Variable speed boiler circulator experiment

[Nofossil]

I've been running for years with the Poor Man's Variable Speed Circulator and decided that it's time to bite the bullet and use a real variable speed circulator, managed by my NFCS.

The control objective is to vary the boiler circulator speed to obtain the desired boiler outlet temp. Lower speed means more dwell time for the water in the boiler means higher outlet temperature - should work like a charm. With hotter fire or higher inlet temps, higher circ speeds would be possible while maintaining the desired outlet temp.

It works wonderfully, except.....

With slower flow through the boiler, you also have slower flow through the zones, which means colder inlet temps, which means you need slower flow through the boiler because you now need more rise.

Same thing happens the other way around when everything is cranking. High flow though the boiler leads to high flow through the zones and a smaller drop, meaning higher inlet temp at the boiler.

Bottom line - variable speed boiler circ doesn't solve all the world's problems.

 

[BoilerMan]

I always look forward to reading your posts NoFo! Informative and scientific with the NFCS. Great expierement, how are the O2 controls coming?

TS

 

[__dan]

I've been running for years with the Poor Man's Variable Speed Circulator and decided that it's time to bite the bullet and use a real variable speed circulator, managed by my NFCS.

The control objective is to vary the boiler circulator speed to obtain the desired boiler outlet temp. Lower speed means more dwell time for the water in the boiler means higher outlet temperature - should work like a charm. With hotter fire or higher inlet temps, higher circ speeds would be possible while maintaining the desired outlet temp.

It works wonderfully, except.....

With slower flow through the boiler, you also have slower flow through the zones, which means colder inlet temps, which means you need slower flow through the boiler because you now need more rise.

Same thing happens the other way around when everything is cranking. High flow though the boiler leads to high flow through the zones and a smaller drop, meaning higher inlet temp at the boiler.

Bottom line - variable speed boiler circ doesn't solve all the world's problems.

I saw something similar on a different system, cooling towers and chillers. Within the operating conditions at that instant, the source capacity and the load are relatively constant, reducing flow around the loop increases the supply/return temp differential, delta T. Increasing loop flow reduces delta T between supply and return temps. Boiler heat production and the load may remain relatively constant as the loop flow varies and supply/return delta T varies inversely with loop flow.

Desired HW supply temp would depend on the load and the heat output firing rate, varying the load or the firing rate.

If the objective is an efficiency gain, I believe there's a big gain using "hot water reset" to vary the load supply temp. The load and the boiler have two different HW temp requirements. Boiler S/R temps have to be maintained above 140 F to avoid flue gas condensation with typically a 10 deg F S/R differential (delta T) for cast iron boilers to avoid thermal shock to the cast iron. Steel boilers generally do not have a thermal shock problem.

Varying supply temp to the load is the big ticket, even with HW baseboards and CI radiators. There's a big savings as the load is satisfied with a lower HW temp, allowing the boiler to achieve setpoint and go off earlier(for oil / gas typically). The load also runs a lot quieter as the flow is on longer at a lower temp instead of cycling on/off at 180 F. Reducing metal expansion/contraction significantly reduces the noise.

I am using outdoor reset variable speed injection pumping to set the load supply temps, independent of the boiler temps, as long as boiler water is available. It runs full speed as the boiler temp is coming up and drops down to almost nothing as the boiler temp goes up above setpoint and the load return temp goes up from a cool start. I put in a Weil Mclain Ultra gas boiler for my brother with outdoor reset built into the boiler, varying the supply temps to his baseboard convectors. It seems to me that with storage your capacity and the load are relatively independent, charge storage and take the load out slowly.

I'd like to second the comment about learning a lot from your posting. Thank You. It was reading the forums here that got me into the Froling.

One thing that is a revelation for me, I cannot understand how the human race has been burning wood for the previous 10,000+ years yet what the system should be, say a gasification boiler with HW distribution, is less than 1% of the market, with the customers completely clueless. Even clean burning stoves can be a tough sell. Keep up the good work.

I will easily be less than three cord this year (and every year after), burning sloppy, prepared wood. I am amazed the market is clueless about this. Zero oil.

I wrote this entry but it has since been rewritten.

http://en.wikipedia.org/wiki/Hot_water_reset

 

[maple1]

I cannot understand how the human race has been burning wood for the previous 10,000+ years yet what the system should be, say a gasification boiler with HW distribution, is less than 1% of the market, with the customers completely clueless. Even clean burning stoves can be a tough sell. Keep up the good work.

I will easily be less than three cord this year (and every year after), burning sloppy, prepared wood. I am amazed the market is clueless about this. Zero oil.

