Rotational Equipment Anomaly Detection Without Asset-Specific Calibration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for monitoring rotational equipment, such as deep learning anomaly prediction models, require significant expertise and are unsuitable for low-cost assets deployed in large numbers due to complexity and asset-specific calibration needs.
Innovation Solution
The development of automatic, asset-agnostic binary anomaly detection models that use frequency domain data from rotational equipment to generate anomaly scores and determine alerts, allowing for cost-effective monitoring of multiple assets without human intervention.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If deep learning anomaly prediction models are used for rotational equipment monitoring, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent replaces complex, expensive deep learning models with simpler, computationally efficient models that can be deployed on low-cost hardware. The invention uses lightweight machine learning models or rule-based systems that achieve sufficient anomaly detection accuracy without requiring the computational resources and complexity of deep learning architectures.
Solution Approach 2:
The patent extracts and uses only the essential features from vibration data that are sufficient for anomaly detection, rather than processing complete raw datasets through complex deep learning models. By selecting and focusing on key frequency domain features and statistical parameters, the system achieves effective monitoring with reduced computational complexity.
2Measurement precision
If asset-specific deep learning models are calibrated, then measurement precision is improved, but ease of operation deteriorates
Solution Approach 1:
The patent develops universal anomaly detection models that can be applied across multiple asset types and operating conditions without requiring asset-specific calibration. The models are designed to generalize across different rotational equipment, eliminating the need for complex calibration procedures and subject matter expertise for each individual asset.
Solution Approach 2:
The system performs automatic anomaly detection without requiring manual calibration or adjustment by operators. The model automatically adapts to different assets and conditions through self-learning or pre-trained universal parameters, eliminating the need for human intervention in the calibration process.
3Measurement precision
If subject matter expertise is applied for model calibration, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The system automatically determines optimal detection thresholds and model parameters without requiring subject matter expertise. Through self-calibration routines, automatic parameter optimization, or pre-configured universal thresholds, the system eliminates the need for expert intervention while maintaining high detection accuracy.
Solution Approach 2:
The patent replaces manual expert calibration processes with automated computational algorithms. Instead of relying on human expertise to tune parameters, the system uses automatic optimization algorithms, statistical methods, or machine learning techniques to determine optimal thresholds and parameters programmatically.
4Measurement precision
If complex heuristics are used for threshold selection, then measurement precision is improved, but ease of operation deteriorates
Solution Approach 1:
The patent replaces complex heuristic threshold selection methods with automated algorithmic approaches. Statistical methods, optimization algorithms, or pre-computed thresholds based on historical data replace manual heuristic adjustments, enabling automatic threshold determination that is both accurate and operationally simple.
Solution Approach 2:
The system automatically selects and adjusts detection thresholds without requiring operator intervention or expertise in complex heuristics. Through self-calibration, automatic parameter tuning, or adaptive thresholding algorithms, the system maintains high measurement precision while keeping operations simple and straightforward.
Data Source
AI summary
A method of detecting deviation from an operational state of a device includes obtaining preprocessed data corresponding to data sensed by one or more sensor devices coupled to the device. The method also includes processing the preprocessed data using a trained anomaly detection model to generate an anomaly score. The method also includes processing the anomaly score using an alert generation model to determine whether to generate an alert.