IoT Edge Current-Signal Fault Detection for Mechanical Devices
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional methods for detecting anomalous operation in mechanical devices are often invasive, costly, and inaccurate, leading to false positives and false negatives, making it difficult to identify impending failures in devices like engines and motors.
Innovation Solution
A computer-implemented method using a deep neural network (DNN), frequency spectrum analysis (FSA), and time series classification (TSC) modules processes current signal data from devices to provide precise indications of normal or anomalous operation, reducing false negatives and false positives through edge computation and data science techniques.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional non-invasive detection techniques are used to monitor operating parameters, then the detection process is non-invasive and cost-effective, but the detection accuracy is low leading to false positives and false negatives
Solution Approach 1:
The detection system is segmented into multiple specialized analysis modules (DNN module for temporal pattern recognition, FSA module for frequency domain analysis, TSC module for time-series classification). Each module processes current signal data through different analytical approaches and provides independent indications, which are then combined through a voting gate to achieve high detection accuracy with reduced false positives and negatives.
2Measurement precision
If multiple analysis modules are used to improve detection accuracy, then detection precision increases, but system complexity increases
Solution Approach 1:
The system divides complex detection tasks into specialized modules (DNN, FSA, TSC), each handling specific analytical functions. This segmentation allows each module to be optimized for its specific function while maintaining overall system manageability through modular architecture.
Solution Approach 2:
Multiple analysis modules are merged into a unified detection system where their outputs are combined through a voting gate mechanism. This merging leverages the strengths of each individual module (temporal analysis, frequency analysis, time-series classification) to achieve detection precision that exceeds what any single module could provide alone.
3Measurement precision
If deep learning models are trained specifically for each device type, then detection accuracy for that device type improves, but training data requirements and model development time increase
Solution Approach 1:
Device-specific DNN models are trained in advance using historical current signal data and failure patterns. This preliminary training action allows the models to learn device-specific temporal patterns and failure signatures before deployment, enabling high detection accuracy when the models are deployed without requiring additional training time during operation.
Solution Approach 2:
The system adapts to different device types by changing the trained model parameters and architecture configurations. Each device type has its own pre-trained DNN model with parameters optimized for that specific device's electrical characteristics, allowing the system to maintain high detection accuracy across diverse device types while avoiding retraining during operation.
Data Source
AI summary
Methods, systems, and computer-readable storage media for receiving, by an anomalous operation detection service, current signal data representing a driving current applied to a device over a time period, processing, by an anomalous operation detection service, the current signal data through a deep neural network (DNN) module, a frequency spectrum analysis (FSA) module, and a time series classifier (TSC) module to provide a set of indications, each indication in the set of indications indicating one of normal operation of the device and anomalous operation of the device, processing, by an anomalous operation detection service, the set of indications through a voting gate to provide an output indication, the output indication indicating one of normal operation of the device and anomalous operation of the device, and selectively transmitting one or more of an alert and a message based on the output indication.