Systems and methods of hierarchical forecasting of solar photovoltaic energy production
View Patent ↗A photovoltaic system can include multiple photovoltaic power inverters that convert sunlight to power. An amount of power for each of the inverters can be measured over a period of time. These measurements, along with other data, can be collected. The collected measurements can be used to generate artificial neural networks that predict the output of each inverter based on input parameters. Using these neural networks, the total solar power generation forecast for the photovoltaic system can be predicted.
1. A method comprising:
collecting a plurality of respective average power productions over a period of time from individual ones of a plurality of photovoltaic power inverters of a photovoltaic system;
generating a plurality of neural networks based at least in part on the plurality of respective average power productions over the period of time for the individual ones of the plurality of photovoltaic power inverters, the plurality of neural networks individually corresponding to the plurality of photovoltaic power inverters; and
predicting a total solar power generation forecast for the photovoltaic system based at least in part on the plurality of neural networks.
2. The method of claim 1 , wherein each of the plurality of neural networks comprises at least one hidden layer.
3. The method of claim 1 , wherein the plurality of neural networks are generated based at least in part on a machine learning algorithm.
4. The method of claim 1 , further comprising determining at least one input parameter, wherein the total solar power generation forecast for the photovoltaic system is further based at least in part on the at least one input parameter.
5. The method of claim 4 , wherein the at least one input parameter comprises at least one of: a time and a meteorological forecast.
6. The method of claim 4 , further comprising predicting, for each of the plurality of photovoltaic power inverters, a respective inverter solar power generation forecast using the at least one input parameter, wherein predicting the total solar power generation forecast for the photovoltaic system comprises summing the respective inverter solar power generation forecast for each of the plurality of photovoltaic power inverters.
7. The method of claim 1 , wherein the plurality of respective average power productions individually correspond to a respective 15 minute average of power production.
8. The method of claim 1 , wherein a respective hourly ahead forecast for each of the plurality of respective average power productions corresponds to a respective hourly average of power production.
9. The method of claim 1 , wherein a respective 24 hour ahead forecast for each of the plurality of respective average power productions corresponds to a respective hourly average of power production.
10. The method of claim 1 , further comprising determining a subset of the plurality of photovoltaic power inverters by an analysis of evolving errors, the subset corresponding to ones of the plurality of photovoltaic power inverters with an impact on PV power plant forecasting.
11. A photovoltaic system, comprising:
a plurality of photovoltaic power inverters;
a plurality of sensing devices individually coupled to ones of the plurality of photovoltaic power inverters; and
at least one computing device configured to at least:
collect a plurality of respective average power productions over a period of time from individual ones of the plurality of photovoltaic power inverters;
generate a plurality of neural networks based at least in part on the plurality of respective average power productions over the period of time, the plurality of neural networks individually corresponding to the plurality of photovoltaic power inverters;
predict, for each of the plurality of photovoltaic power inverters, a respective inverter solar power generation forecast based at least in part on the plurality of respective average power productions over the period of time; and
predict a total solar power generation forecast for a photovoltaic system based at least in part on the respective inverter solar power generation forecast.
12. The photovoltaic system of claim 11 , wherein each of the plurality of neural networks comprises at least one hidden layer.
13. The photovoltaic system of claim 11 , wherein the plurality of neural networks are generated based at least in part on a machine learning algorithm.
14. The photovoltaic system of claim 11 , wherein the at least one computing device is further configured to at least determine at least one input parameter, wherein the respective inverter solar power generation forecast for each of the plurality of photovoltaic power inverters is based at least in part on the at least one input parameter.
15. The photovoltaic system of claim 14 , wherein the at least one input parameter comprises at least one of a time or a meteorological forecast.
16. The photovoltaic system of claim 11 , wherein the total solar power generation forecast is predicted by summing the respective inverter solar power generation forecast for each of the plurality of photovoltaic power inverters.
17. The photovoltaic system of claim 11 , wherein an hourly ahead forecasting of respective average power productions corresponds to an hourly average of power production.
18. The photovoltaic system of claim 11 , wherein a 24 hour ahead forecasting of respective average power productions corresponds to an hourly average of power production.
19. The photovoltaic system of claim 11 , wherein the period of time is greater than or equal to at least one of: a week, a month, and a year.