Hey @fox,

Welcome to the OVO Online Community!

What an interesting post. While I’m no expert on all things heat pumps/electric generators, we do have some Community Members who are really knowledgeable about green tech! @nealmurphy @M.isterW, @jason.lewis do you have any advice here?

You might find some helpful information in these similar threads:

Hope this helps.

@fox  I suspect this would be a very complex question to answer, and you may struggle to get something to undertake this task as with a likely 30amp connection (I understand most ASHPs need this) you would need a massive generator to kick out that much power on a constant feed. 

I am not an electrician or electrically minded, but this sounds very complicated. Are you just looking for a backup heating solution if the power is cut over winter?  You may find a generator might work with an oil filled radiator or two, as an easier solution.

Hi @fox . From what I understand (you should check with an electrical installer), the current consumed is the same - in your case 13A - but spread over the phases rather than on a single phase. The 3 phase supply is more efficient so your heat pump should use less kW than if it was single phase. There are calculations for this involving the phase differences and a 1.73 division factor but that’s way technical. 
The versions I’ve seen don’t have a significant inrush current, powering more when demand is higher so depending on the generator and supply stability needed by the pump it could be ok

Hi, guys.

Thanks for all your answers!

@jason.lewis yes, I want to backup power source for my heat pump in case of power outage in winter season. Oil radiators and other such devices is also a good choice, but for now, main target is a heat pump. 

I did some researches and consultation with an electrician, and share my results here.

Power calculation formula for a three-phase consumer: P = √3 * U * I

For my case it will be: P = √3 * 380 * 13 = 8556 W

So heat pump max power consumption for all three phases will be 8556 W (~8.5 kW).

Inrush current is still unknow for my heat pump model, but it compressor has a "smooth start": when heat pump runs, compressor consume very low current and slowly increase consumption, not just dumb "start - stop loop" on the full throttle. "Start type: Inverter" in specs means this type of compressor launch.

As a result, I can say, electric generator of 10-11 kW of nominal power can be enough to power the heat pump at it max consumption of ~8.5 kW.

At the moment, I have placed an order and am waiting for the delivery of the desired generator. This will take 3-4 weeks. After delivery and installation, I write here about the results.

Thanks again everyone!

So, yesterday was a Day-D in this topic: I started the heat pump from the electric generator. And it works great... the first 1-2 hours.

But first, a few important points about my electric generator and how the energy system is made in my house. 

1. I have a device for monitoring the energy system at home. Among other things, it shows voltage, current, power values and stores them in persistent memory. I will upload pictures from whis device below.
2. I have three voltage monitoring relays installed, for each of the three phases, which turn off the electricity supply if the voltage in the phase(s) exceeds 255 volts or falls below 192 volts. In simple words, its overvoltage protection.
3. As I calculated in a previous post, max power consumption of my heat pump is 8, 556 watts. Electric generator I used for starting the heat pump has 14, 000 watts nominal power. Heat pump compressor inrush current is still unknown, but, looks like, compressor slowly increases current consumption ("inverter" type), so the electric generator is not overloading.

Now for the main problem: what went wrong.

So, the heat pump did not work for three days, a complete "cold start" was performed. After a few hours of operation of the heat pump, all three voltage monitoring relays suddenly tripped. It turned out that the voltage was in the range of 269-272 volts, which greatly exceeds the upper permissible limit.

The heat pump, of course, turned off, and the electric generator continued to work without any problems. A minute later, I turned the heat pump back on and it continued to work successfully. But, after 15-20 minutes, everything repeated: the voltage rose to ~270 volts for a moment for all three phases, the pump turned off, and so on. And so it was a few more times, until I completely stopped the process.

What was done after:

1. Checking the generator for overload. The monitoring device helped here, it showed that the total load was about 4,000 watts, which does not really matter for a generator with a rated power of 14,000 watts.
2. Check for load spikes on the heat pump. The monitoring device is back in business. The values of voltage, current and power grew smoothly, there were no sharp upward jumps. However every 15-20 minutes, the heat pump, until then smoothly increasing its power, abruptly dropped it: the current consumption became near zero. 

Then I put the pieces of the puzzle together and realized the essence of the problem.

