Machine Learning Sensor Failover for Equipment RUL Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems struggle to accurately identify actionable data points for industrial machines operating under varying conditions, particularly in cases of sensor failure or communication breakdown, leading to inaccurate insights and increased maintenance challenges.
Innovation Solution
Implementing machine-learning (ML) techniques to assist in failover and relabeling of sensor signaling channels, using neural networks to generate predicted values and embeddings, and incrementally retrain models to minimize false positives and negatives, enabling event prediction and condition monitoring.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If machine learning models are used to predict sensor values during sensor failure, then the accuracy of identifying machine operating conditions is improved, but the complexity of the system increases
Solution Approach 1:
A machine learning model is introduced as an intermediary component between failed sensors and the condition monitoring system. The model receives data from functional sensors and generates predicted values for failed sensors, enabling the system to maintain accurate condition identification without requiring direct input from all original sensors.
Solution Approach 2:
The machine learning model creates a virtual copy or surrogate for the failed sensor by generating predicted sensor values based on patterns learned from historical data and relationships with other sensors. This allows the system to continue operating with accurate condition monitoring despite the physical sensor failure.
2Reliability
If incremental retraining of machine learning models is performed to minimize false positives and negatives, then the reliability of event prediction is improved, but the loss of time for model maintenance increases
Solution Approach 1:
The machine learning model is pre-trained on comprehensive historical data covering various operating conditions and failure modes before deployment. This preliminary training establishes a strong baseline that reduces the frequency and duration of subsequent retraining operations, as the model already possesses general knowledge of normal and abnormal conditions.
Solution Approach 2:
The system implements a feedback mechanism where prediction results are continuously monitored for false positives and negatives. When performance degradation is detected, the model is automatically retrained using recent data, creating a closed-loop system that progressively improves reliability while minimizing unnecessary retraining operations.
Data Source
AI summary
Techniques are disclosed herein for machine-learning (ML)-assisted event prediction for industrial machines. Sensor data for an industrial machine can be modified by generating imputed values. A trained neural network can be executed on the modified sensor data to generate classifier tags for the modified sensor data. The system can generate a binding between the sensor data and the classifier, and generate a notification based on the classifier. The notification can relate to or include a predicted failure, anomaly, usage profile, or remaining useful life estimate for the industrial machine. The system can also generate additional training data to improve predictive capacity of the trained neural network. The additional training data can include additional classifiers determined using sensor signaling channel information or sensor data (e.g., using payload values from sensor signals, metadata values from sensor signals, or metadata associated with a particular sensor signaling channel).