PR-DT0522-1 - 2022 Digital Transformations, May

Building a Scalable Intelligent System to Advise Predictive Maintenance Operations in a Steel Mill

A. Sahu, R. Chahar, S. Olivar, R. Balasubramanian, A. Gupta, et al.

Download PDF

View Sample Page

Abstract:
The continual increase in demand for high-quality products has spurred the growth of modern factories. One of the key requirements to meet this rapid demand is to optimize operational efficiency through reduced downtime. This is evident from a study by the International Society of Automation, which has shown that unplanned downtime across all industry segments is estimated to cost US$647 billion per year. As a conservative estimate, an hour of downtime in a steel mill could lead to ~US$50,000 in lost revenues and a typical steel mill has unplanned downtime upwards of 500 hours in a year. One major driver behind this is related to the maintenance of industrial equipment. The traditional approaches to maintenance are mostly reactive or periodic in nature based on component usage. The reactive maintenance approach performs maintenance operations only when equipment fails. While this doesn’t entail intermittent equipment downtime, it can lead to expensive capital expenditure and unexpected downtime. The scheduled (or periodic) maintenance approach performs maintenance operations at a regular cadence. This can contribute to extra costs in terms of frequent maintenance interventions that may not be relevant.


Keywords: digital transformation, industry 4.0