Textile manufacturer leverages machine learning-led predictive maintenance solution to reduce asset downtime by 48%

  • Customer

    Leading global manufacturer for short-staple fibre spinning machines

  • Industry


  • Reduced machinery downtime through predictive maintenance
  • Increased asset utilization and machine availability
  • Enhanced line of sight into plant performance, and forecast potential investments needed for maintenance
  • Identified and analyzed key indicators signaling part failure
Business Situation

Declining asset utilization and turnover due to unplanned downtime of machinery. This required the development of a predictive maintenance solution that can forecast the downtimes and trigger required pre-emptive maintenance activity.

Business Challenges

Each machine in the textile mill had to be fit with a “black box”, which supported more than 40-50 sensor connections. The complexity was compounded by the need to build predictive models for key components first, utilizing data from sensors. Data had to be processed into actionable insights for over 2000+ sensor points on a “near-real time” basis to drive proactive maintenance support.

Building these models called for deep domain expertise – from selecting the most critical sensor parameters (temperature, pressure, proximity, etc.) to understanding data flow across the maintenance systems, work orders and backend ERPs.

Birlasoft Solution

Birlasoft‘s proprietary machine learning-backed predictive modeling solution (intelliFactoryTM) helped the client achieve high levels of asset utilization. The accurate models, built by using real-time data in conjunction with the fault conditions, predicted if a machine was approaching failure. The continuous feedback, based on predetermined business logic, helped refine the models with regards to alarm management, ticket generation, maintenance (planned, unplanned, predictive) and reporting.