Operating a solar power generating system

Patent metadata

Field	Value
Publication number	US-10509868-B2
Application number	US-201615011078-A
Country	US
Kind code	B2
Filing date	Jan 29, 2016
Priority date	Jan 29, 2016
Publication date	Dec 17, 2019
Grant date	Dec 17, 2019

Abstract

A method may include obtaining historical irradiance data for each of a first and second locations, each of the first and the second locations including first and second solar power generating devices respectively, and forecasting irradiance at the first and the second locations as a first and a second forecast respectively. The method may also include determining a first and a second confidence interval of the first and the second forecasts respectively, the first and the second confidence interval based on the first historical irradiance data and the first forecast and the second historical irradiance data and the second forecast respectively, and modeling covariance between the first and the second confidence intervals. Based on the modeled covariance, the method may include developing an aggregated forecast of irradiance, and sending a message to a device to modify power generation.

First claim

What is claimed is:

1. A method of modifying energy output, the method comprising:
generating power at a virtual power plant, the virtual power plant including a first solar power generating device at a first location and a second solar power generating device at a second location;
obtaining first historical irradiance data for the first location and second historical irradiance data for the second location, both the first historical irradiance data and the second historical irradiance data normalized to be independent of solar zenith angle;
determining a weather class of a plurality of weather classes for each location data point in the first historical irradiance data and the second historical irradiance data based on each location data point having a predominant characteristic;
forecasting irradiance at the first location as a first forecast and irradiance at the second location as a second forecast;
determining one or more weather classes of the plurality of weather classes for the first forecast and the second forecast based on a respective characteristic of the first forecast and the second forecast;
determining a first confidence interval of the first forecast and a second confidence interval of the second forecast, the first confidence interval based on the weather class of the first forecast and a weather class of a nearest neighbor location to the first location, the second confidence interval based on the weather class of the second forecast and a weather class of a nearest neighbor location to the second location;
modeling covariance between the first confidence interval and the second confidence interval;
based on the modeled covariance between the first confidence interval and the second confidence interval, developing an aggregated forecast of irradiance aggregating forecasted irradiance at the first location and forecasted irradiance at the second location; and
based on accuracy of the aggregated forecast being above or below a threshold, altering operation of the virtual power plant, which includes increasing output or decreasing output of a non-solar power generating device in the virtual power plant proportional to the aggregated forecast.

2. The method of claim 1 , wherein modeling covariance comprises imposing a cardinality budget constraint on a variable d ij representing a relationship between an error at the first location (i) and an error at the second location (j).

3. The method of claim 2 , wherein modeling covariance further comprises narrowing a boundary of a range of the variable d ij according to Jensen's Inequality.

4. The method of claim 1 , wherein altering operation of the virtual power plant includes one of:
of power production at the virtual power plant, increasing output of at least one of the first solar power generating device and the second solar power generating device; or
of power production at the virtual power plant, decreasing output of at least one of the first solar power generating device and the second solar power generating device.

5. The method of claim 1 , wherein developing the aggregated forecast comprises
determining an aggregated confidence interval based on the modeled covariance and using a continuous ranked probability score.

6. The method of claim 5 , wherein:
the virtual power plant further comprises a third solar-power generating device at a third location; and
after determining that a third confidence interval of a third forecast at the third location is independent of the first confidence interval and the second confidence interval, determining an aggregated confidence interval includes aggregating the third confidence interval with the first confidence interval and the second confidence interval with a covariance between the third location and the first and second locations as zero.

7. The method of claim 1 , wherein:
the virtual power plant includes a first plurality of solar power generating devices at the first location distributed in a first density and a second plurality of solar power generating devices at the second location distributed in a second density; and
the aggregated forecast is based on the first and the second densities that correspond to the first location with the first confidence interval of the first forecast and the second location with the second confidence interval of the second forecast, respectively.

8. A system that modifies solar power generation, the system comprising:
a power-distribution network servicing a first customer with a first solar power generating device at a first location and a second customer with a second solar power generating device at a second location;
a non-solar power generating device that generates a modifiable amount of non-solar power in response to one or more operations; and
a control device configured to perform the one or more operations, including:
obtaining first historical irradiance data for the first location and second historical irradiance data for the second location, both the first historical irradiance data and the second historical irradiance data normalized to be independent of solar zenith angle;
determining a weather class of a plurality of weather classes for each location data point in the first historical irradiance data and the second historical irradiance data based on each location data point having a predominant characteristic;
forecasting irradiance at the first location as a first forecast and at the second location as a second forecast;
determining one or more weather classes of the plurality of weather classes for the first forecast and the second forecast based on a respective characteristic of the first forecast and the second forecast;
determining a first confidence interval of the first forecast and a second confidence interval of the second forecast, the first confidence interval based on the weather class of the first forecast and a weather class of a nearest neighbor location to the first location, the second confidence interval based on the weather class of the second forecast and a weather class of a nearest neighbor location to the second location;
modeling covariance between the first confidence interval and the second confidence interval;
based on the modeled covariance between the first confidence interval and the second confidence interval, developing an aggregated forecast of irradiance aggregating forecasted irradiance at the first location and forecasted irradiance at the second location; and
based on accuracy of the aggregated forecast being above a first threshold or below a second threshold, sending a message that causes the non-solar power generating device to modify power generation proportional to the aggregated forecast.

9. The system of claim 8 , wherein the control device is further configured to acquire additional energy from an energy market or sell additional energy to the energy market proportional to the aggregated forecast.

10. The system of claim 8 , wherein the control device is further configured to change a rate for acquiring or selling energy on an energy market proportional to the aggregated forecast.

11. The system of claim 8 , wherein modeling covariance comprises imposing a cardinality budget constraint on a variable d ij representing a relationship between an error at the first location (i) and an error at the second location (j).

12. The system of claim 11 , wherein modeling covariance further comprises narrowing a boundary of a range of the variable d ij according to Jensen's Inequality.

13. The system of claim 8 , wherein developing the aggregated forecast comprises
determining an aggregated confidence interval based on the modeled

Assignees

• Fujitsu Ltd

Inventors

• He Dawei

Classifications

• G01W1/10 Primary

  Devices for predicting weather conditions (computers per se G06; display devices G09) · CPC title

• G06F2119/06

  Power analysis or power optimisation · CPC title

• G06F30/20

  Design optimisation, verification or simulation (optimisation, verification or simulation of circuit designs G06F30/30) · CPC title

• G01W1/12

  Sunshine duration recorders (measuring intensity of sunshine G01J) · CPC title

• G06F2217/78

  Physics · mapped topic