Welcome! Here are the website rules, as well as some tips for using this forum.
Need to contact us? Visit https://heatinghelp.com/contact-us/.
Click here to Find a Contractor in your area.

first stab at heating model

I wouldn't describe my control as custom as much as roll your own. You can't buy one for love or money.

My feeling is that you can't control the house until you have an understanding of how to control one zone. Extending one zone to a house will certinaly have any number of additional issues.

To say that I have to deal with the pressure issue is different than saying it will change. :) I have one constant delta P pump that I could not get to work on my VFD. I have a second trick to get that to work, and am looking at another pump as well. I also own a couple differential pressure senders that I can use to measure drag loss in the pipes. This certainly makes the valve setting more complex.

But the model says it's really knowing the water temperature differential and the emitter output function that are keys for energy input. The third part are things that are very hard to measure. These include the non-heated mean radiant temperature. For this, I'll start by waiving my hands and say this is a strict function of the indoor/outdoor temperature differential and the construction type.

return temp is allowed to vary, and the modulation is varied to deliver desired heat to the circuit. Also, since I am making the modulation decision off the heating supply temperature rather than the boiler temperature, mixing issues involved with P/S plumbing are addresses.

Until I have a proven model, attacking the more complex problem of the whole house is too hard. I take your thoughts on difficulty of applying information learned from one zone to another to heart. If it doesn't work, it's back to the drawing board. Either way I'll report progress as I go.

jerry

Comments

  • jerry scharf_3
    jerry scharf_3 Member Posts: 419
    I've got a first cut to look at

    I'm not sure how many are interested in the ongoing controls project I have. For those that are, I've got a first stab at what the heating model for the house needs to be. If anyone knows of a board where discussing things like this would be more common, please let me know.

    Comments, rotten tomatoes welcome, etc

    jerry
  • Mike T., Swampeast MO
    Mike T., Swampeast MO Member Posts: 6,928


    Re: Heat input

    You've lost me. Are you saying that you will assume that the surface of the tube will behave as if it has a general temperature of supply tube surface temperature - return tube surface temperature * 0.4? How can you use this to derive actual heat input?

    Re: Steady state

    "When we do this [achieve steady state] we end up with three factors involved: the heat loss; the heat input and the emitter delay."

    Here's my take. Sorry that it's so long-winded...

    In true steady-state, time ceases to have meaning so there is no "emitter delay".

    Every set of indoor/outdoor conditions has an associated steady state. Thought of this way, time only enters the equation when either outdoor conditions change (temp, sun, wind, rain, etc) or indoor conditions change (solar gain, occupancy gain, user-requested change, etc.)

    When any change (indoor or outdoor) occurs, the ultimate goal is to change the burner modulation level once and only once to the new output that will achieve steady state under the new conditions. Searching for new steady-state conditions by intentionally varying the modulation rate to find the "sweet spot" will be significantly less efficient than making a one-time change.

    One way to learn how to achieve this is by establishing a simple relationship between outdoor temperature and a "target" temperature for the boiler. With a known (or at least accurately estimated) increase in burner output the boiler can "learn" how to jump between steady-states by analyzing the boiler temperature rise over time and over various outdoor temperatures. Note however that this ONLY works well when the only change is outdoor temperature. Changing the simple equation that relates outdoor temperature to target requires re-learning.

    The problem with this scheme is that the user, not the boiler, is tasked with finding the simple equation that relates outdoor temperature to target. This is a long, tedious process as it takes very significant amounts of time for the boiler to learn how to "jump" from state to state and the structure to again approach true steady-state conditions.

    Now comes the fun part:

    If the user finds a "just adequate" equation and the boiler has learned to accurately "jump" between steady states, ultimate efficiency is achieved. However, it becomes nearly impossible to raise space temperature in any space or zone unless you also "tell" the boiler that you want to do this. Conventional digital thermostats are incapable of this because they are already constantly (or nearly constantly) calling for heat to maintain the steady-state. They can only produce inputs of 1 and 0--not a "big 1".

    Enter the Buderus RC-10 controller--instead of 1 and 0 it delivers a vector to the boiler, e.g. how far away and which direction the actual air temp is from desired. Problem with this is that the controller becomes the ultimate master of the entire structure. Fine if only a single well-balanced single zone, but not so fine for multiple zones that may not be well-balanced and may not be operating at the same desired setpoint.

