About Workshop
Aim
What Participants Will Learn
-
Understand AI-driven predictive maintenance for legacy substations
-
Learn data integration from SCADA, DNP3, Modbus, and other sources
-
Build anomaly detection models and Remaining Useful Life (RUL) predictions
-
Deploy AI solutions via edge-to-cloud pipelines for real-time monitoring
-
Integrate predictive maintenance with CMMS/APM systems for efficient asset management
-
Apply cybersecurity practices in AI solutions
-
Implement scalable AI-powered maintenance systems for improved efficiency and ROI
Structure
📅 Day 1 – Foundations & Data Enablement
- Assets & failure modes: transformers, breakers, relays, CT/PT, batteries
- Data sources: SCADA/DNP3/Modbus, IED logs, DGA, thermography, vibration, PQ, CMMS
- Data readiness: time sync, drift, missing data, weak labels; feature basics (load cycles, gas ratios, wear indices)
- Governance & safety: change control, sign-off
- Hands-on: Build a unified asset + time-series dataset and a baseline Asset Health Index
📅 Day 2 – Diagnostics, RUL & Risk
- Anomaly detection: adaptive baselines, one-class/autoencoder
- Condition diagnostics: DGA + ML, PD patterns, PQ correlation
- RUL modeling: survival/Weibull, gradient boosting hazards, Bayesian updates
- Explainability & calibration: SHAP, reliability, cost-sensitive metrics
- Hands-on: Train anomaly + RUL models; output risk tiers with calibrated thresholds and a short model card
📅 Day 3 – Deployment & Operations
- Edge→cloud pipeline: ingest, streaming features, serving, drift, rollback
- Human-on-the-loop: triage, suppression rules, escalation, UX
- CMMS/APM integration: alert → work order, spares, outage windows, SLA
- Cybersecurity/compliance; value tracking: avoided failures, MTBF/MTTR, ROI
- Hands-on: Package the pipeline and demo “alert → work order” on a lightweight dashboard
Important Dates
Registration Ends
Workshop Dates
What You Will Gain
Outcomes
-
AI-ready dataset for predictive maintenance of legacy substations
-
AI-driven diagnostics and Remaining Useful Life (RUL) models for asset prioritization
-
Practical experience in anomaly detection, RUL modeling, and risk management
-
Deployment of end-to-end predictive maintenance pipelines (edge-to-cloud)
-
Integration of AI solutions with CMMS/APM systems for efficient work order management
-
Measurable improvements in asset health, operational efficiency, and ROI
-
A production-ready system with human oversight and cybersecurity compliance
Who Should Attend
-
Engineers and maintenance managers in power distribution and substation operations
-
Asset management professionals seeking to enhance predictive maintenance strategies
-
Data scientists interested in applying AI/ML to industrial systems
-
Professionals with a background in electrical engineering, industrial maintenance, or data science
-
Individuals looking to implement AI-driven solutions in legacy substation systems
-
Basic understanding of data handling and machine learning concepts is beneficial, but not mandatory