Maintenance is often considered a necessary cost that is almost impossible to control. If organizations operate in run-to-fail mode, it is important to keep high levels of spare parts inventory and technicians on standby (or overtime) to jump like firefighters to get production back on line. If organizations perform preventive maintenance tasks based on a calendar, they are not likely performing the maintenance when it is most effective. And to reach the pinnacle of Condition Based Maintenance, organizations need to undertake exhaustive studies to determine what indicates a failure on specific equipment in its current operating environment. Then, connect the anomaly to the appropriate work task to minimize production loss, quality issues, or other collateral damage. The work task should predefine the appropriate labor, parts, material, safety procedures, etc.
An effective maintenance program requires a combination of reactive, preventive and predictive actions mentioned above. However, it is difficult to determine and measure maintenance effectiveness and to be able to analyze which maintenance strategy to use for specific equipment. Consideration must be given to performance measures that encourage the maintenance organization to prioritize and execute work that benefits the enterprise bottom line.
Example: A manufacturing organization wants to measure MTBF (Mean Time Between Failures) and MTTR (Mean Time to Repair) as a benchmark measure of performance across all facilities. The plants are compared based on a plant wide MTBF and MTTR. This doesn’t make any sense because these are not facility-wide maintenance measures. Operations can be measured by production stops and compared to the amount of time to get the production back to desired capacity. However, the maintenance of equipment would be better measured by MTBF and MTTR for each equipment type to reach a solid metric. But each plant would need to have the same equipment and production requirements to use this metric as a source of comparison across the enterprise. MTTR can be skewed by redundant equipment. If there is a backup and production is not impacted, why rush to get the equipment operational at extra cost?
Rather than MTTR, a much better measure would be Mean Cost to Repair (MCTR). This metric includes production or quality loss while out of commission; scrap created by the failure; labor, parts, and services to repair; and other costs to return to desired production levels. This could be compared to MTBF to create a ratio that provides a full picture of the frequency and cost of reactive repairs that could be further segregated by equipment type. The plant wide ratio could be used to benchmark plants and to provide a comparative measure that shows whether each facility focuses on the highest impact equipment failures and repairing them in the most cost effective way.
It is also critical to measure the failures at the right level in the equipment hierarchy. If work orders are executed at the failed equipment level, this provides a basis for sound analysis. Too high up the hierarchy and the data is unusable. For example, work orders executed at the unit level for piping repairs in the example below would not have appeared in the analysis for piping. The planner is responsible for allocating the work to the appropriate equipment when planning the work tasks.
The graph below shows a breakout of MCTR by equipment type and a further expansion into the specific costs that make up the total cost to repair.
Mean Cost to Repair by Type
By evaluating the ratio of MCTR over the MTBF, you get a good picture of the problem equipment. The high cost and high frequency failures will create a higher ratio. In this case, the Boiler has a lower cost to repair but it fails frequently. The motors have a high MCTR but failures are not very frequent. The piping rarely fails making the ratio very small. See chart below.
Preventive Maintenance should not be included in these numbers for reactive maintenance / failure evaluation. The 9.7 ratio could be used as a plant wide measure of maintenance effectiveness to compare to other plants. The ratio will provide positive feedback for longer MTBF (equipment runs longer) which indicates that equipment is properly maintained and operated correctly. The ratio also provides positive feedback for reduced cost to repair. MCTR is reduced when production loss is minimized; labor is not wasted; overtime is reduced, unused parts are returned for credit back to the work order; contractors are used only as long as necessary; and in general waste is minimized!
The collection and analysis of MCTR is not difficult with an effective Enterprise Asset Management system like Avantis. The metrics will show where you need the most help. Your experience in the maintenance and operations organizations will be essential to decide what actions to take to optimize the business based on these Key Performance Indicators.
Curtis Wilson has been with Schneider Electric (previously Invensys) for the last 5 years as a Principal Asset Management Consultant. He has been providing reliability/maintenance process and technology improvement consulting for global companies for more than 25 years. He is a Six Sigma Black Belt and Reliability Centered Maintenance Practitioner providing extensive background in structured process and continuous improvement analysis as the basis for technology improvements.