    Enter TRVs. Being proportional devices, TRVs vary flow to affect the average temperature of the emitter. By doing this they are nearly always "on" to some degree. They rapidly respond to both internal and external changes. They allow different temperatures to be maintained in different rooms without producing any "zero" calls for heat. Great so far, but they [supposedly] have no way to communicate back to the boiler if you want to raise space temp, so you must tell the boiler that you want to do this.

    Remember: in all three cases I am assuming a nearly perfect boiler target just barely suited to maintain desired indoor conditions. All three of these systems can be adjusted to achieve this reasonably. Most difficult in this regard is multi-zoned systems with digital thermostats; RC-10 with TRVs on the rest of the emitters is easiest; TRVs with outdoor reset a bit more difficult.

    Ultimate efficiency (given nature) can be produced in all of these systems as long as you're willing to foregoe the ability to change to a new indoor state with reasonable speed and without significant adjustments. To some designers this is not an important problem as such changes always lead to inefficiencies.

    At present the limitation is solved the same way in all of these systems: by operating the boiler at a higher-than-necessary target.

    Digitally controlled systems become "more digital".

    Multi-zone, indoor reset systems are set up to such that the space with the controller is forced to produce a higher-than-necessary target in the "master" zone.

    Fully TRVd systems throttle their flow to achieve the same average emitter temperature. Timed daily setback of boiler target can somewhat mask, but not completely solve the problem by the way by introducing a somewhat less inefficient inefficiency...

    ---------------------------------------------------

    It is however possible to:

    1) Determine the most efficient boiler target with minimal user intervention--it won't be a simple line by the way as slope will change somewhat with changing outdoor temps. Somewhat similar to the infinitely variable loudness control on high-end stereo pre-amplifiers based on the physiology of human ears...

    2) Detect the intentions of the user via the water itself and make appropriate output adjustments. While this will involve accepting some inefficiency, such will only occur when wanted with return to the most efficient operation happening as soon as the request is fulfilled.

    Should I continue?









  • Steve D._4
    Steve D._4 Member Posts: 22
    This looks interesting.

    What is it you are trying to do? Is this a product you are developing or a personal learning experience? I am not asking to bust b***s but to figure out what level of sophistication you will be willing to go to. I am not a controls expert but I do work with some.

    Try googling PID control, servo control, automatic control. Also, Tekmar has interesting white papers explaining different control schemes.

    You look like you have the basics under control. You will need to find the time constant of your system. It will be different for different zones, buildings, etc. Your control system should be able to look at the "error signal" or the difference of where you are and where you want to be. Values for proportion, integral and derivative (PID) are set to let the system approach the setpoint gradually or overshoot, according to your wishes.

    The system is basically like your cruise control system. It will need to maintain setpoint on uphill (night, windy) or downhill (late march sunny day, 50*). This will be tough because you have no braking power other than shutdown.

    While I was thermostat hunting this fall I did find that most electronic types were PI type. If it has a lookahead feature for setback recovery, it is "learning" your house to get the PI variables. It does this by guessing, measuring its success, making a change and noting the improvement. As conditions change it can tweak itself to stay on track.

    Good Luck, I will be bookmarking this one.

    Steve D.
  • gasfolk_2
    gasfolk_2 Member Posts: 17
    Braking

    Steve,

    Is "braking" unnecessary because the "flywheel" speed is always adjusted to approximate the load, so the power added by the boiler is always quite close to what is needed by the load, thereby limiting overshoot and improving control?

    gf
  • gasfolk_2
    gasfolk_2 Member Posts: 17
    Optimal Supply Reset

    Jerry,

    If the supply water temp is just enough to satisfy the coldest zone (e.g. a northern bedroom), could excess BTUs from another zone (e.g. a south-facing sunroom) be redistributed by the system? Could the return temp exceed supply temp in a room with substantial solar gain? Or at least, could an optimal supply reset curve prevent significant overheating by minimizing delta-T in other zones?

    Great thread. Great idea.

    gf
  • jerry scharf_3
    jerry scharf_3 Member Posts: 419
    What am I doing

    I wonder this often. :)

    I'm doing control experiments, not building a product. No sane person would deal with the configuration that I have for the simple controls. It has lots of stuff that people would want set or learned for them.

