Case Study: Sterlite Technologies achieves 8% optical fiber cost reduction and 24×7 automated production with MathWorks (MATLAB)

Sterlite Minimizes Optical Fiber Manufacturing Cost Using Machine Learning

Sterlite Technologies, a global specialist in optical fiber and cables, faced high scrap rates in its fiber-drawing process because many interdependent variables affect the thin filament formed from glass preforms. To address this, Sterlite used MathWorks tools—primarily MATLAB along with Statistics and Machine Learning Toolbox and Symbolic Math Toolbox—to ingest diverse test and geometric data and develop a machine‑learning approach to predict and reduce scrap.

Using MATLAB, Sterlite’s team combined measurement and geometric data, computed theoretical fiber profiles, and fed engineered datasets into a machine‑learning model; the workflow was packaged with MATLAB Compiler into a lightweight shop‑floor app. MathWorks’ solution enabled 24×7 automated runtime (eliminating more than 14 FTEs), reduced fiberglass costs by 8%, and delivered an operator-friendly end‑to‑end application.

Sterlite Technologies

Mohit Mathur

Sterlite Technologies