A difficult house to control

[Guess who!]

Hi guys.

A very good friend of myne has this house.

Its located on a hill crest, facing west, looking down on a 20 mile long (2 miles wide) lake. Ho! What a view.

He tells me its a mediteranean design. 4 feet cement basement. Open spaces. West side wall are full of real big windows.

His problem is heating. In winter his electrical heating bill is more than twice has much has myne

The heating is all done through big BUH heaters in the basement. The basement is isolated on all sides except the ceiling which is the floor of the house. The floor heats up and ventillation circulates the floor warmth around the house.

When the sun is up during the day and the heat gets in, the heaters stop. Around 4 o'clock when the sun sets the house starts to be cool. By the time the heaters do there job and the ceramic/cement floor gets hot enough to heat the rest of the house they are freezing in there.

Around midnight it get so hot that they almost have to open windows.

Cycle time is at the least very slow.

We have no other controls but the thermostats in the main floor area.

When its windy (in winter of course) he needs some heat but if clouds get in the way of the sun rays he needs a lot of heat.

We can install a lot of sensors and logics but how would you guys suggest we attack this challenge.


_If it looks like a Cat...

 

[Sliver]

BUH?

Can you give more detail on the BUH heaters.

Its sounds like they are electric space heaters?
Heating the airspace under the house therefore heating the ground floor floor (1 storey?)
I see why there is such a delay in control.
Your system would have to anticipate the heating load required and obviously start heating earlier and stop heating earlier.
Easy to say, not easy to do I think.

I know a fellow with radiant floor heating and he hasn't complained of heating control lag. Perhaps the main delay is in heating the airspace under the house? Can this hot air not be circulated into the upper space rather than rely on the heat transfer through the floor?

 

[Tom Jenkins]

What you need is, I think, called optimized start/stop. It essentially uses outside air temperature to kick in the heaters for a variable time ahead of the internal temperature drop, without waiting for the house to cool down. You are essentially compensating for the big heat sink associated with the floors etc. It would take some playing with, but the pre-start time is going to be longer the colder it is outside.

The HVAC supplier may have something to do this already. If not, an RTD and a micro PLC with a small HMI would be incredibly elegant. You could replace the existing thermostat with another RTD and include night setback functions and so on as well. You could even use decreasing temperature trend in conjunction with the outside temperature to get a dead nuts pre-start.

 

[Sliver]

Pierre?

Did I guess right?

If so and if this house is in Quebec I would assume there is plenty of good maple and oak firewood for sale nearby. Tell your friend to put in a good ole woodstove. I know Hydro-Quebec rates are cheaper than Ontario Hydro (Ontario Generation?) but my electric forced-air furnace main breaker is left open all year. Natural gas is through the roof and heating oil is going to be following it. My dad used to say there were three heats in a stick of wood. One when you cut it, one when you burned it and one when you carried out the ashes. Very economical! Give us the location and we can even organize a wood cutting party. (No beers until the chainsaws are put away!)
If you are not Pierre than completly ignore this post....

Brian

 

[hd_coop]

I think you guessed right ... another forum

 

[akreel]

So he's cheating on us. No wonder he tried to hide his login.

 

[Pierre without password]

Heat

Sliver - "Its sounds like they are electric space heaters?
Heating the airspace under the house therefore heating the ground floor floor (1 storey?) "

You are wright but its 2 storey and open space meanning that the second storey is 75% open from the ground up.

Sliver - "anticipate the heating load required "

For instance I was thincking of installing a thermocouple in a bird house he has in the back. This would tell me if there is sun and wind. Even in freezing temperature the temp. inside this bird house could tell me what the real house is experimenting but with a faster cycle.


Tom - You understood exactly where I want to go. Elegance is on the priority list here and HMI + PLC + optimized heat algorythme.

Sliver - Yes this house is in Quebec around and it already has a fireplace. (no room for a good woodstove).

The problem as Tom mention is this.

T(out) = -10C with sun and no wind, NO HEAT and inside temp gets to 80F in mid afternoon.

T(out) = -10C without sun and lots of wind, NO HEAT and inside temp gets to 50F in mid afternoon.

Samething goes at night.

For the others, I've $crewed-up my password and settings and cannot log anymore. Seems to be no way to get it back

 

[Sliver]

Birdhouse?