    PID controllers are inherently reactive. They balance the ability to react to changes against the ability to tract with stability. When you have environmental changes (such as solar gain) that act much faster than the emitter, no form of PID can correctl track. You need to move to a proactive system, which predicts influences and takes into account the emitter response curve.

    For each zone I have three temp sensors, supply water, return water and room MRT. I also have the room geometry, emitter type and any computer synthesized inputs I want to generate. With all this, I am trying to balance boiler efficiency against target temperature tracking by varying the zone water flow and boiler modulation.

    I am sure there are tekmar engineers who would understand all this, but it's not in any product they or other have.

    jerry
  • jerry scharf_3
    jerry scharf_3 Member Posts: 419
    this is a common misconception

    That you just match the outdoor reset to the load and then everything is perfect. For this discussion, you need to abandon the analogy.

    With outdoor reset set to just make up the current heat loss, there can several problems. First, if you decide you want to increase the room temperature, you have a situation where it can take hours to increase a single degree. Second, the larger the delay between the water temp change and the emitter output reaching the new output level, the worse the temperature tracking is. You also can have problems when there are other heat sources not factored in to the control system (solar gain, people...)

    I'm exploring ways of dealing with these issues.

    jerry
  • Mike T., Swampeast MO
    Mike T., Swampeast MO Member Posts: 6,928


    We're at opposite ends of the sinking Titanic Jerry. I'm just not sure who is on the stern and who is on the bow--I only know the inevitiable result.

    Luciously sweeping second hand of a Rolex or the jumping second hand of a Seiko. Both show seconds of time with equal accuracy.

  • jerry scharf_3
    jerry scharf_3 Member Posts: 419
    I'll break the response into a couple pieces

    This would get very messy very quickly otherwise.

    First, Steady state was the wrong term. The first condition is keeping the inside MRT constant with varying outside air temperature. In this case, when the "factors" are correct, there should be a pair of analogous curves for input and loss, and there should an error in the temperature tracking that should be predicted by the differential of the outside air. In the real world of heating, there is almost never a steady state.

    You say there is a simple relationship between the supply water temp and the outside air. This is classic outdoor reset. There are a couple problems with it. First, I have to control the flow and thus the return water temperature as well as the supply water temperature. Shooting for a fixed delta T during testing is a simple minded way to try to make the energy supplied to the emitter approximately proportional t the supply temp. That's fine for locating parameters, but it's not a functioning control.

    The dynamic case is where either the target is changed or there is an external influence that is not covered by the basic operation.

    jerry
  • jerry scharf_3
    jerry scharf_3 Member Posts: 419
    part 2

    Mike,

    (Don't read this as being harsh or not wanting yoyr input.)

    I think you are being limited by your situation. You have TRVs that do the flow control and you get to play with outdoor reset to control the supply. You also have a single type of emitter.

    You fit yourself into the Viessmann control view, and you are turning the knobs that you have based on the single variable you have, return water temperature. The knobs cause the boiler controls to do various things to try to optimize the results.

    I have a system where I control the burner modulation and the rate of water flow through the emitters. I have a variety of radiant emitters with widely varying response and delay curves. I also have around 40 sensors that are provinding information about various parts of the system, plus any number of simulated inputs. I specificly did not select a Vitodens because I do not want to fit into the Viessmann conrtol view. Also, there is a difference between digital (calculated with numbers) and binary (on/off.) My entire system is digital, but nothing is binary.

    I need to have a model of what heating a zone is like to both implement controls and to evaluate results against. You have to view the entire system under computer control to get where I'm going.

    jerry
  • bob_50
    bob_50 Member Posts: 306
    Accuracy

    Mike, I think you'll find that any of the quartz crystal watches are more accurate than a Rolex. The second hand on a Rolex jumps five times per second. To solve the problem of the boiler knowing which room (zone) has the greatest load you might want to take a look at pneumatic controls. I know you like analog controls. If you replaced your TRVs with single seated N.O. pneumatic radiator valves and wall stats and then hooked in a pneumatic load analyzer (lowest selector). Feed the signal from the analyzer to a dual input reciever controller, which would offset your outdoor reset curve and send a signal to a transducer at the boiler. All analog tried and true technology from the 60's. I don't think the improvment would be worth the cost. I think your Viessmann and cast iron is hard to beat. I wish I could afford a Buderus panel rad set up for my house but I spent all my money on the Rolex. I'll just have to sit here and look impotent and shiver. bob
  • jerry scharf_3
    jerry scharf_3 Member Posts: 419
    doesn't work

    This has been discussed before. I've done the numbers. If you look at the water temp needed to heat the cold room, it will probably be as high or higher than the water temperature you can expect off the hot room. So if you open the valve, you just make the hot room hotter.

