Utility Load Forecasting with 3D Weather Adjustment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current weather normalization methods for electrical load forecasting in utility systems are inadequate, often resulting in poor fits between temperature and load data, with R-squared values typically ranging from 15% to 40%, failing to accurately represent electrical demand sensitivity over periods of weeks, months, or years.
Innovation Solution
A method utilizing a three-dimensional topological surface analysis to generate a polynomial equation that expresses resource usage as a function of hour of day and temperature, allowing for more accurate predictive values by plotting resource usage against these variables, enabling better weather adjustment and load forecasting.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional weather normalization methods are used to forecast electrical load, then the forecasting process is simple, but the accuracy is poor with R-squared values typically ranging from 15% to 40%
Solution Approach 1:
The patent transitions from traditional two-dimensional weather normalization (temperature vs. load) to a three-dimensional topological surface analysis that incorporates hour of day, recorded temperature, and resource usage. This dimensional expansion allows the model to capture the temporal and thermal dynamics of electrical demand more comprehensively, improving R-squared values from 15-40% to significantly higher accuracy while maintaining computational feasibility through polynomial surface fitting.
2Loss of information
If peak hour data is used for grid planning, then the planning process is straightforward, but it fails to capture detailed usage patterns on different equipment
Solution Approach 1:
The patent segments the aggregated peak hour data into equipment-specific usage patterns by analyzing load data across multiple transformers, feeder lines, and customer transformers individually. The three-dimensional topological surface is generated separately for each equipment component, capturing unique thermal and temporal characteristics. This segmentation reveals detailed usage patterns that were previously hidden in aggregate data, enabling targeted infrastructure planning without overwhelming complexity through systematic data organization.
3Measurement precision
If cooling degree days and heating degree days are used to represent heat energy, then the relationship between temperature and load is simplified, but the accuracy in representing actual electrical demand sensitivity is reduced
Solution Approach 1:
The patent transforms the traditional degree day parameters into a continuous three-dimensional topological surface where temperature and hour of day are independent variables and resource usage is the dependent variable. This parameter transformation replaces the simplified degree day metric with a more comprehensive polynomial function that captures non-linear relationships and temporal variations in electrical demand sensitivity to weather conditions, achieving superior accuracy while managing complexity through mathematical modeling.
Data Source
AI summary
A method for weather adjustment of an electrical utility system. Data defining sequential events in the utility system is obtained, each event identified by three coordinate values: hour of day, recorded temperature, and resource usage. A three-dimensional topological surface is generated from the coordinate value, by plotting the resource usage against the hour and temperature in a three-dimensional space. A polynomial equation having calculated coefficients and a highest degree of six for each variable is generated to define the topological surface, wherein the polynomial equation expresses the resource usage as a tenth-degree polynomial function of hour of day and temperature. A future load on components of the system at a particular temperature and hour of day is determined by applying the coefficients to the coordinate values for the particular temperature and hour, and modifying the functionality of components of the electrical utility system based on the determined future load.