The true cost of downtime varies by industry, but analyst estimates put the starting line for a 750,000 square foot warehouse at $10,000 per hour and as high as $500,000 per hour. The price of not meeting service level agreements can be steep and add up quickly, especially when supply chain disruption leads to out-of-stock parts and transportation delays. In critical infrastructure sectors like health care, chemicals or food processing and storage, the penalties are just the start of game changing issues. Then there are the safety issues that often occur as workers try to pick up the workload where machinery has failed. Resilience is the word of the year, and your key to ensuring reliable, resilient operations is predictive maintenance (PdM).
Predictive maintenance should not be confused with preventive maintenance (PM). Preventive maintenance is regular servicing of equipment based on a time interval or usage to reduce the likelihood of failure. The problem with this type of maintenance is that you may end up replacing parts more frequently than necessary, which adds to costs — or not often enough, which requires even more costly reactive maintenance or repairs (RM).
Predictive maintenance relies on data-driven analysis of the condition of the equipment as indicated by sensors that detect misalignment, wear, friction, and stress. There are sensors for detecting vibration, temperature, and sound. Even ultrasonic levels of sound can be detected by certain sensors. PdM is performed based on the real-time behavior of the equipment, not on the manufacturer’s expected component lifetime. By monitoring real-time conditions, you can order parts in advance and ensure technicians are prepared and repairs can be made quickly at a time that works within the operating rhythm of the facility.
Predictive maintenance programs have been shown to lead to a 25-30% reduction in maintenance costs and 70–75% reduction in equipment breakdowns, 35–40% decrease in downtime needed to perform maintenance1.
Predictive Maintenance can:
Some questions to ask yourself to determine whether predictive maintenance is right for you.
There are several steps you can take to get started:
Maintenance Skills Upgrade. The maintenance worker of the future will use technology as their primary tool. If you can’t hire the tech skills needed for the role, you may need to develop them, whether that means upskilling your current team or hiring tech savvy workers who can grow into the role through training programs.
Streamline Spare Parts Inventory. A good PdM program will rely on a comprehensive and efficient spares inventory. Work to identify critical parts that must be maintained on site versus those that can be stored centrally or sourced quickly and reliably. A global parts supply chain adds complexity and transportation costs and times must be factored in.
Implement WES Software. A robust warehouse execution system is critical to assist not only with monitoring of IIoT data, but for the intelligence it provides.
Focus on Critical Assets First. Establish which assets are most critical. Establish a baseline of historical data for those assets. Analyze failure modes and make predictions, then use technology to test and validate your hypothesis in the pilot before moving on to other assets.
Don’t Forget Cybersecurity Measures. Much has been written about the security weakness of the Internet of Things (IoT). Just because the technology is in a warehouse or not tied directly to your financial systems doesn’t mean it can’t be leveraged and shouldn’t be protected.
FORTNA can help you get started with offerings including maintenance team training, spare parts planning and fulfillment, WES software and facility health checks to ensure uptime, performance reliability and seamless operations.
Your warehouses are more than storage facilities. Today’s automated distribution and fulfillment centers are complex machines through which the lifeblood of the organization passes. They generate vast quantities of data that, with the right algorithm, can be tapped into for improved visibility and better decision-making. That’s where Predictive Maintenance meets Predictive Analytics. The crossroads where machine learning transforms data into systems insights, such as why the sorter recirc rate has increased, but also informs the business decisions, like where to place inventory and when. These capabilities are nascent, but rapidly being developed. We’ll explore this more in a future post on Predictive Analytics.
<sup>[1]</sup>https://www.fiixsoftware.com/maintenance-strategies/predictive-maintenance