    It takes a good temperature differential to push the heat through the wall of the tubing and the floor to put enough heat to make up for the room loss.

    jerry
  • Steve D._4
    Steve D._4 Member Posts: 22


    "Braking" is required when the load reverses. In your car it happens when you go down a hill steep enough to accelerate the car past the set point. Your cruise lets go of the car and you will feel it take over when the car hits the cruise speed. Auto mfr's choose to ignore this because it is not a big deal. From a heating standpoint you could get a reversing load when there is a lot of solar gain, woodstove, people gathering, etc. What to do with this error is just one of the decisions Jerry will be making.

    Steve
  • Steve D._4
    Steve D._4 Member Posts: 22
    Jerry,

    "You need to move to a proactive system, which predicts influences and takes into account the emitter response curve."

    If you figure this out, apply for a job on wall st. Since you want to use a pc to do this, could you have the pc go to noaa.gov or reputable weather site and get the forecast for the next day. Excel can get data off the web. If you have tomorrow's temp, sunrise / set, sun/cloud prediction you could generate a control plan for the day that anticipates the environment in advance.

    Steve D.
  • Mike T., Swampeast MO
    Mike T., Swampeast MO Member Posts: 6,928


    Not taking it harshly--hope you're not either as it's not at all intended.

    In my system the TRVs came first--I sincerely searched out all sort of control systems including pneumatic, PID and "custom" similar to yours. Then chose the TRVs because of their low cost, simplicity and reliability. Problem is that I really don't know what they're doing as their control mechanism is mechanical and self-contained.

    Not just one type of emitter by the way--there are also radiant floors with a number of different heat transfer methods, but all engineered to work at the same temperature. This has caused me some problems with lowered target temps as the supply curve is the only control for the radiant.

    Vitodens was the only boiler that could directly drive the system with only a single circulator. All the rest required primary/secondary and weren't really designed around TRVd systems. Control philosophy of the Vito is good, but just doesn't seem designed to work at its' best for US construction, desires, etc.

    Do appreciate what you're trying to do: model one zone very well, then attempt to apply to different zones of the same emitter type by using formulas derived when making the original model.

    Am however really confused by the "Heat input" portion of your paper. Perhaps I'm just dense, but I see nothing there that actually relates to estimating (let alone accurately measuring) the heat input in a single zone of a many zoned system in actual operation.

    Since you're operating multiple zones from a single circulator with digitally controlled variable flow valves, don't forget that changing flow in one zone affects all of the rest. Not to mention that it will also affect the return temperature seen by your boiler (artifact of primary/secondary) which will in turn affect the boiler's output. So, to achieve your "steady-state" measurement in a single zone you really need "steady-state" throughout the entire system.

    Am not saying that this can't be done, but it looks like you'll have assumptions and errors of unknown magnitude at every step of the process. You'll certainly be able to derive a good model for one zone in one set of circumstances--perhaps for one zone in a number of circumstances but I bet you'll have severe problems applying it to other zones.

    I do believe that you'll save a lot of time if you begin with a heat loss assumption significantly lower than what is produced by Manual J.
  • Mike T., Swampeast MO
    Mike T., Swampeast MO Member Posts: 6,928


    I seriously looked at such pneumatic systems. WAY beyond my means.
  • Mike T., Swampeast MO
    Mike T., Swampeast MO Member Posts: 6,928


    Do realize that "steady state" only exists in the laboratory and even then the statement "is at thermal equilibrium with" is the result of an asymptotic calculation that doesn't explain when or how the "equilibrium" has occurred.

