Heating data simulation analysis

I wrote some software that will look at the historical hourly average temperature over various years and simulate energy consumption for heating. It uses the ‘detailed heating data’ sections from the manufacturer spec sheets for various models for the btu output and kwh consumption. Most everything on the list is a 2 stage pump, but I just simulate them as if they’re only ever running in the high stage to avoid failed assumptions as to the logic each model uses to transition stages.

The resulting data is rather interesting. And I think it raised a couple questions:

1. The Goodman Deluxe (D) series was dropped in 2017 and only the G and A series remain today. Could the current 18 seer goodman really be 35% less efficient than the amana?
2. Somehow the 16 seer GSZ always performs considerably better than 18 seer GSZ? Even the 14 seer performs better, but admittedly that is partially a product of the 14 seer being single stage and me only allowing the 18 seer to run in its high stage.
3. This begs the ultimate question.. Do we feel the heating data table that manufacturers provide us is accurate? If so, then it would appear that at least daikin may be gaming the system and tweaking their models to perform well in the standarized seer and hspf testing at the sacrifice of general overall practical efficiency(aka 16 seer is actually better than 18 seer in cold temperatures).

As a homeowner, you should ask your questions in this “AOP Residential HVAC” forum.

Please take note of this new location as the “Moved” icon in the other forum will only last a day.

Originally Posted by btarb24

I wrote some software that will look at the historical hourly average temperature over various years and simulate energy consumption for heating. It uses the ‘detailed heating data’ sections from the manufacturer spec sheets for various models for the btu output and kwh consumption. Most everything on the list is a 2 stage pump, but I just simulate them as if they’re only ever running in the high stage to avoid failed assumptions as to the logic each model uses to transition stages.

The resulting data is rather interesting. And I think it raised a couple questions:

1. The Goodman Deluxe (D) series was dropped in 2017 and only the G and A series remain today. Could the current 18 seer goodman really be 35% less efficient than the amana?
2. Somehow the 16 seer GSZ always performs considerably better than 18 seer GSZ? Even the 14 seer performs better, but admittedly that is partially a product of the 14 seer being single stage and me only allowing the 18 seer to run in its high stage.
3. This begs the ultimate question.. Do we feel the heating data table that manufacturers provide us is accurate? If so, then it would appear that at least daikin may be gaming the system and tweaking their models to perform well in the standarized seer and hspf testing at the sacrifice of general overall practical efficiency(aka 16 seer is actually better than 18 seer in cold temperatures).

Somehow that claim is flawed!

I think there is probably something (or a number of somethings) being missed by your software.

While it’s certainly possible that there’s a bug, I have somewhat strong doubts. One clear limitation is that it’s currently only running the two stage units in their highest stage. Though, that will only overscale the numbers, their relative basis will remain the same so the % of deviation across models should be accurate.

It can also be represented just by looking at the COP at each temp for the models. COP is the measure of efficiency for how much input power was consumed to generate the output heat. It’s derived by (btuH*WattsPerBtu) / watts. wattsPerBtu being a .293 constant.

Here’s just the two 18 seer models and a snippet of the amana’s spec sheet for confirmation. The amana specs are clearly significantly better – at least as reported by the manufacturer in the heating details sections.

Originally Posted by btarb24

While it’s certainly possible that there’s a bug, I have somewhat strong doubts. One clear limitation is that it’s currently only running the two stage units in their highest stage. Though, that will only overscale the numbers, their relative basis will remain the same so the % of deviation across models should be accurate.

It can also be represented just by looking at the COP at each temp for the models. COP is the measure of efficiency for how much input power was consumed to generate the output heat. It’s derived by (btuH*WattsPerBtu) / watts. wattsPerBtu being a .293 constant.

Here’s just the two 18 seer models and a snippet of the amana’s spec sheet for confirmation. The amana specs are clearly significantly better – at least as reported by the manufacturer in the heating details sections.

I’ll trust the manufactures engineers.

And that’s from someone that hates engineers!!!

I trust them as well, which is why I’m pointing out that they are the ones that are stating the grand efficiency difference between the two units.

Those bottom two tables (white) are from their own spec sheets. They were lazy and didn’t put COP values for the goodman, but you can see that my goodman COP numbers are accurate by comparing my amana data to the amana team’s COP spec. Minus some minimal rounding differences they are the same values. My goodman & amana COPs are generated with the same formula and based on their btu & kw data.

I was somewhat hoping for a nice intellectual discussion over the data .. not a grand defensive against blind disregard :/ Curious.. would you trust my posting more if i were an HVAC tech? If so then I would question why you would trust an hvac tech more for data analytics than a seasoned lead software engineer. We can all have skills outside of our primary profession.

Originally Posted by btarb24

I trust them as well, which is why I’m pointing out that they are the ones that are stating the grand efficiency difference between the two units.

Those bottom two tables (white) are from their own spec sheets. They were lazy and didn’t put COP values for the goodman, but you can see that my goodman COP numbers are accurate by comparing my amana data to the amana team’s COP spec. Minus some minimal rounding differences they are the same values. My goodman & amana COPs are generated with the same formula and based on their btu & kw data.

I was somewhat hoping for a nice intellectual discussion over the data .. not a grand defensive against blind disregard :/ Curious.. would you trust my posting more if i were an HVAC tech? If so then I would question why you would trust an hvac tech more for data analytics than a seasoned lead software engineer. We can all have skills outside of our primary profession.

I’m really curious but I will say I have noticed that slightly higher SEER units sometimes have lower cop ratings. I do not view SEER (or the new SEER2) as a valid rating system and I know that’s going to be a very unpopular opinion.

I agree that neither SEER nor HSPF seem like valid metrics for comparison. They really only apply to rather specific environment situations and do not look at the broad spectrum of usage. However, SEER is what’s displayed prominently in all the marketing, so any optimizations within that situational band would likely be the engineering directive. That said, I imagine in many cases, the optimizations for seer marketing would also apply benefits to the broad COP. I do have concerns that this approach would break down once you reach the upper limits of what the scroll technology can handle, after that point i think there’s a possibility that any seer optimizations may regrettably cause sacrifices in other situations.

I also ran some calculations for 2 stage units. Effectively, it always uses stage1 unless the stage cannot supply sufficient btu to cover the heat loss at that specific temperature. After that, stage2 is used. This wont be wholly accurate since it does increase the granularity up to the entire hour rather than a more frequent stage change that could be utilized in real life.

It’s also very interesting that the amana stage1 & 2 have roughly the same efficiency. Though, the goodman has immensely better stage1 efficiency than it has in stage2. Really makes me question if their data collection during stage 2 was somehow flawed.

I would question their data as well as the units are nearly identical….

And remember too, they are testing in a lab and you and I can almost never expect to get the ratings they claim.

As you pointed out seer is more a marketing scam the industry has trained the people to put a lot of weight in.

What about running an Amana 20 seer variable capacity in your numbers.

Sent from my iPhone using Tapatalk