question

You turned on the DESS between 11-12 AM yesterday. From the screenshot, it is not visible when it was taken.

That does change ‘the analysis’ over the situation because if it is little later in the day, then DESS looks at the next day and sees that there is even lower price that is coming up. So the battery is better left empty.

Thank you for explanation. Great work on flow anyway :)

yes I have turned on chargin manually for that hour, thats why it was charged.

But maybe I have misunderstood what the Battery cost means (as I posted question here: Dynamic ESS - configuration - Victron Community (victronenergy.com))

Is the cost meant to include cost of purchase price and expected life? And for default value how did you estimate that? As purchase cost / (capacity * number of expected full cycles ) ?

Tried to set battery charge/discharge price to 0.07 so quite lower than before. But still no charging even today. Slowly SOC is rising but only as part of idle oscilation.

Discharging to grid in the evening - I can understand it.

Maybe suggestion from my side or feature request - I think it would be great if there would be additional value (threshhold) of SOC in battery which will take priority. Like you want to have SOC at 70 % preferably (for grid outage). And there would be 2 different modes of operations under threshhold and over threshhold:

Over threshold - like current operation (selling, buying, iddle state, etc.)

Under threshold - setup gridpoint at such value that battery would reach threshold during calculation timeframe (timeframe there are existing published byu/sell prices and PV Yield estimations). Sell battery to grid/buy from grid or use battery for consumption only if comparison price values are double the normal operation values (for example).

I have made additional test. I think there is some issue with Grid setpoint. Currently I have disabled flow. I have checked during negative price how system works:

Enable grid feed in is false, which is correct.

Scheduled charge for 10 minutes is in place.

Grid setpoint is -9900. And system feeds maximum into grid. So it seems grid setpoint has priority over grid feed in. I have tried to leave flow active, but changed grid setpoint to 0 and it worked as intended until flow rewrote setopint back to -9900.

I dont know how it works internally, but maybe it would also solve that idle battery state. So if grid setpoint would me more dynamic:

1. If battery is full, grid setpoint would be minimum.
2. f batery is under SOC threshhold, but grid feed in is true, grid set point would be set to value min+(SUM(Pv yield - consumption)/number of all known hours in sequence ) - (target SOC in kWh -current SOC in kWh)/number of all known hours in sequence) and set battery to charge
3. if battery is full and grid feed in is false, works normally according to prices
4. if battery is under threshhold soc and feed in is false, compare prices to buy from grid to fill in battery is buy prices are <0 so setup grid setpoint to value similar to point 2, but to get energy on top of consumption based on estimated yield into battery

battery should take in account maximum charge and discharde values. For additional info - current negative price is with broker fee. Without fee, price is positive. Not sure if its important or not.

Dirk-Jan, I have made additional research :)

I have changed battery cost to 0.0870 - (purchase cost)/(8000 cycles * capacity in kWh).

Also I have changed mode to Optimized (With BatteryLife) (active SOC was 65%). In this setup it was slightly better because during discharge into grid, it stopped at active SOC

Battery SOc was very slowly increaing durign whole day, probably as result of high excess, iddle state and balancing peaks.

In the evening, operation was again strange - there was buy price 0.35 in 21-22 period (total buy price), total sell price was like 0.05-0.1 (or something like that). System still used grid for consumption and battery was not used. Even if I take battery cost * 2(charge + discharge) its 0.174. Buy price was 0.35 but system still bought from grid - not sure why.

Today for the test I have change battery cost to 0.04 (your default value). According to visualised graph it seems as the pallned operation is closes to expected - system plans to charge battery at 13 (PY yield still high, sell price very low) and discharde at 19-20. (sell price high)

Maybe isnt there anything missing in calculation formula? Like it works mostly only for default value of 0.04 and as you go further from that price it doesnt work too well.