Course Objective & Scope
Unplanned downtime is expensive, while overly early preventive maintenance wastes parts and labour. This course covers combining sensor streams such as vibration, temperature, pressure and current with maintenance work orders, engineering time- and frequency-domain features, constructing a health index, N-day failure classification and Remaining Useful Life (RUL) regression and survival models.
Asset-level splitting (GroupKFold), rare failure labels, false-alarm costs and maintenance scheduling constraints are emphasised. Participants build an early-warning baseline and a cost-based threshold report on industrial telemetry data.
Who Should Attend?
- Data scientists in manufacturing, energy, transport and field operations
- Reliability, maintenance and process engineers
- Industrial IoT and OT analytics teams
- Prerequisite: Python and basic machine learning experience
Key Takeaways
- Align sensor streams and maintenance logs without asset-level leakage
- Engineer degradation features in the time and frequency domains (RMS, kurtosis, FFT)
- Build horizon-based failure probability and RUL prediction models
- Balance false alarms and missed failures using alert windows and cost matrices
- Connect model outputs to work-order triggering and spare-parts planning
Curriculum
PdM Problem Framing and Data Fusion
- Anomaly, failure classification and RUL formulations
- Aligning sensor telemetry, event logs and maintenance records
Signal and Degradation Feature Engineering
- Rolling statistics, RMS, crest factor and FFT summaries
- Operating-regime normalisation and health indices
Failure Probability and RUL Models
- Warning-window classification with LightGBM
- Weibull/Cox survival and piecewise-linear RUL regression
Asset-Level Validation and Alert Logic
- GroupKFold and temporal backtesting
- Consecutive-alert filtering and lead-time evaluation
Maintenance Decisions and Cost Analysis
- Unplanned downtime vs early maintenance cost matrix
- Early-warning and threshold report for the enterprise scenario
Other Courses in This Category
-
SKU-location demand forecasting covering promotions, price elasticity, stockout censoring adjustment, intermittent demand (Croston/ADIDA), hierarchical reconciliation and probabilistic forecasts tied to safety stock.
- Live Course
- 5 modules
-
Bridging predictive models and mathematical optimisation: LP/MILP formulation, capacity, scheduling, routing and budget/price allocation with OR-Tools and Pyomo, plus stochastic and contextual decision systems under uncertainty.
- Live Course
- 5 modules
-
Mastering tabular gradient boosting with XGBoost, LightGBM and CatBoost: tree-growth strategies, native categorical and missing-value handling, ranking/quantile/custom objectives, monotonic constraints, Optuna and fast inference.
- Live Course
- 5 modules
-
Automated model selection, hyperparameter search and multi-layer stacking with AutoGluon, FLAML and Optuna, covering compute budget management, leakage checks and moving from rapid baselines to production.
- Live Course
- 5 modules
-
Glass-box models (EBM, GAM), TreeSHAP/KernelSHAP, ALE, counterfactual explanations (DiCE), uncertainty quantification and regulatory compliance for producing auditable, actionable explanations in high-stakes decisions.
- Live Course
- 5 modules
-
Moving from correlation to causation with DAGs, propensity scores, Double Machine Learning (EconML/DoWhy), heterogeneous treatment effects (CATE) and Difference-in-Differences to measure price, campaign and policy impact.
- Live Course
- 5 modules
Get in touch
