AI Power System Protection for Fault Detection and Mitigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current power system protection methods, particularly those involving inverter-based distributed energy resources, face challenges in reliability, selectivity, sensitivity, security, and adaptability due to the increasing complexity of the power grid, leading to inadequate detection and mitigation of component failures and non-nominal operations.
Innovation Solution
The implementation of a novel framework that incorporates artificial intelligence (AI) and machine learning (ML) techniques into computing devices for real-time data analysis, enabling the creation and updating of power system models to detect faults, predict operational states, and control components dynamically, thereby enhancing protection systems' reliability and adaptability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional protection methods are used in inverter-based distributed energy resources, then device complexity is reduced, but reliability deteriorates due to decreased detection accuracy and inadequate fault mitigation
Solution Approach 1:
The patent replaces traditional mechanical/electrical protection systems with AI-based computational systems. Machine learning models analyze power system data to detect faults and predict failures, substituting conventional protection relays and settings with intelligent algorithms that process electrical signals and generate protection decisions, thereby improving reliability while managing complexity through software-based solutions
Solution Approach 2:
The patent introduces AI/ML models as intermediary components between raw power system data and protection decisions. These models serve as mediators that process sensor data, identify fault patterns, and generate control signals, acting as an intelligent layer that enhances traditional protection systems without requiring complete system replacement
2Adaptability or versatility
If additional protective settings are added to cover broader scenarios, then adaptability improves, but device complexity increases and selectivity deteriorates
Solution Approach 1:
The patent utilizes AI/ML models that dynamically adjust protection parameters based on real-time system conditions and learned patterns. Instead of fixed protective settings, the system changes parameters adaptively through machine learning algorithms that optimize protection characteristics for different operating scenarios, maintaining adaptability without increasing structural complexity
Solution Approach 2:
The patent implements universal AI-based protection algorithms that can handle multiple fault types and operating scenarios through a single integrated system. The machine learning models are trained to recognize various fault patterns and provide appropriate protection responses, eliminating the need for multiple specialized protection settings and improving versatility while reducing complexity
3Measurement precision
If AI and machine learning techniques are implemented for real-time data analysis, then detection precision improves, but use of energy increases due to computational requirements
Solution Approach 1:
The patent applies AI/ML techniques selectively to critical protection functions rather than processing all power system data with full computational power. The system uses machine learning for key fault detection and prediction tasks where precision is most valuable, while relying on conventional methods for less critical monitoring, thereby achieving high detection precision for essential functions without excessive energy consumption across the entire system
Data Source
AI summary
This disclosure provides a novel framework that enables cost-effective, accurate and scalable detection, prediction, and mitigation of component failure or non-nominal operation in a power system. According to an embodiment, a computing device may implement a power system model that reflects the behavior, performance characteristics, and/or operational state of components in a power system. The computing device may create the power system model using historical and real-time data. The computing device may include a sensing module designed to determine at least one performance characteristic of a component of a power system; a processing module configured to apply a power system model to the at least one characteristic to determine at least one operational state of the component; and a control module configured to effect a change in a component of the power system in response to the at least one operational state.