When a system malfunctions, somebody needs to fix it. The systems we deal with at McKenney’s often have thousands of variables with one, or ten, of those variables being the issue. So how do you solve a random problem when everything is all twisted together? The following is a description of the steps we used to solve a recent problem.
Step 1: Define the Problem
One of our customers was having an issue with their chiller shutting off at odd times when there was plenty of load. The staff was required to manually reset the chiller. Due to the stress of unnecessary cycling, the serviceable lifetime of the chiller was becoming shorter. When defining the problem, it was important to not prematurely simplify the problem. It is easy to think that since the chiller is turning off, there must be an issue with the chiller.
Step 2: What Tools Are Available?
To solve this problem, we used trend data from the building control system to monitor changes of various points from all crucial devices in the system. The sensors find data every 5 seconds, however a data point is only saved in the database every 5 minutes. This 5 second resolution was important to take into account when diagnosing the issue. It required taking the time to watch the changes in the system every 5 seconds, and take screenshots to capture the issue.
Through the 5 second data, we noticed that the flow was drastically changing every 30 seconds on one of the large air handlers. This would never have been seen in the 5 minute trend data.
Step 3: Evaluate All the Options (and test them)
We looked at the pumps, differential pressures, manual parts, valves, temperature sensors, etc. The list seemed endless, but we were finding clues to the solution.
Step 4: Stay Dedicated
It is important to keep your teammates motivated when there doesn’t seem to be an obvious solution to a problem. It can sometimes make people anxious when it takes too long to find a solution.
Step 5: Results
It boiled down not to one, but multiple issues, that were nowhere close to the chiller. There was an air handler balance valve open at 100% which was overflowing the coil, causing cold water to be sent back to the chiller. There was also a manually changed algorithm (that was done incorrectly) in the controls for an air handler. The algorithm is supposed to slow down valve reaction times so that the flow changes more slowly. Lastly, we realized the older screw chiller couldn’t handle the flow swings it was experiencing, which was causing it to shut down.
Fixing the first two problems helped slow down the water flow changes and bring warmer water back to the chiller. Even though it was an older chiller, it was now able to handle the load changes more gracefully, and stay online.
In Summary…
NB: 66% of our chiller problem was solved by looking at another part of the system. For that reason, it’s important to remember to keep an open mind when solving problems, explore all your options, and have fun.
Have a question for our experts? Leave your comment below and check out our website for more information.
