Friday, January 16, 2009

The Misguided Role of Benchmarking


The Chairman of EVO asked me to post his recent thoughts on energy efficiency benchmarking.  I think someone struck a nerve.

From John Cowan, Chairman EVO

All too often these days I hear the uninitiated trying to sound as though they understand energy efficiency.

They say its all about your energy benchmark, or energy intensity:

· kWh per tone of product or widget produced, or

· kWh per square meter of building or per degree day

or some such index. They regard the index as God because it can be used to compare facilities and to track any particular facility from one period of time to the next. I say, well only under very limited conditions. Limits that probably make it useless.

Let’s look at them one at a time:

Comparing facility benchmarks or indexes. Is it fair to compare an old plant with a new plant? That just shows there is a difference in energy intensity based on vintage. And what about the mix of products, or the production volume vs plant capacity, or the type of tenants, or the number/size of windows, or ........ in one facility versus another? Can you compare benchmarks from different plants? Until you have whittled the database for comparison down to only those of comparable design and use, no valid benchmark comparison can be made. Usually by the time the database is so whittled, the comparison is not statistically significant.

Comparing a facility’s benchmark over time. A benchmark is an indication of average energy use. If production increases, average energy use usually drops, due to the constant loads that are unrelated to production. So can you use the energy index from one production level to adjust to another level? Look at some sample plant energy and production data below and the plot of it.

ScreenHunter_01 Jan. 16 11.31

Production Energy Index
Widgets Units  
1 100 100
2 110 55
3 120 40
4 130 33
5 140 28
6 150 25
7 160 23
8 170 21
9 180 20
10 190 19

Most plants (or buildings) have a fixed component in their energy use, as shown see above. At low production rates the index is 30-40+ units/widget. At plant capacity it is 19, seemingly more efficient, based on the index. But its not using less energy, or necessarily saving anything. Let’s figure out savings for a scenario:

Suppose in the first year (baseline) you determined that the plant used 150 units of energy and operating at 6 widgets per period. It had an index of 25. Then in year 2 it operated at its 10 widget capacity and used 160 units of energy.

I think it saved 30 units (baseline use at 10 widgets is 190, minus actual year 2 use of 160).

But the benchmark or index approach would say you saved 90 units (baseline period’s 150 units adjusted for the 4 widget production increase using the index (25 x 4) = total baseline of 250, minus actual year 2 use of 160).

A benchmark is usually only valid at the level of production or activity at which it was derived. It is of no use for determining savings at any other production rate.

So what is the role of a benchmark?

John Cowan