2. Identify strategic assets. Predictive maintenance isn't a practice to be carried out across all assets. Identify your strategic assets — the ones that directly impact revenue. For example, a reactor is strategic if it's essential for making product; its performance and availability affect your output. In addition, consider production throughput to determine to what extent equipment failure would lower revenue. Failure of a highly efficient production line that operates at high throughput may be more tolerable to the business than stopping a production line that struggles to meet throughput requirements. Non-strategic assets often are facilities related, pertaining to the physical building such as lighting, stairwells, etc.
3. Determine best indicators of failure. Failure occurs for different reasons and varies by equipment, environment and operating requirements. A pump handling abrasive slurries may suffer excessive vibration before bearing and seal failure while excessive energy consumption may signal wear problems in another pump. Combining performance history of assets, failure studies and references with intuition based on individual experience, trends and patterns emerge. In addition, strategic assets, because of their importance, may merit monitoring of multiple indicators to minimize production disruptions. Also, watch out for false positives. For example, relating high material usage variances to excessive energy consumption in equipment could be a false positive. In this case, the use of extra energy could stem from poor material or formula quality — and thus wouldn't serve as a leading indicator of an equipment performance problem.
4. Automate analysis. Timely action based on real-time operating data is instrumental to an effective predictive maintenance program. The old method of having staff sort through data is inefficient and may provide an outdated analysis because of the time lag. Plus, manual review and analysis takes staff away from performing maintenance and creates a backlog of activity. Automating the process allows action to be taken on the analysis provided.
Analysis and trending technology can take information and, based on your business conditions and experience, identify issues and trends. Actionable analysis derived from software that includes a trending engine can pinpoint problems, filter false alarms, immediately notify stakeholders, adapt to ever-changing conditions and help drive your asset management practice. In contrast to systems that capture, for example, a slurry pump's real-time performance information and produce reports for an engineer to sift through for answers, a system that supports actionable analysis takes this several steps further. It automatically analyzes pump performance data (e.g., electrical consumption) for predetermined trends over time — such as 10% or more excess energy consumption for more than 60 minutes — and alerts key stakeholders to take action when specific conditions are found. An alert can be in the form of a prescriptive set of steps (e.g., 12-point inspection work plan) pertaining to the pump's condition to guide staff through the diagnosis, repair and restore process.
Furthermore, solutions now can assess the situation in real-time, including identifying stalled work orders and issuing alerts to escalate the matter and ensure work is completed and regulations are satisfied.
5. Measure and refine. It's essential to continually measure and refine your asset management program to achieve better results and ensure it expands to cover additional assets and business processes. You should identify the best opportunities for improvement, monitor the most critical areas, implement enhancements and measure them. Evaluate the impact of process changes across the program, not just at one data point.
With today's leaner supply chains and reduced safety stocks, minimizing time to correct issues and increasing equipment availability are becoming more critical. There're a lot of different approaches to measurement from OEE to MTBF and energy efficiency. There's no single Holy Grail for measurement; each company must find one or a combination of several that best meets its needs.