    "Classic" outdoor reset may be simple and rather crude, but combined with TRVs and a directly driven system it results in about the closest thing to "steady state" attainable. The outdoor reset makes the "general" supply temp assumption; the TRVs do the fine tuning.

    Is it not possible for you to do similar with your valves? Can you not set up a proportional feedback mechanism? You said that the valves used were quite inexpensive. Do they perhaps lack the ability to make quite fine adjustments in flow? Believe you said that many of your zones shut down once or twice a day. If so while some of your control is certainly "digital" the essence remains binary... Do realize this is exactly what you're trying to eliminate, but I believe what you're doing is more suitable to skyscrapers than single-family homes.

    The dynamic case is where either the target is changed or there is an external influence that is not covered by the basic operation.

    Could not agree more! In that regard we're working towards the same goal on that sinking Titanic...

  • Ron Schroeder
    Ron Schroeder Member Posts: 998


    Hi Mike,

    The other boilers don't have to be piped primary/secondary. They can be pumped with a single circ. and a pressure balancing bypass valve. I have 14 zones on one circ. on my own system.

    Ron
  • Mike T., Swampeast MO
    Mike T., Swampeast MO Member Posts: 6,928


    Will you say which condensing/modulating boiler?

    Does this have factory blessing?
  • Ron Schroeder
    Ron Schroeder Member Posts: 998


    Hi Mike,

    Not any specific boiler but I have used pressure balancing bypass even when using tankless hot water heaters as boilers. The purpose is to keep the flow thru the boiler above the minimum spec. I suspect that the manufacturers would approve when supported by flow calculations.
  • gasfolk
    gasfolk Member Posts: 392
    Jerry &/or Steve,

    If outside temp and solar irradiation are continuously differentiable, could you skip NOAA and simply adjust your supply reset curve based on changes in local weather conditions at 5 or 15 or 30 minute sensor interval? Accurate enough?

    gf
  • gasfolk
    gasfolk Member Posts: 392
    Ooops,

    sorry, I think you've also already answered this above.

    Really glad to read what you are doing, and will try to stop asking simplistic homeowner Q's...

    Thanks,

    gf
  • gasfolk_2
    gasfolk_2 Member Posts: 17
    Braking

    Besides the complexity of the piping etc., why couldn't a control sense whether a zone was above or below desired temp and send appropriate temperature supply water to deliver or remove BTUs from that room? Isn't removal of BTUs what some passive solar designs do with their sunrooms? Perhpas not worth the piping and control complexity?
  • Ron Schroeder
    Ron Schroeder Member Posts: 998


    Hi gasfolk,

    Radiators and baseboards are notoriously ineffective for removing heat due to the very small delta T.
  • Rodney Summers
    Rodney Summers Member Posts: 748
    In defense of PID control

    First there was the on-off thermostat. That's like trying to hammer a nail with a blindfold on your eyes - it's easy to miss the target. Then came the heat anticipator, that's more like peeking from under the blindfold - but the degree of customization is highly dubious. Then, the space age gave us the PID control with auto-tuning abilities - finally casting off the blindfold (the math behind the PID control theory is at least several hundred years old, linear algebra and differential equations are a big part of it but without PID theory it would be impossible to send rocket ships to the moon)

    What does the PID (proportional, integral, derivative) see during auto-tuning? well, everything that has to do with the response characteristics of your dynamic system.

    Imagine yourself at the car rental lot. The clerk gives you the keys and you - the Positively Insured Driver - take control of the car. It's a car type you are not at all familiar with and you're ready to go...

    Only James Bond would immediately zoom into heavy traffic. But us, while looking for the Champagne bottle cooler, we'll need to drive once around the lot and see how the car feels and at least locate the turn signal and the horn. It's always important to beep the horn!

    Is the gas pedal touchy? Is the brake spongy? Does the car drag itself? You're gathering many important parameters about the automobile - you're thus auto-tuning.

    It's all rather intuitive and you'll get a very good descriptive feel for your control system.

    There is another way.

    While you're in the parking lot, you could oh so simply measure the car, its weight, its springs, its engine size, its transmission ratios, read the owner's manual, heh heh heh, and theoretically figure out exactly how the car will handle. This is coming up with tuning parameters by hand. Useful if you've got to have a good idea of how the car will handle before even driving it around the lot, and further auto-fine-tuning is still appropriate.

