Question-and-Answer Resource for the Building Energy Modeling Community
Get started with the Help page
Ask Your Question

Thoughts on Efficiency Derates

asked 2016-06-03 10:50:33 -0500

Adam Hilton's avatar

updated 2017-05-02 12:40:23 -0500

I wanted to see peoples thoughts on aged equipment efficiency, such as a HW boiler or DX cooling coil.

I think that without a doubt it can be agreed that the the efficiency of a piece of equipment at the 10 year mark is going to be less than the efficiency of the equipment when it was first installed. This may be attributed to dirt build up on condenser fins, decreased lubrication of bearings, changes in the thermal properties of materials, or general mechanical wear for example. The only way I'm aware of to determine the actual efficiency would be to perform a laboratory test on the piece of equipment which isn't realistic the majority of the time. This is however, a very important concept to capture in the modeling of an existing building. Without this the replacement of an old unit with a new unit may produce artificially low savings.

I've done a few strolls through the interwebs, but wasn't able to find much, especially with respect to energy modeling. I found this report with respect to residential equipment that I think was fairly concise and relevant (even though it's not applied to commercial buildings). It references an NREL report, but I wasn't able to find it as the links that were given were broken. I like the concept of simply factoring the original EER by the square of what's effectively an 'aging' factor directly related to how well the unit was maintained.

Anyway, just wanted to see if anyone was willing to share how they capture aged equipment in their models.

edit retag flag offensive close merge delete

1 Answer

Sort by ยป oldest newest most voted

answered 2016-06-06 05:17:49 -0500

updated 2016-06-06 05:35:41 -0500

Proper field work for existing buildings should necessarily include thorough testing. For example, for boilers you should really do a combustion test at different part loads (unless the burner is On/Off only) to capture the burner efficiency across the modulating range. The idea is to try and nail as many parameters as you can: there will always be some that can't measure or even estimate with more than a broad qualitative judgment, for example infiltration.

I'd also be pretty skeptical about trying to derive general rules of efficiency versus aging. Something old but very well maintained might be working just like on the first day, and generally a system is a combination of individual components. A 50 year-old cast-iron scotch-marine boiler on which you just recently replaced the burner and the jacket insulation shouldn't be much different than an actual brand new boiler (we didn't get that much better at making giant kettles basically). On the other hand a two year old boiler that hasn't been well maintained - no bleeding nor water treatment leading to water-side scale, soot in the fire tubes, burner not tuned across the entire modulation range, controls overridden - might be a lot worse than planned if not a safety hazard (incomplete combustion => CO).

Coils need cleaning, filters need changing, burners need tuning, everything needs servicing.

Anyways, after proper field work, you will need to take an educated guess for the initial value of some parameters you couldn't measure. Start by trying to assess what the rated efficiency of the unit is (nameplate, or common efficiency ratings at the time, look at older codes for example), then derate that based on whatever information you have such as visual inspection (dirty coil, coil fins bashed in, etc)

After that, these parameters will be refined during the calibration process where you'll try to get your model to match the utility bills.

This basically amounts to a linear regression problem in which you have naturally quite a few parameters (some matter a lot, others much less so, a sensitivity analysis is useful for that).

In many cases you'll use 12 monthly bills as the independent variable for the calibration. If you're familiar with the problem of overfitting, this should really stress my earlier point about "nailing every parameter you can".

With four parameters I can fit an elephant, and with five I can make him wiggle his trunk.

-- John von Neumann

(For fun, see a Python implementation of this quote)

edit flag offensive delete link more


I agree with respect to the furnace, the combustion test is 'easy'. A dx coil/condensing unit or chiller I believe to be a different story as I don't believe this is easily field testable. Without having something to compare to it's hard to look at something and and say "That's probably having a 10% impact on efficiency". As far as adjustments during calibration, you might have 30 RTUs that were installed a few at a time between 1996 and 2009. In this case it would be nice to have a feel for some sort of abstracted derate curve form a well-maintained, maintained, and not maintained units.

Adam Hilton's avatar Adam Hilton  ( 2016-06-06 08:21:27 -0500 )edit

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Training Workshops

Question Tools

1 follower


Asked: 2016-06-03 10:50:33 -0500

Seen: 1,375 times

Last updated: Jun 06 '16