Grundfos Alpha Circulators 'Auto Adapt' Program Question

The Grundfos videos I've watched all indicate that over time the Alpha circulator will learn (literally train itself) the optimum output(s) required for your zoned hydronic system. Based on what criteria? Since it is Delta-P and not Delta-T driven, what Delta-P is best for any given install, and how on earth would a circulators internal programming logic come to know this? This circulator is and will always be totally blind to matters such as the zones BTUH demand, the boilers BTUH delivery capabilities, and the optimum temperature drop requirements across various zones, so by what internal logic can it substitute pressure drop for temperature drop, and in so doing, how can it ever know (and thereby ever learn) what the actual optimum requirement for GPM delivery will be for any given zone or combination of zones?

When I look at the pump curve "range" available via 'Auto Adapt', it is a broad and confusing range indeed. In a nutshell, I can fully and easily comprehend what this circulator is doing when it is set in any of its several "Constant Pressure" modes, but not what it is going to do when set to its "Auto Adapt" mode (which seems to be its pinnacle and preferred mode of operation). Is there any detailed literature or discussion on the web related to the logic used in "Auto Adapt" mode? All I can fathom is that it must seek to conform to some internal programming logic that is set to nominally be the best for a broad range of typically encountered requirements. But nominal and optimal can be leagues apart, since typical and actual are often leagues apart.

OK, I just found this important logic tidbit from Grundfos:

AUTOADAPT automatically searches for the lowest possible head to meet system requirements. It will continuously find the setting where optimal comfort meets minimal energy consumption.

The lowest possible head to meet system requirements is a noble ideal indeed, and it just so happens that is just what I need (see my discussion of this on the 007 overkill thread). But how can it ever know my systems optimal comfort requirement? In my case, I've concluded that the Alpha's lowest possible pump curve will work, but what about other peoples systems with different and more demanding requirements?

Here is what I believe (think) to be the case. If the circulator runs forever and the T-Stat(s) are not satisfied, the Alpha circulators programming logic circuitry will say: I've been running forever at my lowest output level possible based on my broad ranging pump curve capabilities, and the T-Stat is not being satisfied. Therefore I must run a bit harder to deliver more BTUH to the zone. And if it repeats this process over and over again, eventually it will reach a point where the T-Stat is satisfied over some anticipated (I.E, fixed into the programming logic) length of time, and it will remember its output. It will then apply this higher output automatically the next time, instead of its minimum output.

You should direct these questions to the manufacturer. Hopefully they don't just tell you it's "magic".

I have a sneaking suspicion that they are not going to be all that willing to publicly reveal the logic behind their magic. So I'm thinking it through for myself.

Larry,
I just posted a unusual Alpha system I used last season over in you 007 thread.

The key to my thinking as to how these Delta-P pumps operate in general is that they assume an optimal time duration for zone satisfaction, and they deliver an output that has been learned in the past to satisfy a zone in said optimal amount of time, while at the same time generating the minimum friction head (Delta-P) across the pump. Thus nominal circulator run time to satisfy a zone replaces the direct achievement of 20 degree F. Delta-T as the essential criteria for optimal zone heating comfort.

But since it may be 30 degrees outside one day and -5 degrees outside the next day, and the circulator will never know this, Its internal learning process must become a never ending learning (actually guessing based initially upon past historical data, and then modifying to suit the current situation at hand) process. What worked optimally once will not necessarily work optimally again, but often enough it may be close. I must assume it learns relatively fast. And on the flip side, if zones are being satisfied in too little time vs, the programming logics currently determined optimal circulator run time, the next firing will be met with a lowered head (output) from the circulator. And so on and so on.... Always learning, but never achieving a complete understanding.

Will Delta-P combined with outside temperature reset logic be the next level (generation)? A level with a more complete understanding?

I thought I read somewhere about temp probes with these pumps?? Maybe I was dreaming...

hotblast1357 wrote:I thought I read somewhere about temp probes with these pumps?? Maybe I was dreaming...

