Bench Talk

I recently commented to a colleague that engineers either learn by “letting the smoke out” of something (breaking it) or they can learn vicariously through another’s mistakes, if the mistake is catastrophic enough. I don’t have a lot of these, but I have noticed that engineering horror stories are effective training tools and often funny. 

When I was an applications engineer for automated building control systems, I heard a story about freezing cold water coils in a large air handling unit. Evidently, and this was one of those stories that travelled in whispers, the engineering manager had at one time had a system running all night long in a test procedure. Running water doesn’t freeze. (Except in Russia when you throw it out your apartment window in sub-zero temps, but technically the water itself isn’t moving, although gravity is moving it as a whole….exploration of this is another blog.) 

In a large building with an evaporative cooling system, a closed loop system passes water through a chiller, and the chiller in turn rejects heat via a heat exchanger that is connected to an open water system in a cooling tower (usually mounted on the roof.) On the other side of the chiller’s heat exchanger is a closed chilled water loop that goes through a coiled portion of piping and sits inside a duct, called a “coil.” Air flows over coils filled with flowing chilled water and heat is exchanged from the air to the coil. The water flows back through the chiller in a closed pipe system via a pump to get chilled again. Some systems use chemical coolants, but larger systems tend to use treated water in a closed loop. The essential point is that there is a pump pushing water through this closed loop that hits the air duct system and returns back to the chiller to start all over again, in a flow. The type of electrical engineering required for a building automation system is macro…not the small components in embedded hardware that you find on PCBs, but made up of temperature sensors with averaging elements (longer than the installer was tall) that you would criss-cross across the duct. I used to wire all of these things up to controllers on a networked controller system, with large electromechanical devices like air temperature sensors, limit switches for sensing valve or damper operation, and current sensing relays for determining run status of a pump. 

Our engineering manager had the system running in test mode with the outside air dampers fully open on a cold night. The temperature inside the duct at the chilled water coil was below 32 by a smidgen. For some reason he shut off the pump (maybe he wanted to test the operation of the current status relay.) The water circulation through the coils stopped, and the water inside a coil froze, expanded, and burst the coil inside the ducts. Water started flowing and freezing, but mostly flowing in large volumes, out of this no-longer-closed system. And it wasn’t just one coil that froze: every coil in the system that was fed by that pump froze. I learned vicariously not to shut down water flow in a coil that’s exposed to freezing temperatures. It’s strange how these things stick in your mind. 

Do you have an engineering horror story to share?

Join the Halloween fun and share your favorite engineering horror story in the Comments below. Setting up a Discus account is quick and easy!