Predictive maintenance & quality solutions provide early warning to what's likely to occur so you can act now to minimize the impact that would have otherwise resulted without this advance insight. There's certainly a premium knowing now what's very likely going to happen in the future. In this second of two posts as part of the AnalyticsZone's newly launched PMQ Blog series, Greg Milwid from the Predictive Maintenance and Quality solutions at IBM, discusses PMQ's pervasive impact focusing on quality. In his previous post, Greg paid special attention to asset failure.
Read Greg's post now:
Predictive analytics is all about identifying an event before it may occur and Predictive Maintenance does just that. By analyzing machine sensor data, Predictive Maintenance helps organizations identify potential asset failures before they may occur and take preventative action. The ability to predict and prevent a potential breakdown has a widespread impact on the business, ranging from cost reduction and optimization to improving customer satisfaction.
We typically think of Predictive Maintenance being relevant to the manufacturing industry, yet there is a groundswell of interest from many other asset-intensive industries. The cost of unscheduled maintenance is driving significant tangible value to adopters of predictive analytics, whether they are trying to prevent unplanned plant and equipment outages, optimize planned maintenance, or reduce scrap and manufacturing defects.
Some sectors are even seeing new horizons opening up to them as a result of these predictive capabilities as they expand their revenue opportunities through new business in equipment servicing. We are seeing this approach being explored as manufacturers look to optimize their businesses around maintaining equipment as varied as aircraft and their engines, mining equipment, farm and construction equipment, and commercial trucks and automobiles.
IBM’s approach to Predictive Maintenance and Quality offers customers two primary use cases: “Asset Failure” and “Quality.” This post will cover Quality. (Click here to read Greg's earlier post on Asset Failure.)
From a Quality perspective, the issue is often that the value of traditional quality early warnings may be limited because Statistical Process Control (SPC) is restricted in its ability to provide useful early detection of subtle problems.
The question posed when trying to predict quality issues is: “Has anything changed enough to require action?” When quality control systems on the production line detect problems, does it indicate fault in only specific items that failed the test or in the entire batch? Commissioning tests to answer this question can be time-consuming and expensive.
IBM implemented Predictive Maintenance and Quality technology in its own manufacturing process to identify faulty patterns and predict outcomes with goal of minimizing inspection costs and improving efficiency of production.
The results have been huge: 97% fault recognition for one specific operation potentially avoids hundreds of thousands of dollars in total costs; 150% ROI expected from fault pattern recognition analytics and 160% ROI expected from improving product quality by controlling humidity. For more information on the implementation of this solution, please see the IBM Bromont case study.
To learn more about Predictive Maintenance and Quality, please join us on June 10th in Indianapolis for “A Taste of Analytics: Discover the Business Value of Predictive Maintenance” events and attend our June 25th webinar, which will cover two Predictive Maintenance case studies – Daimler and IBM Bromont.
I invite you to keep following this blog as colleagues and guest writers dig deeper into various industry applications of Predictive Maintenance and Quality and show you why Predictive Analytics is creating such interest in the asset intensive industries.