    The PID auto-tuning is the par excellence empirical method for measuring the features of your heating model.

    Once the PID controller has figured out the good tuning factors everything becomes pre-emptive not just reactive. We're smart that way, us PIDs.

    The understanding I got from your model, Jerry, is exactly what the auto-tuning features of a PID controller does. Your path makes a lot of sense, it's time to bring rocket ship theory to home heating.

    You start with rough ASHRAE heat loss calculations. Then to get further refinements you analyze your particular system (which a good PID controller will do for you) As you hint, the theoretical aspect can seem to become overwhelming.

    Mathematically speaking we're talking about first order system response for thermal systems. Luckily it could be much much worse. Look for discussions on these topics perhaps.

    To study models and control systems, Labview and Matlab Simulab are both wonderful software tools. They could be the ideal thing for connecting your sensors and coming up with an idealized control scheme, albeit a bit over the top.

    In a re-work of your text, I wouldn't focus on steady-state characteristics all that much, it's the dynamic characteristics that are useful. Also, rather than worry much about temperature I would deal with heat flow (heat loss, heat input, heat output, etc)

    Temperature is only the apparent symptom of the underlying heat movements and the substance of home heating is moving heat around.

    Just like when we drive, it's the odometer count that gets us somewhere, not so much the speedometer. Of course, go explain that to the cop who just clocked you "at a high rate of speed".

    That's all the tomatoes I've got, but thanks to my aim, I'm not sure I hit or missed any of the points you were interested in.

    Like Steve D. I'll be following this thread.

    Vroom vroom. For what Labview can do, go to:

    http://www.ni.com/labview/90sec.htm
  • jerry scharf_3
    jerry scharf_3 Member Posts: 419
    PIDs aren't bad

    Christian,

    They were amazing 40 years ago when I first know of their use. I've done a little bit of playing with fuzzy logic, and haven't had a job yet for neural network based controllers, but have reviewed the literature on these as well. When applied to controls, both are different attacks on the problem with different strengths and weaknesses than PIDs. Fuzzy logic takes quite a bit more time to design, but are good at tracking and are more stable than an optimized PID across wide ranges of input.

    The problem with adaptive PIDs is that when you are exposed to conditions outside the learning, things become dicey. If you train the car in the US, then take it to Germany, you could have unfortunate control lag at times when you need faster response. Stability outside training range is also a significant issue for neural nets. It is also often where well designed fuzzy logic can shine.

    The simple truth is a PID is a relatively simple strategy, cheap to implement and widely applicable. And it was WAY better than the relay logic it superceded. That doesn't make it wrong, just not interesting to me. I want to see what works and doesn't work that uses different control theories.

    jerry
  • jerry scharf_3
    jerry scharf_3 Member Posts: 419
    weather forcasts

    It's on the to do list, but way down the road. Using current cloud cover info from automated airport weather stations to estimate solar gain is on the mid term list. I will also have the output of the solar collectors for input as well.

    As for wall st, I have no magic in predicting group desires of those selling and buying stock. A model for heating a house is much more my style. :)

    jerry
  • Christian Egli_2
    Christian Egli_2 Member Posts: 812
    Still waiting for the Champagne to fizz

    The theory behind control systems is the pearl I was after, no so much the fantastic black boxes it all comes in.

    As you well know, any control model can be made totally useless if you feed it with garbage data. But how to tell the trash apart? we don't use see-through transparent garbage bags...

    Some data is very sensitive and critical to the stability of your control system. Some (probably most) doesn't have much incidence on anything.

    The only way is to tear everything open and analyze. You can do all that with the mathematics and the theory that pulls the strings behind control theory. The concepts are relatively straight forward, the hieroglyphic math is intimidating and the understanding that comes with it is powerful. Keep in mind that graphical analysis with plots can bring you all the way to your destination.

    PID parameters peel themselves right off of the differential equation that describes any system, it's a case of : "let me read the label in your clothes and I'll tell you how big you are, and I can do this without even looking at you, just your clothes" spooky, no?.

    Fuzzy logic is another step down the same line, with an added endearing title. Who wouldn't love something fuzzy? However, fuzzy principles are exactly what was hiding behind old fashioned microswitches and relay systems of the past. They are the ultimate pure reactive systems and behave with bureaucratic rigor.

