I'd like to propose a new unit for natural gas energy demand.

This would be a unit of power and could be expressed as:

• therms_per_hour, therm/h

If we step into the wayback machine and look at Forum Post #4: Units cubic feet per hour comes up as well. Since ft³_gas is already an accepted unit for natural gas energy, I think it would be worth adding this as well:

• cubic_feet_per_hour_natural_gas
• ft³_gas/h or cfh_gas

And/Or

• cubic_feet_per_minute_natural_gas
• ft³_gas/min or cfm_gas

We already have cubic_feet_per_hour_natural_gas:

cubic_meters_natural_gas, m³_gas; kg1*m2*sec-2; 37313432.83582089
cubic_feet_natural_gas, ft³_gas; kg1*m2*sec-2; 1086498

In order to add therms_per_hour, we need to know the conversion factor to the normalized unit of power which is the watt.

We have BTU/hr as:

btus_per_hour, BTU/h; kg1*m2*sec-3; 0.292875

So would therm_per_hour be?

therms_per_hour, therm/h; kg1*m2*sec-3; 29280.50878

That would lead to these equations:

1therm/h.to(1W)     =>  29,281 W
1therm/h.to(1kW)    =>  29.281 kW
1therm/h.to(1BTU/h) =>  99,976 BTU/h 

We already have cubic_feet_per_hour_natural_gas.

Maybe I'm missing it, but I can only find cubic_feet_natural_gas and not cubic_feet_per_hour_natural_gas.

It looks like the above definition of 1therm/h doesn't quite line up with the existing definition of a therm in units.txt, we now have:

therm; kg1*m2*sec-2; 1.05506E8

If we use from above

therms_per_hour, therm/h; kg1*m2*sec-3; 29280.50878

and have something running at 1 therm/h for 1 hour, we would end up with 0.9991 therm.

Would it make more sense to base the unit off the existing definition of 1therm? This would lead to:

therms_per_hour, therm/h; kg1*m2*sec-3; 29307.2222

And convert to:

1therm/h.to(1BTU/h) =>  100067 BTU/h 

Maybe I'm missing it, but I can only find cubic_feet_natural_gas and not cubic_feet_per_hour_natural_gas.

You are right - we have energy unit, but not the power unit. Can you propose definition to test?

If you query google for "therm per hour to btu per hour" you get 99,976.129 which is pretty much what my equations worked out.

The Wikipedia article on therm says one therm == 29.3 kilowatt/hrs, so that looks right to me too.

It looks like it depends if base the unit on the existing Therm definition or the existing power units' definitions. Either way, they are really close to the same number. A Therm is defined as 100,000 BTU, I think it would make sense to base the Therm/h on the existing BTU/h.

So if

btus_per_hour, BTU/h; kg1*m2*sec-3; 0.292875

then

therms_per_hour, therm/h; kg1*m2*sec-3; 29287.5

This gives us the following equations:

1therm/h.to(1W)     =>  29,287.5 W
1therm/h.to(1kW)    =>  29.2875 kW
1therm/h.to(1BTU/h) =>  100,000 BTU/h 

This definition and the one Brian first proposed are really, very close. Either will work in most cases.

The design (based on oBIX) has units are defined against SI base units. But really just based on how many digits we take it out. I took your definition and pushed it to the mercurial repo