BEOpt: Cooling Load Hours not Met

asked 2021-08-05 16:00:16 -0600

Forever_Student gravatar image

I have a BEOpt model of a single family home in North Carolina, about 3,200 square feet. BEOpt autosizes the cooling system to be a 2.5 ton system, which comes with over 500 hours of cooling loads not met. A home of this size in NC might have two units, but probably has 5-tons or more, maybe as much as 8 total tons.

I tried manually oversizing the AC system to 6, then 8, tons, with absolutely no change to number of cooling hours with loads not met, only a reduction in cooling energy consumption.

Duct sizing is autosized by BEOpt. The cooling setpoint is 72F and the heating setpoint is 65-68F, but that shouldn't matter anyway because of BEOpts monthly average HVAC system auto-choosing.

What other home parameters might drive the hours of cooling loads not met? How can I address these?

edit retag flag offensive close merge delete


Does the home have a conditioned basement?

shorowit gravatar image shorowit  ( 2021-08-05 16:54:40 -0600 )edit

You might generate hourly output and look at indoor temperatures compared to setpoints. Even though there are many hours of unmet load, it's possible that the temperature excursions are small. It also might show that cooling is deactivated for certain times of the year, due to the automatic heating/cooling seasons built into BEopt.

shorowit gravatar image shorowit  ( 2021-08-05 16:57:02 -0600 )edit

@shorowit, Excellent insight! The home does not have a conditioned basement, but your other insight was spot on. The temperature excursions during the cooling season are very small, a couple here and there and only fractions of a degree. Your insight led me to discover that the hours of "cooling load not met" are hours in non-cooling seasons, November to April. Problem solved!

Forever_Student gravatar image Forever_Student  ( 2021-08-05 17:07:18 -0600 )edit

@shorowit, If you might, I have a follow-up question because you seem to understand the auto seasons. I'm calibrating the model to actual consumption data and AMY. One major challenge stems from BEOpt's auto heating/cooling seasons. October is apparently just warm enough that it triggers cooling in BEOpt in October, which causes this consumption spike in October in the model that's not represented in the actual consumption. Is there a way to "force" BEOpt to not cool in October? Or to only cool for the first week of October?

Forever_Student gravatar image Forever_Student  ( 2021-08-05 17:10:22 -0600 )edit

I second @shorowit's response to check if the temperature excursions are small.

sashadf1 gravatar image sashadf1  ( 2021-08-05 17:45:18 -0600 )edit