Temperature probes are the Taco way of achieving variable speed control. This logic goes by the name 'Delta-T'. Roughly a year ago I concluded on this forum that these circulators (which work well for oil and gas boilers) may prove highly problematic in a much slower to respond coal boiler environment.

Hold the presses. In reading the install manual for the AquaMotionSMART Einstein ECM variable speed circulator, if I'm understanding it correctly, I noticed that it does not have a fully automatic 'Auto Adapt' (learning capable) mode like the Grundfos Alpha, and it also does not have the multiple flat line "Constant Pressure" modes that the Alpha offers, but rather it uniquely has something called "Proportional Pressure Curves". There are 4 of these proportional pressure curves that can be selected. It will not jump from curve to curve, and with that, you have the confidence that the circulator will deliver output that is both 'variable speed' and 'fully defined', as circulator output will assuredly reside right along one of the 4 selectable yet proportional (as opposed to flat line) pressure pump curves, and it will do so via computer logic varying the pumps speed and energy consumption (which ranges from 5 to 38 Watts). It will not roam around across a broad range of virtually infinite yet unknown and thereby unpredictable (ever learning) curves. This is a uniquely different concept that I find has advantages for someone trying to be able to actually calculate and know with far more certainty right up front what the circulator will deliver. This eliminates nigh on all confusion. In my mind it is a game changer. And a winning one at that.

With the Alpha pump here is no magic to it...its just a pressure sensor hooked to a motor speed controller.....pressure drops slightly below set point, motor ramps up until pressure sensor is happy, pressure climbs slightly above set point motor slows until pressure sensor is happy again.....it does this continuously with the controller constantly chasing the pressure set point...at a constant load these swings will get smaller and smaller until they practically stop....hence the "learning" bit....as soon as a TRV or zone valve moves its back to chasing the pressure set point.

One thing to remember with the Alpha pump is that it was designed to operate on a European style heating system i.e pump permanently running with one feed and one return pipe loop with radiators controlled by TRV's that constantly modulate the water flow thru them.....no zone valves that can suddenly open and increase pump requirements by 25-50% or more.....

An Alpha pump will still work and maintain a constant head pressure when operated with on/off zone valves.....but alot of the 'smartness' is lost to the dumbness of the on/off zone valves.

Callum

unhippy wrote:With the Alpha pump here is no magic to it...its just a pressure sensor hooked to a motor speed controller.....pressure drops slightly below set point, motor ramps up until pressure sensor is happy, pressure climbs slightly above set point motor slows until pressure sensor is happy again.....it does this continuously with the controller constantly chasing the pressure set point...at a constant load these swings will get smaller and smaller until they practically stop....hence the "learning" bit....as soon as a TRV or zone valve moves its back to chasing the pressure set point.

This seems to perfectly describe the function of the Alpha circulator when set to one of its "Constant Pressure" modes. Not when set to Auto Adapt. That is, unless Auto Adapt is actually just another constant pressure mode in disguise.

lsayre wrote:That is, unless Auto Adapt is actually just another constant pressure mode in disguise.

Thats pretty much it in a nutshell...

I have 5 zones, each with a 007 taco...Would it be of any benefit for me to replace them with the Grundfos circulators that you are discussing? If so, what would the benefit be?

johnjoseph wrote:I have 5 zones, each with a 007 taco...Would it be of any benefit for me to replace them with the Grundfos circulators that you are discussing? If so, what would the benefit be?

Without a system that is configured/piped in such a way as to have normally ever changing zone circuit water flow volumes and accompanying pressure changes , then no real benefit as to making a change.

johnjoseph wrote:I have 5 zones, each with a 007 taco...Would it be of any benefit for me to replace them with the Grundfos circulators that you are discussing? If so, what would the benefit be?

Keep running the 007's. By the time they need to be replaced there will probably be more ECM circulators available, and at a lower cost.