Most companies have discovered to some degree the benefits of condition based maintenance (CBM). By definition, CBM entails performing maintenance tasks, not on a scheduled interval basis (such as operating hours, miles, cycles, etc.) but rather based on data gathered from certain predictive maintenance tasks such as oil analysis, vibration analysis, thermography or ultrasonics. The key benefit to a condition-based maintenance strategy is that maintenance tasks get done only when required, based on data, optimizing the utilization of increasingly scarce maintenance resources. Done properly, there is little doubt that CBM can and does work.
While CBM has proved successful for many companies, all too often they miss the most basic and often the most effective form of predictive data gathering: routine operator-based inspections. While some companies struggle, other companies have jumped in with both feet, anxious to follow the success of Toyota and others that have successfully deployed autonomous maintenance in support of lean manufacturing initiatives. In fact, as many as 43 percent of all U.S.-based manufacturers are actively pursuing lean maintenance initiatives with varying degrees of success.
Just like conventional maintenance practices, autonomous maintenance is predicated on predictive maintenance and, in particular, routine inspections. These routine inspections are usually part of regular operator rounds that are performed each week, day or shift. But even those companies that are having success with operator-assisted maintenance sometimes miss the mark; and while their desires are well-intentioned, their execution of inspections routes is lacking.
From my experience, the main reasons for this are a lack of detail and engineering design up front in defining what needs to be inspected, how it should be inspected and documenting the required inspection routes. Inspections routes need to be specific. It's not sufficient to simply state a task such as "check pump". For some, check pump may mean to feel the housing for temperature. Others may look for signs of a packing leak, while others may simply check yes, indicating the pump is indeed still in the same place it was yesterday!
Making the List
The best inspection lists include clear, concise, task-specific details for each asset. For lubrication, this includes details such as: check shaft seals for signs of oil leakage and check oil level gauge for correct operating oil level. Basic predictive maintenance data should also be collected. This might include the differential pressure gauge reading across the filters, the color and clarity of the oil in the sight glass or the color of the desiccant in the breather.
Check Yes or No
Wherever possible, questions should be stated in a simple yes/no format. For example, "oil appears clean and clear?" with an appropriate check box for yes or no. Training should be provided so that operators can easily determine what cloudy or discolored oil looks like along with basic tools such as an infrared temperature gun to measure bearing temperature or a laser pointer to check for oil clarity.
The inspection routes should be captured in a simple check sheet format, either electronically (preferred) or manually. They should be easily accessible at the machine, but should also require sign-off to avoid pencil-whipping. While a simple sign-off process doesn't necessarily guarantee that the task was performed, human nature is such that we naturally feel more accountable when physically signing a document or form.
While paper-based inspection routes do work, all too often, the extra time required to capture manual route sheets and input into a tracking spreadsheet or database means that good intentions often go astray as more pressing tasks arise. The most successful inspection routes involve electronic data collection in the field, using either a rugged, industrial personal digital assistant (PDA) or tablet personal computer (PC).
Not only does this offer the advantage that the information gathered is already databased, but the PDA or PC can also be used to include simple drawings or pictures indicating where inspection points are located and what normal vs. abnormal conditions look like.
Once databased, a process needs to be put in place to periodically review the data, and where necessary, create appropriate maintenance work orders to correct the problem. There is no easier way to kill an inspection program than to send the message that nobody is looking at the data. In every case where I've seen this happen, the operators simply stopped performing the inspections. None of us likes to feel like we're wasting our time, and operators are no different.
Inspections should be cross-disciplinary. They should include lubrication, mechanical maintenance, electrical, safety and operational inspections. It makes little senses to conduct one survey for lubrication, followed by a similar survey for electrical systems on the same machine. If your plant has different maintenance planners for different maintenance functions (mechanical, electrical, production, etc.), inspections can easily be divided once the information has been gathered. The critical path is getting good data: the rest will fall in place accordingly.
When it comes to lubrication, industry studies reveal that, on average, 30 to 60 percent of all maintenance problems can be traced to lubrication. It is likely that 70 to 80 percent of those could have been avoided simply by looking for the correct oil level or condition, evidence of a seal or breather failure, or the observation that an automatic greasing system is not functioning correctly.
Take advantage of the people who are in front of machines eight to 12 hours a day and involve operators as part of the solution, rather than blaming them as part of the problem.
As always, this is my opinion. I'm interested in yours.