Azure based Predictive Maintenance: 10wk PoC

This contextual proof of concept implementation will be developed to provide predictive maintenance insights using Azure Advanced Analytics Solutions customised to a native set-up of inter-connected devices and/or sensor networks for live industrial scenarios.

We will implement core Azure capabilities using Azure ML, Azure IoT Hub and Azure Logic Apps to develop a customised predictive maintenance solution.

Benefits:

• PRE-EMPTIVE MAINTENANCE SCHEDULING:
• REDUCE OPERATIONAL COSTS: Lower maintenance frequency and higher uptime for assets will result in lower operational cost
• DEMAND ESTIMATION OF CRITICAL SPARE PARTS: Optimized capital expenditures by improving asset component replacement decisions by analysing asset operational conditions and predicting spare part demand patterns
• RISK BASED PRODUCTION PLANNING: The estimated asset deterioration can serve as an input to the production planning systems to improve the performance of an asset or phase out assets with high failure risks

Agenda:

• Week 1: Data Understanding & Data Dictionary Creation
• Week 2: Analytical Dataset Creation Using Sensor Log Data
• Week 3: Data Sanitization
• Week 4: EDA & Feature Engineering
• Week 5: Predictive Maintenance Model Development & Iterations
• Week 6: Model Validation & Testing
• Week 7: UX/BI Design - Mock Up
• Week 8: BI Development - Power BI
• Week 9-10: Deployment

Outcome:

Solution Framework which analyses historical data collected from the IoT devices to predict when an asset is likely to fail and identify the type of failure for an asset in order to initiate preventive maintenance actions for various products and services across industries.