Current Supervisory Circuit for Predictive Fault Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing current supervisory devices primarily rely on threshold-based detection for overcurrent events, which are inefficient in predicting or detecting faults promptly, lacking advanced functionality for timely intervention.
Innovation Solution
A current supervisory device that utilizes a processing circuit to analyze a time series of current values through functions such as difference, derivative, integral, frequency representation, and correlation, generating an event signal for quicker fault detection and prediction, optionally using a neural network for enhanced accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If threshold-based detection is used for overcurrent events, then the device structure is simple, but the fault detection speed and prediction capability are insufficient
Solution Approach 1:
The system performs preliminary actions by storing multiple historical current values and pre-defining multiple analysis functions (difference, derivative, integral, frequency domain, envelope, correlation) before faults occur. When a fault event is detected, the processing circuit can immediately apply the appropriate function to the stored data subset, enabling rapid fault detection and prediction without complex real-time calculations.
Solution Approach 2:
The system dynamically selects different analysis functions based on the type of fault event detected. The processing circuit receives a subset of stored current values and applies appropriate functions from the predefined set (difference for rate of change, derivative for instantaneous rate, integral for accumulated energy, frequency domain for oscillation patterns, envelope for modulation detection, correlation for pattern matching). This dynamic function selection enables adaptive fault detection while maintaining a relatively simple overall device structure.
2Measurement precision
If multiple current values are analyzed using various functions, then the fault detection accuracy is improved, but the processing complexity increases
Solution Approach 1:
The system segments the fault detection task by dividing it into multiple independent analysis functions, each targeting specific fault characteristics. The processing circuit receives a subset of stored current values and applies different functions (difference, derivative, integral, frequency domain, envelope, correlation) to detect different types of anomalies. This segmentation allows accurate fault detection through specialized analysis while keeping each processing module relatively simple and manageable.
Solution Approach 2:
The system changes analysis parameters by applying different mathematical functions to the same set of current values. Instead of increasing hardware complexity, the processing circuit varies the computational approach (different functions) to extract different characteristics from the stored current data. This parameter-based differentiation enables accurate detection of various fault types using the same hardware infrastructure.
3Reliability
If a neural network is used for current value analysis, then the prediction accuracy is enhanced, but the device complexity and computational requirements increase
Solution Approach 1:
The system uses an intermediary approach by first applying simpler mathematical functions (difference, derivative, integral, frequency domain, envelope, correlation) to preprocess the current values and extract meaningful features. These processed features then serve as input to the neural network, which performs the final prediction. This intermediary preprocessing step reduces the computational burden on the neural network while maintaining high prediction accuracy, and allows the system to achieve enhanced reliability without excessive computational complexity.
Data Source
AI summary
The present invention relates to a current sensor device comprising: current sensing means for sensing a current, a circuit arranged to convert a signal received from the current sensing means into a signal indicative of the sensed current, storage means for storing a plurality of values of the signal indicative of the sensed current, a processing circuit arranged for receiving a subset of the plurality of values stored in the storage means, for detecting or predicting an event based on a function of the subset, said function being stored in the storage means, and for generating a corresponding event signal, said current supervisory device further comprising an interface circuit arranged to output the corresponding event signal.