Ok Pierre.

No T/C please. They are for much higher ranges than ambient Quebec temperatures. You need an RTD.

Neither T/C or RTD by themselves can compensate for windchill. Ambient air temp at no wind = ambient air temp at 200 KPH wind.

And to take into account the sun?

HMMMMMM....

OK. First of all there has to be a bird in the birdhouse.
He, She, it must stand a small but determined distance from the RTD.
The heat generated by the bird radiates to the RTD.
The Birdhouse is on a swivel like a weather vane.
The front door always points into the wind, the back window away.
There is a plexi roof on the birdhouse so that the sun will shine in.
The RTD is painted flat black.

When the sun is out and there is no wind the RTD sees the greatest temp.

When the wind blows the heat transfered from the bird decreases.

When the sun goes behind a cloud..... you get the idea.

Substitute 10 watt light bulb for the bird, and you have a single analog input that can be tweaked to give you a reasonable, if not elegant predictor of heating load.

This started as "Tongue in Cheek" but I think I better call the patent office...

Brian.

 

[Peter Nachtwey]

Re: Heat

Sliver - "anticipate the heating load required "

Or a feed forward or a model.

For instance I was thincking of installing a thermocouple in a bird house he has in the back. This would tell me if there is sun and wind. Even in freezing temperature the temp. inside this bird house could tell me what the real house is experimenting but with a faster cycle.
[/B]

More sensors are good but I don't think this is you main problem.

Tom - You understood exactly where I want to go. Elegance is on the priority list here and HMI + PLC + optimized heat algorythme.
[/B]

Are you serious? I think the situation can be made better but it can't really be fixed.

Underground the heaters kick in and the basement temp. raises to 90C. By the time the room temp rises 1 degree we almost cannot walk on the floor.
[/B]

I assume you mean 90 degrees F. 90C would be hazardous to the health.

This is your problem. The heat is NOT being removed radiated from the floor so it must build up a big delta T to get the heat flow. That floor is also acting like a big heat capacitor so it takes time to heat it up and to cool if off. You must find a way to remove the heat from the floor and get it into the rest of the house.

Because the floor has a huge heat capacity and is apparently insulated the time constant of the floor may be on the order of hours instead of minutes. Add this to other lags and you have a system that CAN NOT be controled by a simple on/off thermostat.

First find out why the floor must be so hot. Find a way to get the cool air in the house to exchange heat with the floor. Otherwise there will always be a big difference in temperatures and slow responses to changes in heat loads. After that you can attack the controls.

We have a problem like this in reverse where I work. When it gets really hot the air conditioner cools the air but the fan is not big enough to blow enough air over the cooling coils. Parts in the air condition unit start to freeze but that doesn't help get cool air to the working spaces. We need a bigger fan or bigger heat exchanger. So does your friend.

 

[elevmike]

I cant help but wonder...

So I know this doctor. When he was a resident he was sent to work at a community health clinic. He never really saw patients, but just "friends" of patients requesting advice for their "friends". So eventually the doc caught on and started dispensing antibiotics for the "friends" and telling the "friend" to have their "friends" girlfriend to come see him also....

So Pierre, who setup the HVAC system on your house?

 

[Bitmore]

Elev...beerchug... doctor
Now we know!
I think you have had some good advice already, my only thought is what has already been sugested, a "look ahead" or timer function "might" work to your satisfaction. IMHO the system was sold as a "bill of goods"
Peter brought up a good point, it seems the system is not sized wrong, just the method for BTU exchange.
Not thought out or engineered properly. Extra air circulation "may" eleviate some hot spots. Look into circulating some of the heating fluid into a coil type heat exchanger (air conditioner type) and the force the heated air into the space (house). There are some very inovated designs that use existing frame work as ducts!

 

[XPierre]

Peter Nachtwey - Like most of the time I ask a question, I already have an answer. The thing is I am looking for options. Peter you are the closest one to what I thinck of doing.

1. The basement I'm telling about is not a place for living creature. My poor mastering of the english language induces some imprecisions in most of my descriptions of the problems. I sort'a always count on you guys smarts to figure out what I'm saying and the temp is 90 Celcius.... 90C.

