Protective Device Fault Cause Determination via Neural Network
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current electrical power supply network fault detection systems can identify faults but fail to automatically determine their cause, leading to delayed response times and inappropriate corrective actions, especially in cases of potential theft or natural occurrences.
Innovation Solution
A method utilizing a neural network trained on external data processing systems to analyze time profiles of measured variables, recognizing patterns and assigning probable causes of faults based on predefined criteria, allowing protective devices to generate signals indicating the likely cause of errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If protective devices only detect fault presence without determining cause, then device complexity is reduced, but response time and troubleshooting efficiency deteriorate
Solution Approach 1:
The system performs preliminary analysis by continuously monitoring measured variables and pre-evaluating potential fault causes before a fault actually occurs. The neural network is pre-trained with extensive fault data, enabling it to rapidly identify fault causes when events occur, thereby reducing troubleshooting time without requiring complex real-time analysis infrastructure.
Solution Approach 2:
An external data processing system serves as an intermediary between the protective device and the fault analysis function. This intermediary handles the computationally intensive neural network processing, allowing the protective device itself to remain relatively simple while still achieving advanced fault diagnosis capabilities through the intermediary's analysis of measured variables.
2Productivity
If protective devices determine fault causes using complex analysis, then troubleshooting efficiency is improved, but device complexity increases
Solution Approach 1:
The patent employs an external data processing system as an intermediary to perform the complex neural network analysis. This intermediary handles the computationally intensive tasks of pattern recognition and fault cause determination, allowing the protective device to achieve high troubleshooting efficiency without bearing the full complexity burden itself. The intermediary processes measured variables and returns fault cause information to the protective device.
Solution Approach 2:
The system enables protective devices to automatically determine fault causes without requiring operator intervention for analysis. The neural network autonomously evaluates measured variables, identifies patterns, and determines fault causes, making the system self-sufficient in diagnostic functions and thereby improving troubleshooting efficiency while keeping the user interface simple.
3Speed
If pattern recognition is performed locally in protective devices, then response speed is improved, but computational resource requirements increase
Solution Approach 1:
The external data processing system acts as an intermediary that performs the computationally intensive pattern recognition tasks. This intermediary has access to greater computational resources and can process neural network calculations more efficiently than individual protective devices, thereby reducing the energy consumption at the protective device level while maintaining fast overall response times through optimized intermediary processing.
Data Source
Figure 1~2
Figure 3
Figure 4~5
AI summary
The invention relates to a method for determining the cause of a fault in an electrical power supply network (11, 21a, 21b), in which electrical measured quantities are recorded at at least one measuring point (12) of the power supply network (11, 21a, 21b), and the electrical measured quantities are supplied to at least one protective device (13), which examines the measured quantities with regard to the presence of a fault in the electrical power supply network (11, 21a, 21b) and generates a fault signal in the event of a detected fault.In order to determine not only the detection of a fault in a power supply network but also its cause, it is proposed that, after the detection of a fault, the temporal profiles of the measured variables recorded by the at least one protective device (13) immediately before and/or during the fault are examined with regard to the presence of certain criteria. Possible causes of faults are assigned to the individual criteria and/or combinations of criteria. One or more criteria are identified that are wholly or partially fulfilled by the examined profiles of the measured variables. Depending on the identified criteria and the degree to which they are fulfilled by the profiles, a cause of the fault is determined. The invention also relates to a protective device (13) for carrying out such a method.