Absolutely spot on. That was the thing that was most aggravating when I was doing my researching. There is widespread wood burning around here, has been going on for eons, and I have 3 boiler manufacturers within an hours drive of me - none of which seem interested in advancing their wood burning tech beyond the stone age. Frustrating to say the least. Kerr had a spurt with the Jetstream back then, but that has since gone by the wayside. They seem so clueless about this I just can't understand it.

Then there's all the yards I drive by with piles of freshly half-split wood in the yard....

 

[jebatty]

The control objective is to vary the boiler circulator speed to obtain the desired boiler outlet temp.

This single objective probably is not the ideal objective with a wood fired boiler. Would not the ideal objectives be to obtain combined maximum efficiency in the burn and in heat transfer to the water, plus sufficient supply of hot water to meet system demands under varying conditions? To do this 1) fan speed + O2 control to maximize the burn efficiency, 2) hx design with circulator speed (flow) control to maximize heat transfer while maintaining non-condensing stack temperature, and 3) sufficient storage with flow control to supply hot water to meet demands under varying conditions regardless of boiler operating circumstances.

The "sophisticated" gasifiers, like the lamda controlled boilers, come close to this, but I haven't seen any with a controlled variable speed circulator-loading unit designed to maximize heat transfer, particularly at the high end of boiler operation with low delta-T. It seems we pretty much agree that wood burns best and most efficiently when allowed to burn at its maximum burn rate, and while the burn rate may be controlled within limits and achieve reasonable efficiency, it cannot be modulated like a natural gas burn can be modulated (btu output) through control of the fuel supply. Therefore, a gasifier may be best "modulated" (btu output) at the storage-supply end.

What do you think?

 

[Nofossil]

I'm actually working towards three closed loops:

1. Varying blower speed to maintain optimum burn (lambda will help - not there yet). Probably will have separate primary & secondary blowers at some point.
2. Varying boiler circulator speed to maintain target outlet temp. The outlet target itself varies depending on the load - lower for hot tub, higher for baseboards, highest for topping off storage. I'm not doing outdoor reset, although I certainly could at this point. That will probably wait on my radiant main floor zone.
3. Varying bypass circulator speed to maintain target inlet temperature. I don't have that in place yet, but it will help by more accurately varying the net flow through the zones.

I'm hoping that my contributions here and on my site might help in some small way to erode the mountain of ignorance about the science of wood burning.

 

[maple1]

1. Varying blower speed to maintain optimum burn (lambda will help - not there yet). Probably will have separate primary & secondary blowers at some point.

I haven't seen any figures anywhere with respect to measured actual airflow rates to each of primary & secondary burns. Are there any? From fiddling with & watching my boiler, it seems that - in my case at least - it uses way more air in the primary fire than the secondary burn. And therefore I also think there is more to be gained by regulating primary air flow than secondary though a burn. My secondary air flow adjustment is set to almost closed, and it stays there all the time. The only real noticeable change I can get by opening more is that sometimes (not very often though), it blows the gassification fire out. Whereas very noticeable changes are seen by regulating primary air (say by cracking the door for more, or blocking the inlets by hand for less - taking to extremes for observation purposes). That is just seat of the pants stuff with no fancy exhaust gas analysis or anything like that - so I might be off base. But I guess my point of all that talk was - I wonder if a separate secondary blower would be worth it?

 

[jebatty]

... watching my boiler, it seems that - in my case at least - it uses way more air in the primary fire than the secondary burn....

This is where a universal rule/control may not work, as this may be boiler design specific. For my Tarm, for example, I set the damper on the draft fan, which supplies both primary and secondary air, at a point that produces good burn and draft, and I don't fiddle with it. In fairly warm weather, I'm somewhat draft challenged, but during the typical winter heating season draft is very good. Then I set the secondary air control at the midpoint and don't touch or tweak it. This diverts air from primary to secondary, so both are now fixed. Based on my wood, mostly pine and some aspen, and very dry, all works well, at least well enough so that I am not induced to vary the adjustments. Obviously, except by accident, some adjustment would make it work better. And this is where a more automatic control system could be of real benefit.

 

[maple1]

Yes - boiler designs can vary widely. And on some second thoughts, I think mine may actually get a fair amount of secondary air via primary air that gets through the nozzle. Perhaps that supplies a 'base amount', and my secondary control is for more as needed?

Maybe.

This whole gassification process is still pretty new to me - but very interesting.

 

[Nofossil]

I keep hoping that someone with a lambda controlled boiler will share a plot of primary and secondary flow rates over the course of a fire. I have lots of theories about what optimum would look like, but no data.

 

[ewdudley]

I keep hoping that someone with a lambda controlled boiler will share a plot of primary and secondary flow rates over the course of a fire. I have lots of theories about what optimum would look like, but no data.

I think Tom's setup varies proportion of secondary to primary, so red line in first graph is more secondary as red line increases:

https://www.hearth.com/talk/threads/not-your-grandpas-garn-graphs.80962/

So increasing secondary air until plateau, then decreasing as volatiles peter-out.

