Ess with dc load for overdimensioned mppts

Hi,

3~ mp2 3kVA’s 48Vdc, 6 mppts with 265A total (all of them having plenty of pv), 20 kWh pylontechs. Loads, gridmeter, and pv inverter in mp2’s input side.
There will be an nodered relay controlled water heater on ac side for periods of good solar/cheap/zero utility prices.

Planning to implement also dc water heater option. Is it necessary to have an smartshunt or something like that for 1,5 kw dc element? I’m not interested whether it shows up in vrm correctly or at all. Would that kind of load mess system behavior too much while having unused mppt power?

Dc contactor and 20m cabling for element etc isn’t specially cheap. Should i consider using contactor for switching one of the 60A mppt units ~160-200Vdc pv input to some separate pv heating element solution for periods of good solar?

Thanks,

Switching a heating load to utilise excess or cheap power should not mess the system behavior.
If the cost of the dc contactor and cabling is high, consider a dedicated inverter and AC heater, then the control can be done in software, or with a small relay for the inverter remote input?

Actually it is/was 6 mppts with 265A that are partially waiting for installation. This secondary heater would be (almost) solely using excess dc line power available. Investigating more it seems that secondary ve.bus (single mp2 5kVA etc) system connected with mk3-usb to cerbo could be the way to go. I think it should be able to switch secondary ve.bus unit on/off with node-red. There could be like few on/off cycles in a day, so probably no need for any contactor with that one. Maybe just manual switch if implementing some other parallel use cases. This approach would be better for system/battery charging/discharging behavior compared to simple dc load?

I wonder if the cerbos computing power could come an issue. 6 mppts are ve.direct versions. I’m pretty new with node-red, should i perhaps limit the “control loop”/flow interval compared to for example measured gridpower output? One of the mppt units would be for somewhat tracked pv field. Also thinking of utilizing cerbo/maybe gx-io extender for tracking control unit. Thread subject is getting bit off.