I have seen the remarks about big loads influencing the forecast. But thought that using a known load (evcs ns) should lead to better/dependable load kWh/hour forecasts.
I guess that the algorithm is not open/known. To me it seems that any ‘big load’ or at least the evcs charge is BOTH ‘spread over the 24hrs’ AND ‘input on the known hours’ leading to a prediction that is higher by ‘one historical charge session’.
Is there a way to give it a ‘better expectation per day’ for my known baseload?