 

[TCaldwell]

Thanks EW for locating the graphs, as stated the red line represents the secondary damper output, in my case the primary air is the percentage difference with a combined total =100%.This arrangement is based on the direct acting secondary damper with a 90 degree stroke = 0-100% output, the o2 or process value scale is 0-20.9%., 0 =rich, 20.9= ambient air. The primary air creates co2, a average co2 fuel factor max for wood is approx between 17 and 20%, 0=lean and 20= rich. effectively for simplicity I consider both scale's spans the same but reversed, as the primary opens fueling the fire to create wood gas, the secondary closes the same ammt to reburn the process gas, and together closing in on setpoint. This is the concept to control for most boilers, and that is where the fun begins. All control strategies are based on combustor and airflow designs and how far the manufacturer wants to go to get to market. Zirconia o2 sensors and flue temp used to calculate a co2 value, some manufacturers independantly control p/s dampers, but the p damper works from a calculation and not a actual sensor measurement. The benefit of independant control is in pid or fuzzy parameters for each damper., As nofo is experiencing there is a stabling influence when temp is part of the control. Add that every wood load is different, and it becomes a challenge.

 

[Nofossil]

That's awesome! If I'm reading that right, your secondary air reached about 55% of total during the height of the burn, then tailed off as the fire burned out. Couple of questions:

1. Total air volume is a constant in this setup?
2. The 'secondary damper output' line is position (percent of 90 degree stroke) rather than any calibrated flow ratio?

I'm interested in learning whether combustion temperature and time since refueling can act as a reasonable proxy for primary/secondary ratio. Lambda control is nice, but there will always be a lot of boilers out there without it. If I can make a control that provides most of the advantages AND can be easily added to a lesser boiler, I think that would be a Good Thing.

 

[TCaldwell]

Please note the graphs are with a garn[uninterrupted burn], maybe someone with external storage that happens to go into idle mode will have a burn trend.
A batch burn has at least 3 states, start-up, steady state and char. with the most difficult being start-up, for o2 control. Starting a fire with a secondary burn chamber below gasification temps, will read a lower o2%, if o2 control with pid only is tried at this time, the proportional gain will be too strong, causing oscillations. This is where temperature is used to guide the controller output to 1200degf + and handed off to o2 control for the balance of the burn. Once in the steady state the controller output movement is minimal, but needs to be robust enough to handle wood shifting disturbances. Pid will reduce controller output[ secondary air] to a low limit output, about 10% to finish off the char stage. Startup can also be helped by starting the controller output at 35% or more, 1, the fire escalates slowly, 2 the damper position at 35% is closer to a fuel air neutral position, better able to respond to instability.

 

[TCaldwell]

Yes, in my case the total air volume is static throught the burn, probably the biggest contribution in the whole setup was linearizing the damper airflows to controller output. I first realized that both airflows were not equal at 50%, or alternating each one at 100%, although the physical ducting was the same. A small flow reducer in one brought them together. This was a little bit of a challenge, the total combined air volume at different controller outputs were not equal. So in reality each controller output [combined p/s has a slightly different total air volume, but if the secondary has 70% total air volume, then the primary has the remaining 30%. I linearized 10 points from 0-100%. It was suprising how far off equating controller output to damper swing can be. From what I see fan modulation is more for heat output modulation, lower fan speed when closing in on water temp setpoint, to hold off idle. I am currently working on a feedforward pid control strategy,o2 as the primary loop and secondary burn temp as the feedforward signal. basically the o2 controller output and the secondary burn temp are used to configure a modified controller output to the dampers. If this works out possibly a controller output can be derived froma modified temp signal.

 

[Nofossil]

I'm impressed - linearizing this type of setup is tedious. Ask me how I know

My PID algorithm already allows boundaries (thou shalt not allow the output to go outside this range). I also added periodic purge cycles as an option, since running primary air too low for too long can cause problems. This can happen if you need to reduce output - in my case, because I can't dump heat into storage fast enough when storage is getting 'full'.

Because of the purge cycle, a single fan and a controllable damper might be easier. With two fans, I'd have to figure out how to coordinate purge cycles.

