Making PdM work in the cloud

Sometimes PdM and condition monitoring seem like a giant puzzle with missing pieces. To take the necessary readings we need the right measurement tools in the right, properly trained, hands. Moreover we need maintenance and reliability planning that spells out what needs to be measured, when and how. We also need maintenance organizations with the skill and will to execute effective measurement programs regardless of pressure from competing activities.

We don’t just need most of the things on this list to get our readings. We need them all. If the plan is missing, there will be missing readings and unmonitored machines. If enough training is not in place, the readings won’t happen, or the data quality will suffer. If the PdM program buckles to demands from emergency work or other priorities, there goes the program. Or if the people who know how to get it done retire, as so many are today, there may be nobody to take the readings.

The rest of the story is even worse. Once PdM measurements are taken and the data created, the job is only about half way done. The new data must still be organized into useful information, reviewed to identify any new equipment issues, and then stored where it can be easily retrieved along with earlier readings and baseline data on each piece of equipment. And once all that is accomplished, management must be trained and committed to read the information developed by the PdM system and make informed asset health care decisions based upon it.

Let’s see – train maintenance; initiate the monitoring program; identify key assets; plan the readings; make the schedule; take all the readings; record the data; analyze the data; package the results for management; read the data and make the right decisions; make a plan to execute corrections; perform needed maintenance. We have something like a 12 step process to go from PdM initiation to delivery of condition-based maintenance. Those of us who have beaten our heads against the reliability wall for a few decades will tell you that even 95% delivery on each of these elements is unlikely. If we do achieve it, the likelihood of successful completion for a machine is 95% to the 12th power, or 54%.

This isn’t an airtight quantitative analysis, but hopefully it demonstrates an important point. Without some kind of help, an all-out PdM program will probably deliver about half the potential it was designed to fulfill. Fortunately there is help available, but it is seldom gathered and used to good effect through all the stages of the PdM program.

For the first stage of the effort, capturing the data, a mix of time-tested tools is already at hand. A CMMS can store equipment criticality and monitoring plans. With proper staffing plans, training can be delivered to crews in time to take the needed measurements. Bar codes or radio frequency identification (RFID) tools can help ensure that all the right readings are taken on schedule. Even though actual use is pretty spotty, these tools are universally available and widely understood.

What will be special about 2014 will be some new capabilities enabling data to be transferred straight from condition monitoring equipment to cloud-based data reservoirs. If coupled to the barcode-controlled maintenance schedule, these data could be readily grouped around asset numbers and indexed in the CMMS as series of readings for analysis of equipment life cycle status.

Planners will still have to plan, and technicians will still need to visit the equipment and take readings, but scheduling the work and marshalling the resulting information can be simplified and, in some cases, even automated. Analysts will still need to read the data, and managers will still need to digest the analysis and make sound decisions, but they won’t be fighting endless file drawers of paper reports from the shop floor. Analysts and managers will be able to work as decision makers instead of librarians. In cases where outside skill is required for analysis, the data will be accessible, organized around equipment numbers and condition monitoring events.

This is a long march from where most companies are today, but the good news is that it can be done with minor financial investment. Bar codes tags and hand-held monitoring equipment are now available at reasonable cost. Soon they will be feeding results directly to the cloud. In 2014, monitoring equipment will also begin to unravel some of the complexity of taking and understanding readings, making maintenance staff more effective.

To make a long story short, 2014 could be the year that hand-held equipment developments, combined with improved use of data management tools we already have, could bring a new level of condition monitoring to medium-sized plants and their managers. This quality of coverage has been available for years on some megabuck equipment with dedicated instrumentation feeding data to outside analysts, so we know it works. In 2014, we will see the advent of tools that will make high quality coverage attainable for key equipment in most manufacturing facilities.

Watch for the product bulletins and have a Happy New Year!