Electric Grid Load Forecasting With Linear Autoregressive Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing prediction methods for electric energy consumption in electric grids, particularly those based on linear regression analysis, often provide poor reliability and accuracy, and are unsuitable for computing systems with limited computational and data storage resources, such as Edge computing architectures.
Innovation Solution
A prediction method using a linear auto-regressive mathematical model that processes both endogenous and exogenous input values, including periodic functions to approximate energy consumption patterns over different time windows, allowing for accurate predictions even with limited resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If machine learning-based prediction methods are used, then prediction accuracy is improved, but computational resource requirements increase
Solution Approach 1:
The patent replaces complex, resource-intensive machine learning models with simpler, computationally efficient mathematical models (ARIMA, exponential smoothing) that consume fewer computational resources while maintaining acceptable prediction accuracy for edge computing environments with limited capabilities
Solution Approach 2:
The patent changes the fundamental parameters of the prediction approach by transitioning from data-intensive machine learning algorithms to parameter-based mathematical models that rely on statistical parameters and historical patterns, thereby reducing computational complexity while preserving prediction functionality
2Measurement precision
If machine learning-based prediction methods are used, then prediction accuracy is improved, but data storage requirements increase
Solution Approach 1:
The patent adopts lightweight mathematical models that require minimal data storage compared to machine learning models, using only essential historical consumption data and calendar information to generate accurate predictions in resource-constrained edge computing systems
3Device complexity
If simpler prediction methods are used, then computational resource requirements are reduced, but prediction accuracy deteriorates
Solution Approach 1:
The patent segments the prediction problem into multiple components handled by different mathematical models (trend component, seasonal component, residual component), allowing each model to specialize in specific aspects of energy consumption patterns and achieve high accuracy without requiring computationally intensive unified machine learning approaches
Solution Approach 2:
The patent introduces calendar data as an intermediary element that bridges simple mathematical models with complex energy consumption patterns, enabling accurate predictions by incorporating temporal patterns, holidays, and seasonal variations without requiring complex machine learning algorithms
4Ease of manufacture
If linear regression-based prediction methods are used, then ease of implementation is improved, but prediction accuracy deteriorates
Solution Approach 1:
The patent creates a composite prediction system that combines multiple mathematical models (ARIMA for autocorrelation, exponential smoothing for trend, regression for external factors) into a unified framework, achieving high prediction accuracy while maintaining relative ease of implementation through modular model integration
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method for predicting electric energy consumption in an electric grid, which employs a linear auto-regressive model to calculate prediction data related to the electric energy consumption in an electric grid. The prediction method ensures high level performances in terms of prediction accuracy and it can be easily implemented even when limited computational and data storage resources are available.