Thermostats and boiler cycling

In_New_England

I've enjoyed the many discussions of boiler cycling here and have gained some understanding of the mechanisms. The discussions I have read have focused on boiler size relative to the heat load of the house.

What role does the thermostat play. Mod-con boilers can modulate to adjust to the heat load, but thermostats are still bang-bang control. Once the temperature reaches the desired level, it clicks off, and the boiler switches off. Depending on the heat loss and heat mass of the house, the temperature drops and the cycle begins again.

What features in a thermostat could give a more continuous operation for a heating system?

Tinman

A common way to achieve that is to use the thermostat as a high limit and let the boiler control’s ODR handle the rest. 

Dave Carpentier

Although Im just starting on my modcon journey, my understanding is that a "perfect" setup would have continuous flow into your emitters where the btu being released equals the btu being lost. The btu being released is varied by the supply temperature via ODR at the boiler.
In my old conventional setup, the mixer just sent 110f (or so) water down the zones, so the room tstats had to switch on and off. Most of the tstats have anticipator type controls too, so they would cycle off regardless. I might have to swap out my old stats to ones more compatible with continuous ODR temperature flows. Just a high-limit like Steve said would be functionally appropriate, since a reg stat opens and stays open until it drops down to the swing point then goes back on. This would conflict with ODR range changes.

Dave Carpentier

Thinking more on that.. a person could balance multi-zones out with flow rates and they would respond to outdoor temp (losses) and all would be fine, but what happens when someone wants a change (cooler or warmer) to a room (zone) ? You could carefully reset the flow for that room, but it would likely throw off the other carefully adjusted rooms. Otherwise, adjusting the high-limit in a room would cause that stat to bounce off the limit (opening and closing the zone frequently). If you build in a delta to the high-limit, then thats just a tstat.
Short of installing variable speed circ pumps per zone, maybe we're stuck with stats (but get rid of any anticipator circuits).
There could be a point though where the stats are spending too much time "off" , which would be a signal to recurve the ODR temperature downward I think.
Collecting zone switching info over time (and temperature) would be useful when tweaking a setup ?

Jamie Hall

The primary setting on a thermostat which affects the length of a heating call for non-modulating (note: the boiler may cycle on other controls, such as an aquastat for hot water systems or pressure, for steam systems -- which is a completely different form of cycling) is what is variously referred to as dead band (in control terminology) or "swing". What this is the difference between the temperature IN THE THERMOSTAT at which it turns off and the temperature at which it turns on. The wider the dead band, the longer both the off cycle and the on cycle will be. Dead band might appear to be undesirable, but it is needed so that the boiler isn't constantly switching on and off at very short intervals.

Now... that's not the end of the story. Except for forced air, which responds very rapidly to a command for heat, it takes time for the radiation to heat up after the boiler turns on -- and time for it to stop radiating once the boiler turns off. If one added no other tweaks, the result would be that the space would overshoot the settings in both directions, which really isn't desirable. Various thermostat designs do add tweaks, however, either electronically in the newer digital thermostats, or by tricking the thermostat into thinking it's warm enough or cool enough when the space really isn't. This latter trickery is referred to as an anticipator, and works by having a very small, adjustable heater in the thermostat which heats when the boiler is on, and thus makes the thermostat warm up faster (and, correspondingly, cool down faster) than the space. Digital versions -- which work more or less well depending on the sophistication of the circuitry -- similarly can be set to turn the thermostat off at a preset time (or back on) before the actual set point temperature is reached.

Most digital thermostats have a setting for system type -- and this setting controls that time anticipation. The details vary with manufacturer, and even within a manufacturer's various models.

Anticipators are a little more difficult to set -- it's not a nifty menu control -- but both approaches, if set properly, can result in the actual space temperature dead band being considerably less than the intrinsic thermostat dead band.

