Context
Hindustan Zinc — part of Vedanta Resources — runs some of the largest zinc operations in the world, and its compressors are the kind of assets whose downtime is measured in production, not inconvenience. Inside the Business Excellence Office’s Center of Excellence in BI, our principal built the reliability layer.
The problem
Calendar-based maintenance overhauls healthy machines and misses degrading ones. The office wanted condition-based decisions — but condition lived in scattered sensor streams nobody had shaped into an answer to the only question that matters: how long until this machine fails?
What we built
A real-time reliability module — a working digital twin of the compressor fleet — ingesting sensor telemetry and computing remaining useful life for every compressor. RUL turned raw vibration and process data into a number a planner can schedule against.
Around it: KPI-focused BI dashboards that maintenance and overhaul teams actually used, and an automated cost-benefit report translating reliability decisions into financial terms — capturing opportunity costs of up to ₹40 lakh a year.
Key decisions
Leading with money, not ML. The twin earned its keep because every recommendation arrived with a rupee figure attached — the difference between an analytics demo and a tool that changes planning meetings.
Outcome
Condition-based maintenance for the compressor fleet, quantified savings, and the template for JES’s digital-twin practice: telemetry in, defensible operational decisions out.