    Fuzzy theory is a great at helping the designer figure what the desired output should be, at the same time, there are no tailored clothes to look at, fuzzy logic theory offers little insight at how the system behaves or what makes it tick. The designer has to provide that.

    I read somewhere that there is apparently a big fight going on between control people who advocate either PID or fuzzy logic. A thermal system would seem to scream PID because everything related to heat conduction, convection, radiation and accumulation is very reliably determined by a differential equation while there aren't many or any seemingly random inputs.

    But that's a bigger debate and I don't doubt your issues about optimization and wide ranging stability.

    You're on the right track, keep going, trust your intuition and verify it with a bubbling splash of theory. Weed out the junk and keep us posted on what you find. I'm interested.

    This all seems to be lots of fun. What data acquisition software works best for you? Will you use ladder logic and programmable logic controllers?

    Thanks.
  • Steve D._4
    Steve D._4 Member Posts: 22
    solar gain info

    If weather lookups are a distant future item, could you use a photovoltaic cell to provide feedback for the solar situation?

    It has become clear to me that this is a site not for the above average heating contractor, but the way above average. Keep plugging away and spreading the information.

    Steve
  • Kevin McAfee
    Kevin McAfee Member Posts: 6
    Heat gain back stored back into the \"system \"

    Jerry,

    I have been reading this post, and have a few thoughts as well as questions for you to ponder. First, please know that I am not a professional in this field. I am well read homeowner/builder with an Engineering Science and Mechanics degree from our good old PSU (88). Anyway, I have installed warmboard in my house (2 floors) and cast tubing in the floor of my basement. I designed my home to be passive solar, with the "correct" glazing and sited properly for the maximum benefit. In addition the house is well insulated. This is my first heating season, and so far things are performing pretty muc as expected. I am controling a propane boiler with a tekmar 361. I am running constant circulation. Enough about the system.

    When the sun comes out for extended periods, the house grows quite warm. I have been conisdering different ways to store this heat for later use. Since the house stays warm, the heat doesn't come on for long periods of time. During that time, the water temperature genarlly will rise several degrees (system capacity is alittle over 30 gallons) Back (north) rooms stay very comfortable. I poured the slab in the basement to be 6" deep. I considering doing reverse hydro air to warm up the flooe in the basement to stoor the heat. Currently, I keep the basement temp around 62-65 degrees. With the addition of a couple of valves and triggering off of the cooling contact on my thermostat, I am thinking I can dump the excess heat into the basement floor for use later by using an air to H20 heat exchanger. Any thoughts? I have been intrigued by the idea, so when I assembled the system, I included the appropriate connections required to do this.
  • Kevin_in_Denver_2
    Kevin_in_Denver_2 Member Posts: 588
    My house is set up similarly

    I'd say save your money on the fan coils. Even if you were feeding the fan coil 75F water, the air will feel cool blowing around. And it's very unlikely that a day's worth of solar gain will even get that 6" slab up to 75F. You need temperature differences to move heat around efficiently, and your delta T's are quite small.

    However, an approach that is definitely worth trying is simply to provide continous circulation to the entire house and basement from 10am to about 8pm (depends on solar gain and outdoor temp). (Forgive all the following if you've already tried this) Under this scenario, warmth from the passively heated rooms/floors will automatically migrate to the basement slab. The Warmboard acts like an active solar collector, and the basement slab acts like the storage tank. After the sun goes down the heat automatically migrates from the slab to the rest of the house. Another way to look at it is constant circulation is keeping the whole house at a more constant, comfortable average temperature.

    Again, the delta T's may not be enough to drive the heat around sufficiently to even make this theoretical improvement noticeable. Therefore, don't spend too much time or money on the control strategy.

    I never got around to implementing this on my house because I really don't have any overheating upstairs even on the sunniest winter days. This is mainly due to the 2" concrete floors upstairs. Putting the thermal mass right where I need it was helpful in my case.

    Superinsulated Passive solar house, Buderus in floor backup heat by Mark Eatherton, 3KW grid-tied PV system, various solar thermal experiments
  • Constantin
    Constantin Member Posts: 3,796
    Agreed...

