify MPGe with various subscripts (or superscripts). For example, MPGe$ would correspond to equivalent $ to some particular $/gal of gas and $/kWh of energy. This is what the consumer is expecting to see with MPGe. It won't be a number, but a table like I show in my blog post. If it's a number, it must be qualified with $/gal AND $/kWh. For most consumers, this is the only MPGe that matters. Once again, here's the link to MPGe table.

Another would be MPGeEGG which would correspond to equivalent energy in gallon of gasoline. This is probably what EPA is getting, although it's impossible to compute at this time with given information. It's easy to remember, too; its a meaningless number of EGGheads.

Another would be MPGeEFF which would correspond to equivalent energy in fossil fuel used compared to gas car's gallon of gas use. Electricity for most parts come from fossil fuel, primarily natural gas and coal. Since electricity generation is far more efficient than gas cars, especially the combined cycle generators, one would need less fossil fuel to drive EV. This becomes especially important when non-fossil fuel is used to generate electricity. It's also easy to remember, EFF as in F as in we're fracked when we have to import oil. I will explore this in more detail in future blog post.

Another would be MPGeP which would correspond to equivalent in pollution compared to gas car's gallon of gas use. If coal is used for this calculation, EV would be worse (or awful if one considers coal ash). Radioactive pollution would be bad, too. This is more nebulous merit which must assign somewhat subjective weight to pollution from each source of energy. It may be discussed further in future blog post, but it'll be messy.


In science, an experiment must be replicated by independent sources and reasonably close result obtained. Asking people to believe in something without such scrutiny is religion. As an EV advocate and science advocate, EV should not be a religion, although many people seem to treat it this way (ie, EV is NOT zero pollution, despite what Nissan Leaf prints on its doors). That means MPGe number must be something meaningful and derived from experiments from independent parties, not a  number handed down by the EPA. I hope true out of pocket cost MPGe$ as table will be printed on windows of EV and educate the public rather than continuing with the religion that it is now.

No comments:

Post a Comment