My buddy's concern is not really about the electrical bill. He can surely afford it but who would not try to improve?

Imagine the house. (Sorry cannot post image without logging in) Its 40 feet x 40 feet. Two storeys. The ceiling is 45 degree. At each end on the second floor there is a small room 15 wide from the end. This leaves a wide open space in the centre.

There is two ceiling fan in the center. They are about 20 feet from the ceramic floor of the ground floor. They do the circulation.

When you walk in there the floor is warm.

The dead and time delay is long (of course).

My "solution" is based on the outside bird-house sensor AND a sensor IN the floor.

We have access to PID and Fuzzy logic. This ends up in a cycling ON/OFF trigger but it still will be a lot better than what he has now.

For instance. If the ambiant is OK at 75F and we keep the floor at 75 instead of shutting down the heaters...

And also, we could have an ambiant thermostat in cascade with the floor thermostat itself cascading to the basement thermostat.

Ex: When the chill factor would be XX Watts/SqInch, 75F Ta would need 80F T(floor) which would need 100F T(basement) when its stabilised.

I tend to beleive that the 2 important numbers come from the chill factor outside AND the floor Temp.

Other options?

If it looks like a Cat...

 

[elevmike]

I don't think any amout of control can ever make up for a poorly designed mechanical system. I think Peter's right, and that you have to move the heat from the basement to the living space faster. The floor heat/load capacity is nothing but a barrier. The only way this is going to be done is by placing vent holes in the floor, allowing the heated air into the living space.

I would be tempted to install some ducting and a circulation blower to be controlled by a conventional thermostat. Otherwise any type of control system would be required to accuratly anticipate the weather hours in advance. If you can do that, please feel free to provde me with next weeks lottory numbers.....

Winters a-comin..... Mike.

 

[XPierre]

Sorry elevmike but maybee your experise is not in the heating industries.

I've manage good results with farr more complicated systems.

See, the big mass which appears to be causing us a problem can be reversed as being a good damper for our enrgigy fluctuations.

As for ALL heat systems we have to deal with what comes IN, what goes OUT and the LOAD.

What comes IN is pretty much stable.
What goes OUT suffers variations.
The LOAD is pretty much stable.

When the losses or the load varies, there is no way to have good control but to use a feed-forward algorythme. Or is there?

Engineer: A person who knows a great deal about very little and who goes along knowing more and more about less and less, until finally he knows practically everything about nothing.

Lets be imaginative

 

[Tom Jenkins]

OK, Pierre, it looks like you hve several problems here. Fortunately, I think you can hae some fun and give the guy a system that works.

As others have pointed out, wind chill doesn't affect the sensor temp. It does, however, affect the rate of infiltration - how much cold outside air is getting into the house. It also affects the rate of heat loss through walls and roof etc. You could put a cheap wind speed sensor (Dwyer has them) and compensate, but I think that level of elegance is not necessary.

I haven't done HVAC for decades, and my text is buried in the basement someplace, but from memory here are some thoughts (for what they are worth).

I would have three control "loops". First, anytime the outside temperature is below 60°F or so kick on the in floor heating and use a simple thermostat control to maintain an baseline floor temp of around 70°F or so with a deadband of maybe -5°F.

Second, use a simple thermostat control based on inside air temperature so that when the inside air temp is below 68°F or whatever the setting is, a higher setpoint is sent to the in floor heating to get it up to the 90°F or 100°F or whatever is required to get adequate heat transfer to the room air. This would correspond to your current heating cycle operation.

Third is the outside air temp tie in. Have a setting that is proportional to the difference between inside and outside air temperature. The greater the difference the greater the setting. This amount would be added to the baseline floor temperature setting to minimize the response delay of the heating system and should minimize inside temperature fluctuations on very cold days.

For reference, the optimized start stop I referred to above is actually used to pre-start the heating or cooling system when changing from a nighttime to a daytime thermostat setting. It does compensate for thermal mass and delays, but only with respaect to changing inside temperatures. I don't think it would work for your objcetive, but if you wanted to add nightime or unoccupied setback on the inside air temp setpoint you could incorporate it. Essentially you have a time delay to start the heating syetem before the set time, and the amount of time is proportional to the difference between outside and inside air temperature.