Systems and methods for improving load energy forecasting in the presence of distributed energy resources

1. A method comprising:

(A) accessing, by one or more computers, one or more electronic databases, stored on one or more computer readable media, the one or more databases comprising:

(i) forecasted meteorological conditions data associated with a geographical area encompassing an electric power grid;

(ii) forecasted energy load data associated with energy load on the electric power grid;

(iii) time data comprising at least one of time of day data or season data associated with the electric power grid;

(iv) historical data comprising historical load data, historical forecasted meteorological conditions data, historical forecasted wind and solar energy generated by distributed energy resources made available to the electric power grid, and historical time data corresponding to the historical forecasted meteorological conditions data;

(B) calculating, by the one or more computers, a revised load forecast for the electric power grid based on the forecasted meteorological conditions data, forecasted wind and solar energy generation, the forecasted energy load data, the time data and time-series variables determined based on an analysis of the historical data, wherein the calculating step comprises:

(i) estimating, by the one or more computers, a structural equation of electricity load on the electric power grid based on the historical data, including the historical forecasted wind and solar energy generated by distributed energy resources made available to the electric power grid, where the electricity load is a dependent variable of the structural equation;

(ii) analyzing, by the one or more computers, the structural equation to determine whether the electricity load as the dependent variable requires transformation;

(iii) performing, by the one or more computers, a multivariable fractional polynomial analysis of explanatory variables in the structural equation with the dependent variable transformed or not transformed as determined in step (B)(ii) to determine whether each of the explanatory variables have a linear or non-linear specification;

(iv) performing, by the one or more computers, a time series analysis of the structural equation with the dependent variable transformed or not transformed as determined in step (B)(ii) and with each of the explanatory variables having a linear or non-linear specification as determined in step (B)(iii), wherein the time series analysis comprises use of an autoregressive-moving average (ARMA) model represented by ARMA(p,q), where p is a number of lag lengths for the auto-regressive component of the ARMA model and q is a number of lag lengths for the moving average component of the ARMA model, and values for p and q are 100 or more;

(v) generating, by the one or more computers, a load prediction equation based on the time series analysis of the structural equation with the dependent variables transformed or not transformed as determined in step (B)(ii) and with each of the explanatory variables having a linear or non-linear specification as determined in step (B)(iii); and

(vi) calculating, by the one or more computers, a revised forecast of energy load by inputting the forecasted meteorological conditions data, the forecasted load energy data, the forecasted wind and solar energy generated by distributed energy resources made available to the electric power grid, and the time data to the load prediction equation;

C) providing, by the one or more computers, to an energy management computer system, the revised load forecast; and

D) adjusting levels of generation of conventional energy using the electric power grid based on the provided revised load forecast.

2. The method of claim 1 , wherein the forecasted meteorological conditions comprise one or more of the following: forecasted temperature; forecasted wind speed; forecasted humidity; forecasted solar radiation; forecasted air pressure; forecasted air density, forecasted wind density; forecasted dewpoint; forecasted visibility; forecasted probability of precipitation; and forecasted sky conditions.

3. The method of claim 1 , wherein the load prediction equation is a combination of the structural equation with the dependent variable transformed or not transformed as determined in step (B)(ii), the structural equation with each of the explanatory variables having a linear or non-linear specification as determined in step (B)(iii) and time series variables as determined by the time series analysis as performed in step (B)(iv).

4. The method of claim 1 , wherein step (B)(ii) comprises a Box-Cox analysis.

5. The method of claim 1 , wherein the explanatory variables comprise one or more coefficients and exponents.

6. A system comprising:

one or more data processing apparatus;

a computer-readable medium coupled to the one or more data processing apparatus having instructions stored thereon which, when executed by the one or more data processing apparatus, cause the one or more data apparatus to perform a method comprising:

(A) accessing, by one or more computers, one or more electronic databases, stored on one or more computer readable media, the one or more databases comprising:

(i) forecasted meteorological conditions data associated with a geographical area encompassing an electric power grid;

(ii) forecasted energy load data associated with energy load on the electric power grid;

(iii) time data comprising at least one of time of day data or season data associated with the electric power grid;

(iv) historical data comprising historical load data, historical forecasted meteorological conditions data, historical forecasted wind and solar energy generated by distributed energy resources made available to the electric power grid, and historical time data corresponding to the historical forecasted meteorological conditions data;

(B) calculating, by the one or more computers, a revised load forecast for the electric power grid based on the forecasted meteorological conditions data, forecasted wind and solar energy generation, the forecasted energy load data, the time data and time-series variables determined based on an analysis of the historical data, wherein the calculating step comprises:

(i) estimating, by the one or more computers, a structural equation of electricity load on the electric power grid based on the historical data, including the historical forecasted wind and solar energy generated by distributed energy resources made available to the electric power grid, where the electricity load is a dependent variable of the structural equation;

(ii) analyzing, by the one or more computers, the structural equation to determine whether the electricity load as the dependent variable requires transformation;

(iii) performing, by the one or more computers, a multivariable fractional polynomial analysis of explanatory variables in the structural equation with the dependent variable transformed or not transformed as determined in step (B)(ii) to determine whether each of the explanatory variables have a linear or non-linear specification;

(iv) performing, by the one or more computers, a time series analysis of the structural equation with the dependent variable transformed or not transformed as determined in step (B)(ii) and with each of the explanatory variables having a linear or non-linear specification as determined in step (B)(iii), wherein the time series analysis comprises use of an autoregressive-moving average (ARMA) model represented by ARMA(p,q), where p is a number of lag lengths for the auto-regressive component of the ARMA model and q is a number of lag lengths for the moving average component of the ARMA model, and values for p and q are 100 or more;

(v) generating, by the one or more computers, a load prediction equation based on the time series analysis of the structural equation with the dependent variables transformed or not transformed as determined in step (B)(ii) and with each of the explanatory variables having a linear or non-linear specification as determined in step (B)(iii); and

(vi) calculating, by the one or more computers, a revised forecast of energy load by inputting the forecasted meteorological conditions data, the forecasted load energy data, the forecasted wind and solar energy generated by distributed energy resources made available to the electric power grid, and the time data to the load prediction equation;

C) providing, by the one or more computers, to an energy management computer system, the revised load forecast; and

D) adjusting levels of generation of conventional energy using the electric power grid based on the provided revised load forecast.

7. The system of claim 6 , wherein the forecasted meteorological conditions comprise one or more of the following: forecasted temperature; forecasted wind speed; forecasted humidity; forecasted solar radiation; forecasted air pressure; forecasted air density, forecasted wind density; forecasted dewpoint; forecasted visibility; forecasted probability of precipitation; and forecasted sky conditions.

8. The system of claim 6 , wherein the load prediction equation is a combination of the structural equation with the dependent variable transformed or not transformed as determined in step (B)(ii), the structural equation with each of the explanatory variables having a linear or non-linear specification as determined in step (B)(iii) and time series variables as determined by the time series analysis as performed in step (B)(v).

9. The system of claim 6 , wherein step (B)(ii) comprises a Box-Cox analysis.

10. The system of claim 6 , wherein the explanatory variables comprise one or more coefficients and exponents.