A large and powerful generator has a large and powerful engine with a large and heavy flywheel with a large moment of inertia. When the load from the generator abruptly leaves - it was 4,000 watts, and in a next moment it became about 50 watts - the engine and flywheel does not have time to slow down, and the alternator continues to turn at high speeds. Which leads to a sharp voltage surge, due to which the voltage control relays are then triggered and turn off the current supply to the heat pump.

I tried to determine what are generally possible magnitudes of such power surges? Is ~270 volts the maximum possible value for this generator or it can be 290 volts, 320 volts or more?

I found the specification of the alternator, it indicated the AVR model. AVR - Automatic Voltage Regulator, device inside alternator that controls output voltage stability and range. Next I looking for the specification for this AVR model and the following was indicated there:

Voltage accuracy:

 ± 0.5% from no-load to rated load in steady state conditions, with any power factor and for frequency variations ranging from -5% to +20% of the rated value

Doing simple calculations:

• generator nominal output voltage is 230 volts, due to spec
• 20% from 230 volts is 46 volts
• 230 + 46 = 276 volts - the maximum voltage that the AVR can deliver. Matches what I got in practice...

The graphs from the monitoring device below illustrate the ongoing processes.

Finally we got to the end, you can exhale :)

The main question is what to do with it? I have already considered several options, but have not come up with a good solution.

1. I can't put voltage stabilizers on every phase, as it is generally accepted that the voltage stabilizers are very harmful to the electric generator. To a greater extent, it harms when the voltage drops, but also when it is exceeded too. Although I could be wrong, and perhaps there are some special stabilizers for this case.
2. I can't increase the upper voltage threshold on the voltage control relay to ~ 270 volts, as this will lead to the fact that sensitive equipment, especially LED lamps PSU, may be damaged.

I would appreciate any help or advice.

Thank you for reading.

Partial success then .. 

I presume there is no way to adjust the output down a range to allow for the peaks?

What about a voltage clamp on the output .. power zener diode .. so that the excess is sunk to ground. Or an AVR (automatic voltage regulator) which is often designed to work with the output from a generator to prevent what your experiencing. 
Edit: just noticed you mentioned the AVR
It sort of feels like there is a piece of the output control missing as generators are often used to power a wide load variation. 
There are also a series of Thyristor Control banks which give access to phase angle and burst firing to control the output. United Automation manufacture a range of these

I presume there is no way to adjust the output down a range to allow for the peaks?

Nope.
 

It sort of feels like there is a piece of the output control missing

No, I think. I talked to several experts in the field of electrical generators, they said that AVR with 20% of "overvolrage" is a very common thing, and is not something unique or specific to my electric generator device. 
 

Generators are often used to power a wide load variation. 

True, but you can’t cheat with physics: engine flywheel (and the alternator that it rotates) will not stop immediately as soon as the load is gone. However, here we can discuss the quality of the AVR and that, hypothetically, in some models, these 20% of volrage range can be 10% or 5% or something else.

Maybe I'm just using the generator incorrectly: in order to avoid power surges, the power must (relatively) smoothly increase and smoothly go away, and heat pump cant do this so it is “incompatible” with electric generator. Or the power of the generator (the engine and alternator) should be less than, say, 10,000 watts, to avoid such phenomena.

There are also a series of Thyristor Control banks which give access to phase angle and burst firing to control the output. United Automation manufacture a range of these

Can you please provide a direct link to such products? Or at least model numbers or something.

Thank you.

https://www.united-automation.com/product/pr3-e-105kw/

I read the documentation for this device, but it seems that I did not fully understand how it would help me. Too many complicated terms for me, but if I understand correctly, this device allows you to implement a "soft start" of a heat pump, BUT not a "soft stop", required in my case.

By soft start or stop, I mean a gradual increase in the output power (current) from the device and also its gradual decrease, something like a "reversed" PWM or VFD control.

But, the heat pump is controlled by its own control board, and if it gives the "stop" command, then no external device like this will help in this case... Or I'm wrong.

Can you please explain in more detail how this device works and how it will help in my case?

Also, yor link is for "PR3-E-105KW". Looks like 105KW is too many for my 14KW electrig generator. Maybe "PR3-E-18kW" will be better?

Sorry @fox but I don’t know enough of your plant or situation to design the thing for you. All I can do is make suggestions to your questions. 
The burst control or phase angle firing enable a reduced power mode control which might be of use (can’t be sure) and provide the finer control needed