A little more technically speaking, this on and off control by the thermostat is a form of modulation, called pulse width modulation. Just as in a "modulating" boiler, what is happening is that the boiler is asked to run just enough so that, averaged over time, the power from the boiler is equal to the heat loss of the structure. There are two parameters: duty cycle (the ratio of "on" time to the total time between successive "on" commands) and the period (or inverse, frequency), which is the actual length of time between one "on" command and the next. The duty cycle is determined by the ratio between the "on" power output of the boiler and the heating demand. The optimum period is determined by the characteristics of the entire system, not just the boiler.

In principle, a digital thermostat with enough computing power can compare the space temperature error -- how far it is from the set point -- and the rate at which the actual temperature is changing, and can "learn" how rapidly the heat output from the radiation falls off when the boiler stops, and can calculate (it's not hard) when to turn the boiler off to give zero overshoot. Older thermostats with anticipators do this automatically, if the anticipator is set correctly.

Not sure any of this is what you were after?

Dave Carpentier

Some of these boilers feature a room-sensor that offsets the ODR to try and guess the best temperature to feed. I assume thats can smooth out over/undershoot cycling and just get constant flow, if it's configured properly. Not much help with multiple zones on different demands though.

In_New_England

These are great insights, thank you.

To me, thermal equilibrium has been reached when the temperature difference between the incoming and outgoing water remains fixed i.e. the boiler is putting in exactly the energy that is being lost.

So we'd like to combine the thermostat reading (room temperature difference from desired, dTr) with knowledge of the rate of change of the difference in incoming and outgoing boiler water (dTw).

When we start up the system (cold house) we have a large dTr and we are OK with the dTw increasing, as we need to heat up the water and pump some heat into the rooms.

As dTr decreases we ramp down our output, reducing our rise in dTw.

When dTr = 0 (we are at our desired temperature) we crank down firing rate so that our dTw, which turns out to be the right dTw to keep the house in equilibrium, remains constant.

If dTr increases we crank up our firing, if dTr decreases we crank down our firing, adjusting for changes in the heat loss of the building.

A quick search on the internet reveals the following things.

Interesting keywords: "sensorless outdoor reset", "adaptive reset ratio"

Interesting article: https://www.pmengineer.com/articles/86699-outdoor-reset-control-no-quality-hydronic-system-should-be-without-it with a ton of theory and remarks on algorithms that could be used to achieve zero-cycle heating with mod-cons.

Super interesting boiler from US Boiler: The Atla https://www.usboiler.net/product/alta-high-efficiency-boiler which has the following information in the manual

Sensorless Reset
Control system monitors recent firing rate and
burner cycle data to infer current building heat
loss. Target supply water temperature is then
adjusted to match this heat load. No outdoor
sensor mounting or wiring is required. Boost
feature increases operating temperature setpoint
by 10°F (5.6°C) every 20 minutes CH demand
is not satisfied. This process will continue until
heat demand is satisfied (indoor air is at desired
temperature), or CH setpoint is reached. Once
heat demand is satisfied, operating setpoint
reverts to value determined by the Outdoor Reset
settings.

I don't know if the Lochinvar Noble has this, but my mind is already going to a microprocessor device that runs a similar algorithm to exploit the reset curve.

This device would look at the dTr and supply a fake signal to the ODR pin to adjust the target water temperature. (Ideally we'd be looking for a target dTw, but ODR on the Noble just lets us target a SWT).

As long as the dTr is decreasing (room temp rising) we'd slide the SWT down (using the fake ODR temp) and approach our asymptotic target of having a SWT that leads to dTr = 0.

When we have perturbations, our dTr deviates from zero, and we move the SWT up and down and, as long as we are in the dynamic range of the boiler, we can keep it running continuously.

The tricky thing would be to adjust the SWT in a manner that minimizes boiler cycling, we could achieve this by using a larger temperature band perhaps.

