Power Grid Fault Prediction Using IED Event Patterns
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting evolving faults in electric power grids are challenged by the use of filtered data from SCADA systems, making it difficult to accurately predict and respond to potential disturbances.
Innovation Solution
A system utilizing intelligent electronic devices to generate events with time stamps, combined with a machine learning-based predictor unit that analyzes historic event data to forecast disturbance occurrences, enabling early detection of evolving faults through real-time or near-real-time predictions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If SCADA system data is used for fault detection, then data collection is simplified, but measurement precision deteriorates due to filtered data
Solution Approach 1:
The patent segments the data collection process by utilizing multiple independent intelligent electronic devices (IEDs) distributed across the power grid. Each IED independently collects and generates events locally, then these events are aggregated for analysis. This segmentation allows the system to maintain operational simplicity while capturing granular, high-precision data from multiple sources, overcoming the limitation of filtered SCADA data.
Solution Approach 2:
The patent introduces an intermediary layer consisting of intelligent electronic devices that sit between the physical power grid components and the central analysis system. These IEDs act as mediators that capture raw events directly from grid operations and translate them into structured data, preserving measurement precision while maintaining ease of data collection through standardized event interfaces.
2Device complexity
If traditional fault detection methods are used, then system complexity is reduced, but reliability deteriorates due to inability to detect evolving faults
Solution Approach 1:
The patent implements preliminary action by using a machine learning model to predict future disturbance events before they actually occur. The system analyzes historical event patterns and generates predictions about potential faults, allowing operators to take preventive measures. This predictive capability enhances reliability while maintaining relatively simple system architecture by adding only a prediction layer on top of existing event collection infrastructure.
Solution Approach 2:
The patent incorporates feedback mechanisms where predicted disturbance events are continuously compared with actual events, and the machine learning model is retrained using historical event data. This feedback loop improves the reliability of fault detection over time while keeping the system architecture manageable through automated learning processes rather than complex manual rule systems.
3Loss of time
If real-time fault detection is implemented, then response time is improved, but device complexity increases due to additional processing requirements
Solution Approach 1:
The patent implements self-service by enabling the system to automatically process and analyze events without requiring complex external intervention. The machine learning model autonomously predicts disturbances, and the system automatically retrains using historical data. This automation achieves rapid real-time response while controlling complexity by eliminating the need for manual analysis processes and reducing dependency on complex external systems.
Data Source
Figure 1~2
Figure 3~6
Figure 7~10
AI summary
For detecting evolving faults in a electric power grid, a machine learning based model is trained and the trained model used using events with time information as input. An event is generated by a function in an intelligent electronic device based on one or more values measured from the electric power grid, the event indicating at least the function and its output. In the training, a plurality of event patterns that are extracted from event history data are used, an event pattern comprising in occurrence order events preceding within a time span a disturbance event. The disturbance event is an event resulting to an interruption in power supply in the electric power grid. The trained predictor is outputting predictions for occurrence times of disturbance events in the electric power grid, and they are displayed for detecting evolving faults in the electric power grid.