While the cost of unplanned downtime is clear, most maintenance strategies still do not strike the right balance between planned and unplanned downtime. Under-maintenance leads to unplanned downtime while over-maintenance results in excessive planned downtime.
The key to reducing both planned and unplanned downtime is predictive maintenance. Predictive maintenance is based on actively monitoring the performance and condition of equipment during normal operation and then analyzing that data to predict and prevent equipment failure through corrective maintenance.
Here are a few best practices for using predictive maintenance to reduce downtime in your organization.
1. Conduct a criticality assessment
What machines are most important to your organization? If you don’t know, it’s time to conduct a criticality assessment, which will help you rank your assets from most critical to least critical. Doing so can give you valuable data on how to identify and prioritize your most important assets, improve the reliability of your equipment, and reduce unnecessary inventory.
You will need to assess the assets in your facility from a reliability standpoint. Then choose the appropriate maintenance strategy for all of the different machines and components that are critical to the facility’s production process, as well as any utility and supporting equipment.
The criticality assessment should be part of your overall maintenance strategy, reviewed every 18-24 months, and involve multiple stakeholders at your organization.
2. Choose the right condition monitoring method(s)
Multiple condition monitoring methods exist, each with its own merits. Using the right combination of methods is essential to ensure that you’re getting the right information at the right time to make a decision. There are four major monitoring methods:
Vibration analysisallows a technician to monitor a machine’s vibrations with a handheld analyzer or real-time sensors built into the equipment. Vibration analysis is the most commonly used condition monitoring method, and it should be considered for any critical rotating machinery including high-speed rotating equipment, such as bearings, gears, shafts and free wheels. Ultrasonic testingis performed by listening to the sound characteristics of a component and noting any changes that can provide a warning flag for pending problems and failures. Ultrasonic analysis is primarily used for leak detection and surveys of electrical equipment but can also be used for listening to rotating components, such as bearings. It is not as diagnostically accurate as vibration analysis. Infrared analysisis a non-intrusive testing technology using infrared cameras to detect high temperatures in equipment. Infrared analysis is primarily used to survey electrical panels, wires, and cabinets for inspections. It can be used to pick up hot spots which indicate developing issues for rotating machinery. When it comes to diagnosing issues on rotating equipment, it is not considered to be more effective than vibration analysis. Oil analysis enables a technician to determine if particles or contaminants are present in the oil within a machine. Some tests can also show the viscosity, presence of water or wear metals, and particle counts.Oil analysis can be used to determine fluid properties by evaluating the condition of the lubricant, which could affect oil change intervals. It is typically used to validate other diagnostic conclusions that are obtained with technologies like vibration or ultrasound.
3. Deploy condition monitoring solutions appropriately
Every plant and production line is unique. It’s important to deploy the right mix of condition-monitoring technology based on the results of your specific criticality assessment and the monitoring methods available to you. The information derived from those two factors will help you understand your assets and how to run them for peak efficiency and performance.
AI-driven diagnostics can make this process more efficient by eliminating guesswork for your team. When the AI is purpose-built and trained on a large pool of trusted machine hours, predictive, prescriptive analytics can quickly and accurately identify the root cause triggering an alert and advise on how to fix it. AI-generated insights outline a course of action to prevent machine failure. Instead of drowning in data, maintenance and reliability experts can directly address problems in a timely manner and avoid the impacts of unplanned shutdowns.
4. Implement smart, selective redundancy
A backup system can help prevent downtime. When one production line, clean room or machine is shut down, a second can be activated. Mitigation strategies can be particularly helpful in the pharmaceutical industry, where people can’t afford delays on life-saving medications. These include: multiple monitoring systems on crucial systems, backup air purification units, multiple chillers, and, in very rare cases, even a full backup of production lines.
However, the cost of a backup system can be significant, given that the redundant system must be maintained in working condition and operated periodically to check its functionality. A smarter approach is to choose exactly how and where to implement redundancy for your operations. AI-driven machine health systems provide data and actionable insights allowing you to see which machines need redundancy and which machines don’t. Using this data will help you develop the optimal backup strategy for your operations as well as inform smarter resource allocation and budgeting exercises from capital expenditures to obsolescence planning.
5. Get everyone on board
Predictive maintenance can be an ideal solution for unplanned downtime, but there may be resistance within the company, especially if the value isn’t clearly communicated or understood. To encourage cultural change, a change management program should be thoroughly planned and the value of the new maintenance strategy clearly explained to all stakeholders.
For more tips, download the Guide to Decreasing Downtime eBook