Five steps to applying predictive maintenance
Ensuring the right work is performed at the right time to minimize downtime on production assets is crucial to productivity, but avoiding costly unplanned maintenance is paramount as well. It can be expensive to remove a piece of equipment from production to perform routine preventive maintenance, only to find a function dictated by the Preventive Maintenance (PM) task was not necessary. Predictive Maintenance (PdM) is based on the operating conditions of the asset and the set parameters of its function, rather than time- or calendar-based PMs.
The following steps are predicated on the assumption that all assets in a facility have been identified within a functioning PM program that uses a computerized maintenance management system (CMMS), which includes all planned and unplanned maintenance activities, labor and material costs accurately.
1. Follow the dollars
First, you need to determine where financial and human resources are expended and itemize the assets that are vital to your process. Concentrate first on assets lacking a spare. Aircraft manufacturing is one of the few industries where multiple redundancies are engineered into a product. Chances are, you have a few backups such as supplementary pumps for critical processes. But no one can afford an idle production line just in case it is needed if something breaks down on the primary line.
Using data extracted from the CMMS, identify the high-cost and/or frequent incidents that are consuming the maintenance budget and generating lost production. Examine 2to 3yrof data to ensure all potential issues are addressed. Include the cost of lost product, whether it’s pieces, cases, gallons, or tons. Many times these costs are neglected in the calculations, and an important piece of the cost picture is lost. Occasionally, the cost of lost product exceeds that of the repair itself. Just because an incident has a low occurrence rate doesn’t mean it should be discounted from a PdM program.
2. Perform analysis
Once you have identified the bad actors by cost/frequency, conduct an analysis to examine the mean time between failure/rate (MTBF/R), and identify whether the mean time is increasing or decreasing. Perhaps you’ve already undertaken some remedial actions and improved the MTBF/R on a particular unit. However, don’t dismiss these incidents even if the number of incidents is decreasing. The time and costs consumed can still be significant and warrant investigation
3. Determine where to use PdM
Some assets are, and will generally be expendable and can be run to failure. A ¼-hp motor on a conveyor segment comes to mind. These motors are likely in stock or can be obtained quickly and easily. However, a 200-hp dc motor on a reversing rougher in a steel mill would be a prime candidate for PdM, as well as its gearbox.
The intent is to identify the assets that would cause a catastrophic impact should they unexpectedly fail. While some assets will immediately come to mind, others may come as a surprise as to their difficulty to detect an impending failure.
To better understand where PdM should be deployed, a failure mode effects analysis (FMEA) exercise should be conducted, with the equipment ranked by their susceptibility of failure based on severity, occurrence, and detection criteria. This will generate the risk priority number on which to prioritize the assets. This requires accurate data from a CMMS to trace past breakdowns.
4. Define the technologies to be used
Not all machines need be included in PdM strategies. However, rotating equipment, gearboxes, and electrical distribution components are prime targets for PdM. Vibration monitoring, oil/wear particle analysis, and thermography are the most prevalent PdM strategies deployed, regardless of the industry. Bearing temperature analysis is also a common tool. The failure that the site is attempting to detect drives the technology employed.
Obviously, there is an initial investment to acquire the needed technology and tools to implement PdM. Prudent research and common sense need to be applied. For example, only one infrared thermography camera needs to be purchased, as it can be used sitewide. The same would be true for vibration monitoring, since a portable unit can be used throughout the facility. Generally, oil analysis is conducted offsite by a firm specializing in tribology, sometimes by the actual oil vendor.
Another factor to remember is training staff in using data collection tools and data analysis. Acquisition of data is meaningless unless there is trained staff to analyze and interpret the results, and actions are taken to correct the potential failures identified. If a site is not going to train staff or use the information obtained, it is better off outsourcing the PdM function.
5. Deploy tools
After the tools have been purchased and the staff trained in their use, you can acquire the initial baseline data and capture the results in the tracking program to plot it against performance standards developed by the site or manufacturer. These results should be continually plotted on an asset deterioration curve. Many software packages are available that will perform this function. Bear in mind, the validity of a deterioration model is based on the accuracy and reliability of the data employed.
When the impending failure has been detected, planning should begin to address the problem at the appropriate time, based on production, business conditions, and the amount of deterioration revealed. A technological strategy for today’s high-tech equipment is a sound approach to maintaining assets with minimal production loss. A PdM program can reduce costs and downtime to undertake inspections, as it uses noninvasive tools and procedures without interrupting production. Assets that are properly monitored and maintained via PdM rarely fail and generate unplanned downtime. PdM has the potential to generate savings of about 30% in maintenance expenditures. Too little or too much maintenance is avoided, freeing staff to pursue other activities such as reducing the backlog or performing capital projects.
What’s not to like about that?
Mike Johnston, CMRP, is a senior consultant for T.A. Cook.