I break a fire into multiple stages with different control actions:

1. Fire Detected. Flue temperature just jumped and combustion temp is cold. Assume bottom door and bypass damper are open. Actions:
   • Switch to wood mode.
   • Turn on tank and boiler circs to full speed to preheat boiler water jacket.
2. Phase I - getting to gasification. Operator has just enabled fan and combustion temp is low. Assume fire is burning, but not reached gasification. Actions:
   • Fan at full speed
   • Wood boiler circ at full speed
   • Bypass zone valve open
   • All other zone valves closed
3. Phase II - full burn. We were in Phase I but combustion temp has now reached gasification levels. Actions:
   • PID control of fan to hold target combustion temp
   • Manage bypass zone valve to control inlet temp
   • PID control of boiler circ to manage outlet temp
   • Manage zones to match load to output
   • Change combustion target temp if needed to limit outlet temp (avoid overheating)
4. Phase III - coals. We were in Phase II but haven't hit target combustion temp for a while. Assume we're burning out; Actions:
   • Reduce fan to low level
   • Manage bypass zone valve to control inlet temp
   • PID control of boiler circ to manage outlet temp
   • Manage zones to match load to output
5. Scavenge - fire is done - combustion temp has hit low threshold and we can't get useful outlet temp. Lets' get all the residual heat out of the boiler. Actions:
   • Fan off.
   • Bypass zone valve closed
   • 'Dump' zone selected (basement or main floor)
   • Circulate through dump zone until outlet drops to 125 degrees.
6. Done. Fire is out. Boiler is cold (125 degrees). Actions:
   • Circ off
   • Switch to Storage mode.

I have a manual 'Fan Disable' switch that physically interrupts power to the fan. The controller detects this and reverts to Phase I when the fan is turned back on, assuming that refueling has occurred. It may by in Phase I for only a couple of seconds, but a normal fire involves a couple of cycles of Phase I -> Phase II -> Phase III -> Refuel -> Phase I.

 

[maple1]

So you're using a zone valve rather than a thermostatic diverting valve for return temp protection?

Just wondering on the advantages or reasons for chosing to do that (with what would seem like added complexity involved)?

 

[Nofossil]

So you're using a zone valve rather than a thermostatic diverting valve for return temp protection?

Just wondering on the advantages or reasons for chosing to do that (with what would seem like added complexity involved)?

Well, I wasn't going to admit this in an open forum, but I'm a control freak AND a cheapskate. For me, the problems with a thermostatic diverting valve are (1) lack of ability to directly manage the temperature and performance, and (2) cost. Back in the dark ages when I did my installation, I couldn't find a 1 1/4" themostatic valve under $350. Zone valve was $35 off of eBay.

Complexity? I LOVE complexity! Another thing to control and monitor - what could be better?

Next time I do plumbing, I'm planning on replacing the bypass zone valve with a PID controlled bypass circulator which will give me even MORE control.

 

[TCaldwell]

my controller does not have a linearization block, alot do, so I used a eurotherm q498 signal conditioner/math module and a hotwire anemometer, converted cfm to mA and made the linearization table. Sounds like your next nfcs might incorporate combustion control?

 

[Nofossil]

my controller does not have a linearization block, alot do, so I used a eurotherm q498 signal conditioner/math module and a hotwire anemometer, converted cfm to mA and made the linearization table. Sounds like your next nfcs might incorporate combustion control?

Yes, we're building a more commercial version which will likely have a list of sample applications including rule sets. The actual hardware won't change from a functional point of view, but the user won't have to figure out everything from scratch. One example near and dear to my heart is the BFCS (Beer Fermentation Control System) that I've used to good effect over the last couple of years. Same hardware, different application.

In the NFCS, any input or output can use a canned (or user-provided) lookup table (40 points over the full scale range) and it does linear interpolation between the defined points. I use it for thermistor sensors, for instance.

 

[heaterman]

I cannot understand how the human race has been burning wood for the previous 10,000+ years yet what the system should be, say a gasification boiler with HW distribution, is less than 1% of the market, with the customers completely clueless. Even clean burning stoves can be a tough sell. Keep up the good work.

I will easily be less than three cord this year (and every year after), burning sloppy, prepared wood. I am amazed the market is clueless about this. Zero oil.

Pretty simple really. Our culture here in the US is totally programmed to consider only first cost rather than life cycle cost. Whatever is cheapest is what gets picked 9 times out of 10.

 

[BoilerMan]

I cannot understand how the human race has been burning wood for the previous 10,000+ years yet what the system should be, say a gasification boiler with HW distribution, is less than 1% of the market, with the customers completely clueless. Even clean burning stoves can be a tough sell. Keep up the good work.

I will easily be less than three cord this year (and every year after), burning sloppy, prepared wood. I am amazed the market is clueless about this. Zero oil.

Pretty simple really. Our culture here in the US is totally programmed to consider only first cost rather than life cycle cost. Whatever is cheapest is what gets picked 9 times out of 10.

Thats why customers want high temp fin-tube baseboard 99% of the time in new construction. In the radinat thread I spoke fo this cheapo (or uninformed and un-willing to learn) principal.

TS

 

[bigburner]

Doing a large residential system now, using nice stuff - Parts our cost are over 9.00 Sq ft. put labor on that and of course the granite counter tops and hot tub come first.