Hi,
I have 5 kVA Muti II unit and I would like to use excess power when my battery is full to heat water so nothing special. So I started thinking about what would be the most gentle way to regulate resistive heater (3kW) element from the point of view of multi.
The methods of regulation I’ve come up with are:
Autotransformer
VFD
Phase-fired control
Zero-cross switching
Feel free to add more.
The first two were ruled out. Autotransformer would be pricy for the loads I intend to use, slow, etc. The VFD would probably be the most gentle solution in the sense that the current spikes, noise etc would or at least should be minimal, also the regulation should be really smooth (small step). The problem is price at least from reputable manufacturers. Buying some Chinese VDF or vevor kind would be much cheaper but I’m not sure about their longevity and also the chances of not having a proper input filter limiting “garbage” going back to the “grid”(multi) are much higher.
Phase-fired control seems like a nice option but the interference, current spikes etc really isn’t something you want to send on its way to the multi. Especially when I can clearly see what switching 2kW water heater does with the lights.
So from my point of view, the zero-cross switching regulation seems like a winner. Transients should be a lot smaller than in the case of phase-fired control. Flicker should be a lot smaller and mainly it should be a lot more gentle on the multi.
So my question is… Is there any recommended way of regulating resistive load on the AC side of the multi that won’t negatively affect it? My main concern is using regulation that would in the long term be damaging to the multi and shorten its lifespan.
In order to regulate power with on/off switching at zero crossings, you may be looking at sub-one second cycling of the load. I have an espresso maker that does this and the inverter can’t keep up with the load changes. It survives but there are pretty large swings from inserting to feeding the grid/battery charging.
The typical triac slicing of the AC waveform does create current steps but since it’s resistive there really aren’t any spikes.
The way some power regulation is done is to slice the AC waveform at a very high frequency then filter it to remove the steps. That’s probably what the VFD does and would be the best solution.
Yes, I was planning to use bust mode because I thought it would cause less flicker etc due to the switching in zero so no sudden current spikes shouldn’t occur. But after some searching, I’ve found out the burst mode should be even worse than phase-fired control. Weird.
So maybe the triac slicing doesn’t create many spikes but I can only imagine how much flicker it would create. At the moment I have a 2 kW water heater connected to multi and when the thermostat switches the heater on you can visibly see the light bulbs dimming. Yes, those bulbs are the old filament type but they are nicely visualizing the transients.
The last suggestion seems like a winner. The VFD works a little bit differently to my knowledge (AC→DC→H bridge AC with adjusting sine wave frequency). But I wasn’t aware there is such a thing as PWM for AC. After some reading, it should create less “garbage” than burst mode and triac slicing. It still seems strange to me.
So probably PWM let’s say around 30 kHz with additional filtering on the input side to limit noise getting back to the multi seems like a way to go.
The problem isn’t when to start the load with some logic (mentioned NodeRed). The problem is the… let’s call it effectivity.
If the PV charges my batteries to 100% SoC and then is able to provide an excess 1kW of power excluding the current consumption it would not be ideal to switch 3 kW heating element. It would result in discharging the 2kW from the battery and 1kW from the PV. It would be possible to work with solar irradiance predictions and implement predictive regulation that would basically roughly ensure that at the end of the day, the battery would be at 100% SoC but in that case, you would unnecessarily cycle the battery. The better solution, at least for me is to regulate the power to the load and basically use the power from PV directly in the heater than going PV→Battery→Heater.
I do exactly this. I have a PID loop, where the MV is the dump load and the PV (process variable) is the battery power. When the battery is full, set the SP to a slightly negative value and the solar will ramp up to compensate
That is interesting, I would expect some flicker. Do you have any old filament type bulbs? Newer LEDs can tolerate a little bit more than classical light bulbs.
I will probably go with the PWM method mentioned above. Unless someone will convince me otherwise (higher transients, more flicker, etc)
Interesting. I’m still a little bit worried about how will the multi handle phase angle regulation in the long term. I’m not thinking about it failing in two or even five years, but if it could work for an extra year let’s say ten years in the future it would be nice. In the end, the transformer will most likely still be good but I don’t expect to be able to buy a replacement motherboard ten years in the future. I’m not even sure that you are able to buy them now separately.
Basically, the whole thread could be solved by Victron staff showing in the comments and saying that the phase angle regulation won’t hurt Multi in any way
I was planning on doing it the same way as you describe it (PID regulator, maybe 2DOF).
