Electrical Load Forecasting for Overload-Aware Equipment Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional load forecasting techniques for electrical equipment, such as transformers, are complex and require significant computing power, making it difficult for many types of equipment to accurately estimate future loads, which is essential for effective load management and preventing overload conditions.
Innovation Solution
A method using a processor circuit to predict load parameter values for future times based on machine learning models trained on historical data, allowing for the calculation of overload capability and adjustment of equipment parameters to manage loads effectively, without the need for specialized high-powered computing or external information like holidays or events.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional statistical tools (ARIMA) are used for load forecasting, then prediction accuracy is improved, but device complexity and computing power requirements increase
Solution Approach 1:
The patent extracts and removes complex statistical modeling components (trend decomposition, seasonal adjustment, multiple correlated parameters) from the forecasting system. Instead, it retains only the essential historical load data elements needed for prediction, thereby reducing model complexity while preserving forecasting accuracy through a streamlined machine learning approach
Solution Approach 2:
The patent employs lightweight machine learning models that require minimal computing resources compared to conventional ARIMA implementations. These simplified models can be deployed on devices with limited processing power, making the solution accessible to a broader range of electrical equipment without requiring specialized high-performance computing infrastructure
2Measurement precision
If additional correlated parameters (temperature, seasonality, events) are incorporated, then prediction accuracy is improved, but ease of operation and data requirements worsen
Solution Approach 1:
The patent removes the requirement for collecting and processing multiple external correlated parameters (ambient temperature, seasonal indicators, festival data, sports events). By extracting only the necessary historical load data from the electrical equipment itself, the system simplifies data collection operations while maintaining effective load forecasting capability
Solution Approach 2:
The system uses only data that the electrical equipment generates and stores itself (historical load data), eliminating the need to externally source additional parameters. This self-service approach simplifies operation by requiring no external data collection infrastructure or specialized knowledge about regional cultural events and climate patterns
3Measurement precision
If complex models with multiple components are used, then prediction accuracy is improved, but productivity and implementation speed worsen
Solution Approach 1:
The patent extracts and eliminates time-consuming model components such as trend decomposition, seasonal variation analysis, and random variation separation. By using a simplified machine learning model that directly processes historical load data, the system achieves faster implementation and deployment while maintaining prediction accuracy
Solution Approach 2:
The patent applies a simplified modeling approach that uses only the essential portion of historical data needed for accurate forecasting, without performing exhaustive analysis of all possible data components. This partial action approach accelerates implementation while achieving sufficient prediction accuracy for practical load management decisions
Data Source
AI summary
Embodiments are disclosed for predicting, by a processor circuit, a load parameter value of an electrical equipment for a future time based on at least one machine learning model and multiple load parameter values including a set of predefined number of load parameter values extracted from a time series data stream of load parameter values obtained for the electrical equipment. Thereafter calculating, by the processor circuit, an overload capability for the future time based on the predicted load parameter value and changing, by the processor circuit, at least one parameter associated with the electrical equipment at the present time based on the calculated overload capability for the future time.