    Dumping the excess "heat" into the thermal battery of the basement is a great way to reduce the overall load on the building and quite simple to implement as long as you do not mind the possibility that the basement temp will overshoot its setpoint temperature.

    Along the same lines, you can "absorb" some of the summertime insolation to reduce the "peakiness" of the thermal spikes it can produce and hence potentially downsize the AC system. Being 6" thick, the slab ought to have a fairly impressive thermal storage capacity.

    I suppose one could use some zone or 3-way valves to enable a direct feed of Warmboard return water to the slab, but I agree that the simpler approach is to enable whole-house circulation whenever "cooling" upstairs is required.

    I am going to have to review the WWSD-mode on my Vitotronic 200 to see if the 4-way shuts all the way down and isolates the boiler from the radiant loops (as I suspect it does). Then, I'll have to see if the Tekmar 369 can be programmed for a "cooling mode" where circs come on when the temperatures in the zones overshoot. If a way could be found to redistribute the heat accordingly, that would be great.

    Trouble is, the only way to make this work in a situation where one room needs heat (i.e. 4-way valve is open) and another room is overshooting (due to insolation) is to have a pretty complicated dual manifold system where each zone can source its water from different "suppliers", i.e. the boiler or the returns from another zone.

    Thus, putting the mass where it is needed (i.e. in the rooms) is probably the smartest approach of them all because it's the simplest one to implement.
  • Weezbo
    Weezbo Member Posts: 6,232
    have you given any time to focusing on something known,

    *~/:) as bubble logic?

    it has a few years on the drawing board by the late 60's.

    it is explainable as a metaphor and as an analogy.

    Think of a calm crystal clear pool of water,the surface is calm yet there are certain factors (laws) which are in operation just in order for it to Be a calm crystal clear pool of water, every body of water must have a source.

    from That source comes is emitted a single bubble... as it raises to the surface more laws it must pass through on its way to the surface..... say it hits a law (a dense object such as a rock) and it cuts the bubble in two... now you have two bubbles rising to the surface yet smaller yet with the same information operating under the same set of rules to escape the container...as it comes up from the depths on its journey it encounters other small bubbles that is the basic H of bubble logic .....

    its application is little bit useful for say a logic learning say for Artificial Intelligence. there are studies that are teaching computer generated "beings" to evolve. These beings have something a little bit different going for them, They have Degrees of freedom in behaviour outside of certain tasks they must learn to circumnavigate or preform depending on their choice in expenditure of energy. or to say it another way,there are solutions that are derived from more than say linear interpretations ....or compound numbers of simultaneous processes of linear arrangements....

    if this sounds interesting to you ,maybe we as individuals could look into this more in depth.

    *~/:)

    It also has roots in the medical field this bubble logic in areas such as cyber electrochemical "learning" for problem solving mobile physical beings vs say a robot with preprogrammed responses and tasks with lineal processors for a "Brain".

    Mathematics is a field that is interesting to me. some perspectives are limited by completely overlooking some very important variable, science experimentation is like make up your own best guess and then go about a series of procedures that are recorded and reproducible to "prove" that best guess in a way that may be understood.

    i also like Vedic studies and often see very direct co relationships yet have no means to reveal what my thought might have looked like at the time :)

    to solve for a solution mathematics might not be the least cumbersome means to prove a result for your question is what it might look like in words. You have my apologies on not being more comprehensive in the description . were you to have built a working model and then reasoned backwards as it were, perhaps, then... "how to" would be more succinctly formulated. sorry that my mind is like ,all over the place ,expressing this rather dim point of view....

    you have expressed what i think is wanting to use some single function for variable and multiple processes .the exactitude you are searching for is within the Bubble :) why? because it hasn't reached the surface :)

    Oh Sh The Moose is banging into stuff outside....i like them buh they sure are clunky people :)
  • Kevin_in_Denver_2
    Kevin_in_Denver_2 Member Posts: 588
    storing heat in the basement slab

    I should have added that if you use this approach, the boiler needs to stay off during the constant circulation mode during the sunny day. In a typical passive solar layout, though, it's reasonable to assume that it will naturally stay off once the sun starts pouring in.
    Superinsulated Passive solar house, Buderus in floor backup heat by Mark Eatherton, 3KW grid-tied PV system, various solar thermal experiments
This discussion has been closed.