Distributed Solar Forecasting With Regularized Reverse Power Flow Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for forecasting net renewable power generation in distributed solar power systems face challenges in achieving both computational scalability and accuracy, leading to unpredictable reverse power flows through power substations, particularly under varying weather conditions.
Innovation Solution
A method combining meso-scale weather forecasting with parameter regularization techniques, such as LASSO, to correct for overfitting and reduce parameters, creating scalable and accurate power flow models that predict reverse power flows.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing forecasting methods are used to improve accuracy, then forecasting precision improves, but computational complexity increases and scalability deteriorates
Solution Approach 1:
The patent segments the forecasting problem by dividing the geographic distribution area into multiple portions and applying separate power flow models to each portion. This segmentation allows the system to maintain high forecasting accuracy for each local area while reducing overall computational complexity through distributed processing, directly resolving the contradiction between accuracy and computational complexity
Solution Approach 2:
The patent applies parameter regularization techniques that modify the parameters of the forecasting models by correcting for overfitting and reducing parameter correlations. This parameter transformation maintains model accuracy while reducing computational complexity, as regularized models require fewer computational resources while preserving predictive performance
2Measurement precision
If existing forecasting methods are used to improve accuracy, then forecasting precision improves, but computational efficiency deteriorates
Solution Approach 1:
By dividing the distribution area into multiple portions and training separate power flow models for each, the system achieves high forecasting accuracy locally while improving computational efficiency through parallel processing. Each model processes a smaller subset of data, enabling faster training and execution while maintaining overall system-wide accuracy
Solution Approach 2:
Parameter regularization transforms the model parameters to reduce overfitting and correlations, which improves computational efficiency by reducing the computational burden of processing highly correlated features while preserving the accuracy needed for reliable forecasting
3Adaptability or versatility
If parameter regularization is applied to reduce parameters, then scalability improves, but model complexity increases
Solution Approach 1:
The patent applies parameter regularization that transforms and reduces the number of parameters in the forecasting models. This parameter reduction directly improves scalability by enabling the system to handle larger geographic areas and more data points, while the systematic approach to parameter regularization actually simplifies the overall model structure by eliminating redundant parameters
Data Source
AI summary
An example method comprises receiving first historical meso-scale numerical weather predictions (NWP) and power flow information for a geographic distribution area, correcting for overfitting of the historical NWP predictions, reducing parameters in the first historical NWP predictions, training first power flow models using the first reduced, corrected historical NWP predictions and the historical power flow information for all or parts of the first geographic distribution area, receiving current NWP predictions for the first geographic distribution area, applying any number of first power flow models to the current NWP predictions to generate any number of power flow predictions, comparing one or more of the any number of power flow predictions to one or more first thresholds to determine significance of reverse power flows, and generating a first report including at least one prediction of the reverse power flow and identifying the first geographic distribution area.