Quantile-based feature extraction improves predictive maintenance accuracy
Quantile-Led Feature Extraction for Multi-Horizon Predictive Maintenance in Industrial Manufacturing Systems
Artificial Intelligence
Summary
Predictive maintenance helps spot machine problems before they cause failures. Usually, feature extraction from sensor data is fixed even if the prediction time changes. The authors introduce a new method that adapts feature extraction based on the forecasting horizon, using quantiles of sensor data distributions. This method works better for both short and long-term predictions, improving accuracy across different timescales.
What this means in practice
- •For industrial maintenance teams: Improve machine failure predictions by using adaptive, horizon-dependent feature extraction from sensor data.
- •For automated manufacturing operators: Optimize scheduling of predictive maintenance across short and long time frames with horizon-specific sensor data features.
Authors
David J Poland, Daniele Ravi, Na Helian
Abstract
In data-driven predictive maintenance (PdM), feature extraction is usually treated as fixed preprocessing: a descriptor set is chosen once and reused while the downstream model or forecasting horizon changes. This paper isolates the representation-learning stage and presents a quantile-led feature-extraction framework based on a dual-stage MLP-QRNN hierarchy. QRNN1 learns a broad ten-quantile conditional distribution for each sensor channel, while skip-connected QRNN2 refines a retained mid-tail quantile set into compact, channel-resolved, distribution-aware features. A fixed thirteen-pipeline ablation spans 1-hour, 70-hour, and 30-day regimes across 72 machines in 9 industrial facilities, with the downstream temporal classifier held fixed within each regime. Increasing the retained mid-tail set from two to four quantiles improves 30- and 60-minute F1-score, reaching 75.92% and 72.44% with attention enabled. The results also show that representations do not transfer reliably beyond their design horizon unless feature capacity, temporal embedding, activation strategy, and sensor breadth are scaled with the forecasting task. The unmodified short-horizon extractor falls to 42.90% F1 at 70 hours, whereas horizon-conditioned extractors reach 60.38% at 70 hours and 79.97% at 30 days. The framework therefore supports treating PdM feature extraction as a horizon-dependent representational stage rather than fixed preprocessing.