Related Experiment Videos
Hierarchical Residual Attribution with Supervised Channel Shortlisting for Sensor-Local and Process Fault Diagnosis
Yuchen Wang1, Xinran Lu1, Jun Wang1
1College of Electronics and Information Engineering, Sichuan University, Chengdu 610065, China.
Sensors (Basel, Switzerland)
|August 13, 2026
Summary
This study introduces Hierarchical Residual Attribution with supervised channel shortlisting (HRA-SL) for pinpointing sensor faults. HRA-SL effectively diagnoses process-level events and localizes affected sensor channels, improving system maintenance.
Area of Science:
- Multivariate data analysis
- Sensor network diagnostics
- Machine learning for industrial systems
Background:
- Binary anomaly alarms in sensor systems lack specificity for maintenance.
- Distinguishing between sensor channel issues and process-level events is crucial for effective fault diagnosis.
- Existing methods struggle with precise localization of faulty sensor channels.
Purpose of the Study:
- To propose and evaluate a novel diagnostic framework, Hierarchical Residual Attribution with supervised channel shortlisting (HRA-SL).
- To enable accurate source-level diagnosis and affected-channel localization in multivariate sensor data.
- To establish a controlled-injection benchmark for sensor fault diagnosis.
Main Methods:
- Developed HRA-SL framework utilizing calibration-based cross-channel consistency residuals.
- Employed a controlled injected-perturbation protocol for evaluation.
- Utilized synthetic, UCI HAR, and UCI Air Quality datasets with split-before-injection protocols.
Main Results:
- HRA + spectral configuration achieved superior source-level macro-F1 scores compared to multiple baselines.
- HRA-SL demonstrated high top-1 affected-channel localization accuracy (e.g., 0.916 on synthetic data).
- HRA-SL performance was competitive with, and in some cases superior to, advanced autoencoder-based methods.
Conclusions:
- HRA-SL provides an effective solution for sensor fault diagnosis and channel localization.
- The framework offers significant improvements over existing methods, particularly in identifying specific faulty channels.
- The established benchmark facilitates reproducible research and aids in pre-deployment evidence gathering for field applications.
Related Concept Videos
01:11Residuals and Least-Squares Property
The vertical distance between the actual value of y and the estimated value of y. In other words, it measures the vertical distance between the actual data point and the predicted point on the line
If the observed data point lies above the line, the residual is positive, and the line underestimates the actual data value for y. If the observed data point lies below the line, the residual is negative, and the line overestimates the actual data value for y.
The process of fitting the best-fit...
If the observed data point lies above the line, the residual is positive, and the line underestimates the actual data value for y. If the observed data point lies below the line, the residual is negative, and the line overestimates the actual data value for y.
The process of fitting the best-fit...
01:04Random Error
Random or indeterminate errors originate from various uncontrollable variables, such as variations in environmental conditions, instrument imperfections, or the inherent variability of the phenomena being measured. Usually, these errors cannot be predicted, estimated, or characterized because their direction and magnitude often vary in magnitude and direction even during consecutive measurements. As a result, they are difficult to eliminate. However, the aggregate effect of these errors can be...