Get Inspired Cost Control

Cost Control Through Predictive Maintenance

The challenge: Getting insight into cost savings when implementing predictive maintenance

Having access to critical business resources when necessary is crucial to improve productivity and deliver quality customer service. However, unexpected downtime is a recurring issue that keeps maintenance teams up at night and costs the world’s top companies billions in revenue every year. Downtime costs include loss in staff productivity, time spent repairing customer trust due to delayed deliveries, and unexpected repair costs for equipment. Costs escalate further if downtime in one factory impacts a company’s total production flow. Preventing machine breakdown is crucial, especially in an increasingly automated industry. For a global producer in the food industry, machine breakdowns are currently handled ad hoc, costing them large amounts of money. To keep these costs under control, a global food producer wanted to know how implementing predictive maintenance on an older production line in one of their largest factories could help.

How we created value: Implementing a predictive maintenance stream to reduce downtime

So far, this food producer has been scheduling maintenance based on input from operators. Preventative maintenance is scheduled in time intervals and is based on the experience of the maintenance planner. We proposed a predictive maintenance stream where sensor data is continuously monitored by a selflearning model and notifications are sent to the maintenance planner when detected asset behavior is anomalous. This enables the planner to take the required maintenance actions and prevent unexpected downtime and costs. To detect anomalous behavior, we fed the model data corresponding to healthy (=non anomalous) behavior. New sensor data deviating from that healthy behavior is indicative of impending failure, triggering a notification for the maintenance planner. The proposed predictive maintenance stream results in an immediate reduction of unexpected breakdowns and enables the global producer to stretch the intervals for preventive maintenance – leading to lower costs and production stability.

Better results: €2.3m in expected annual savings due to predictive maintenance

To determine the potential savings, we analyzed historic data on maintenance jobs with the local reliability engineer. We also defined whether ad hoc maintenance could have been prevented using the predictive model. Our analysis led to a significant amount of high priority jobs that could have been planned in advance, resulting in expected annual savings of:

■ 27% (€ 2.1M) in technical downtime
■ 10% (€ 0.2M) in preventive maintenance

Let's Get in Touch and Share Thoughts

We would love to join you in your journey toward data-driven decision making.
Let’s get in touch and share thoughts!