Peak generation surplus rule of thumb

[Engineer]

I've had my system for a month or two and we are in prime generation time here in the US, so I should be generating a surplus to use during winter. Anybody have a rule of thumb for how much surplus should be generated? Obviously it depends on ...

• Location
• Panel Orientation
• Weather patterns

One thought would be to use PVwatts, however it predicts 949 kWh generation in May whereas I produced 1065 kWh, in what seemed like a good month (just a few cloudy days early on), so maybe it's just a matter of looking at the ratio between good months to poor months. Like "need to bank a credit equal to 1/2 months energy in the good months to make up in the poor months", or something. However I need to take into account TOU credits, ack.

Any real world examples?

[bcroe]

I've had my system for a month or two and we are in prime generation time here in the US, so I should be generating a surplus to use during winter. Anybody have a rule of thumb for how much surplus should be generated? Obviously it depends on ...

• Location
• Panel Orientation
• Weather patterns

One thought would be to use PVwatts, however it predicts 949 kWh generation in May whereas I produced 1065 kWh, in what seemed like a good month (just a few cloudy days early on), so maybe it's just a matter of looking at the ratio between good months to poor months. Like "need to bank a credit equal to 1/2 months energy in the good months to make up in the poor months", or something. However I need to take into account TOU credits, ack.

Any real world examples?

Research your past consumption, over different weather years. Then you will have a starting point. A problem here is that
the severity of winters varies greatly, so one answer will never fit all. The decision was made to replace propane heat with
electric heat, with a gallon of propane about the heat energy of 27 KWH (resistance heat). Heat pumps will change the
ratio, and also the temperatures will considerably affect the pump. Dec generation here is about 1/3 of June. Bruce Roe

[J.P.M.]

A bit late to the party on this thread, but FWIW, PVWatts is an estimator for long term average annual output. As the help/info screens describe, expecting PVWatts to predict output for short periods will lead to problems. Also, and often, the PVWatts default of 14 % for system losses is too conservative. A lot of users find 10 % or less a # that seems to match system output better when using the hourly output option for known clear days, provided the rest of the input is a fair approximation of reality including amb. temp. and other weather data.