Cautionary Tales of Building Data Analysis
Published by Greg Swiss
Most of the buildings we work with have building automation systems (BAS) and most sizeable new commercial buildings do as well. A BAS system often allows you to see what’s happening in real time or in the past with your building. This can lead to a lot of data; a standard single air handler could have 15 different points. If you start recording (aka trending) the data it becomes a lot to sift through. How do you make sense of all of it and translate the data to what’s actually happening?
A key part of that is not being fooled by what you see and knowing from experience what is most likely happening. I’ll go through a couple common traps.
Know that if you are looking at historical data it’s probably going to look messy. With historical data you are only taking a snapshot of the data every 5 minutes or 15 minutes and a lot can happen in-between (aka aliasing). To help sniff out what’s going on, start with looking at a week’s or day’s worth of data at a time rather than a month.
First off, what are some good procedures for setting up trends? Intervals of 5 to 15 minutes often make the most sense for points – any longer and you may start to miss out on what’s happening, any smaller and you could run into data storage issues and not gain much more detail. For any points that you want to trend, it’s good practice to trend the current value and setpoint. For example, take a look at the following graph showing the supply air temperature for an air handler.
We can see the system turn on at 7 am supplying 55°F, but then at 4 pm all the sudden we jump up to 58°F. We know we’ve seen this on other days too, adding further speculation. Is this the sticky outdoor air damper we’ve had before or are rooms calling for warmer supply air? If we were smart enough we would have trended the setpoint too and we could overlay the two and not be scratching our heads.
Resolution! Our operator did adjust the setpoint, and sometimes discoveries like this lead to additional answers. We deployed room temperature sensors in these zones and might have an idea why many of those rooms were consistently above setpoint that next week… nobody changed back the supply air setpoint to 55°F. Seems like good evidence for implementing a supply air reset, doesn’t it?
Outside air damper operation is often an investigation point for analysis. The outside air dampers have to operate in the “Goldilocks” zone: not too much and not too little. Unfortunately, just the right amount varies throughout each day and season. A good tool in the tool chest for analysis is the % OA Equation.
RAT = Return Air Temperature
MAT = Mixed Air Temperature
OAT = Outdoor Air Temperature
RAT and OAT are easy to get and record – no mysteries there. MAT is the troublemaker. Getting the MAT seems easy since most air handlers are already reporting one. However, there’s a big difference between what the MAT sensor reports and what the MAT actually is. It comes down to mixing. The mixed air section almost certainly has stratification, meaning if you measure the temperature in 10 different areas you will get 10 different results. Sometimes the MAT sensor is in a good location and gets an accurate reading but if the economizer just opened up, it could skew your reading.
What’s the solution then? First off, take the mixed air temperature with a grain of salt. If you are getting %OA that makes sense there’s a good chance you’re in luck. Inspect the mixed air section of the air handler and ask how the flow through would affect the reading. I like using a thermometer/anemometer combo that reads air temperature and flow. They often have a telescoping sensor letting you read across a large area.
If proper mixing seems like an issue consider installing a baffle to help promote proper mixing. Ensure that the baffle doesn’t create too much pressure drop by ensuring the velocity through them doesn’t exceed 800-1,000 fpm. You can see below how the MA sensor on the left will be seeing a temperature more biased towards the outdoor air. With the baffle, the airstreams are forced to blend.
Our buildings are telling us what’s happening. We just have to learn how to interpret them.