Jamie Hall

My comments, @In_New_England , applied to constant firing rate boilers. ONLY. I was not aware that you were interested in the interaction between outdoor thermostats, space thermostats, and continuously modulating boilers.

If you are using continuously modulating boilers, then you really shouldn't be using on/off thermostats for any of the control functions, at least not to get the best results. You need to be using temperature sensors instead. Your control algorithm needs at least two inputs, which must be continuously variable, and preferably three. You need the space temperature, the circulating water temperature, and you may or may not benefit from the outdoor temperature. Your computing algorithm needs to compute the target error (the distance from the setpoint and the actual temperature and the rate of change of the actual temperature. It also needs to know the characteristics of the system as a whole in terms of lags jn response to changes in inputs (these are constants in most cases). Outdoor reset is helpful in setting an approximate value for power which you then trim with the indoor error and error rate signals.

If you have target temperature error and error rate available, the difficulties in control are related entirely to the lags in the system response. If these are properly characterised, then it is possible to control space temperature (or anything else, for that matter -- the control theory is the same) within a very narrow band -- in principle a band as narrow as desired -- without, in this instance, cycling the boiler at all.

If, however, you are missing the target error and error rate signals, as is inevitable with an on/off thermostat, you will not be able to prevent cycling.

hot_rod

ODR is not the perfect control in and of itself. You need a good indoor feedback, room by room really, if you want near perfect control. The indoor feedback accounts for all the internal gains. Cooking, showering, people, computers, lighting. In well built, tight super insulated homes, a heating system is barely needed, and high mass systems can be a bugger to control with heat loads in the single digits. So the heat emitters need to match the home and occupants expectations. Many argue light, low mass systems like panel rads are ideal. Much of Europe used panel rads, constant circ, and TRV controls. Even homes on district systems.

A true constant recirc has the circulator running all heating season, and heat energy gets injected as needed. ODR can get real close to constant circulation with a lot of monitoring and adjusting. In a high mass system, the ODR can still over-shoot. So the indoor reset helps also. Some stats connect and adjust based on changes in the expected weather, if you want even another factor. Download some of the tekmar essays on ODR with indoor feedback.

Constant circulation is nice for homes with radiant slabs and solar gain with south facing glass, as you keep the slab temperature very consistent with non stop circulation.
Here is how I did my one zone shop slab with constant circ. A Lochinvar Knight with step firing, ODR, and boost function as mentioned above. This could be a buffer tank instead of a hydro sep.

I'd caution against building a frankenstein control system

In_New_England

These are all insightful comments, thank you.

On a practical note - I will be having a Lochinvar Noble installed. I currently have a simple on/off digital thermostat. Is there a modulating thermostat I should be considering that will work with the Lochinvar Noble?

I looked over their manual but there is no mention of OpenTerm or any advanced thermostat protocol. It seems that the system only accepts a on/off type thermostat?

hot_rod

just a 24V stat to that connection. Stats run from mild to wild, programmable, wifi. If you have any high mass slab areas, a slab sensor option may be nice to help with over-shooting.

Most of the fine tuning is within boiler controls these days. ODR at the very least.
I went with the Lochinvar Knight as it does have a bit more control options. I use the ODR, ramp delay, boost and limiting functions. I also have constant circulation, so the stat just opens a 3 way zone valve on heat call.

The Knight also allows the Con x Us option to watch and adjust the boiler remotely.

An Ecobee stat allows me to have a small reset, vacation mode for my frequent travel schedule. I can monitor and adjust temperatures remotely.

So some basic off the shelf components give me all the adjustability I can stand.

The Knight allows 0-10V input through the BMS function, The variable speed boiler circulator, included, can also be controlled with a 0-10V signal. Best to download the Knight manual to see how that can be configured. I have never used that control logic.

pedmec

because the boiler supply temperature is based on outdoor temperature to try and match the heat loss of the structure. if you set up your algorithm correctly your boiler will basically try and match the loss of the structure.