After reading the Plant Services April cover story on ultrasound and vibration monitoring, I recalled an interesting experience from a few years ago in Australia. A client there showed me a bank of file cabinets in a storage room behind his reliability department offices. “Here it is,” he said, “our predictive maintenance (PdM) and root cause failure analysis (RCA) data library. We call it the ‘holistic asset healthcare (HAH) system.’”
OK, they didn’t really call it, “HAH,” but the rest is true. The data in the cabinets were filed by asset number, and there was a parallel electronic version on the intranet that held monthly readings from a huge number of condition-monitoring samples, taken using several technologies. The two libraries contained most of the information needed to predict upcoming failures of most of the plant’s key equipment. Some of it had been read and analyzed and had maintenance orders written on it. But the files did not say which information had been summarized and transmitted, either as reports or as work orders. Most of it had been simply recorded and neatly filed. There was little other communication with the departments that owned the equipment and no direct connection to the computer maintenance management system (CMMS).
Implicit in the design of this “system” was that the departments owning the equipment would take the initiative to discover the files and make weekly or monthly pilgrimages to the file locations to read raw data and RCA results about their equipment. Of course the traffic couldn’t have been less if they’d wrapped the whole thing in barbed wire. I think there were a few visits by engineers who were preparing projects or shutdowns in the production areas, but that was about all the use the costly data saw.
[pullquote]The HAH data system was an extreme example, but it serves to demonstrate an industry-wide trend. Data gathering and analysis technology have outstripped the training offered by most employers. As a result, condition-monitoring data are not receiving the attention required for users to reap the business benefits of all the information they are paying for.
In other words, a technician in the company often understands PdM equipment and has the data to predict equipment failures. Unfortunately, he or she often doesn’t really know how to read the data and create useful information for management. In cases where good, useful information is being created and transmitted, management often doesn’t know how to react to the readings provided by the technicians. If there is no reaction to the data, failures occur as predicted, forcing reactive maintenance and creating other forms of chaos.
When this happens, the company is already spending 90% of the time and cost required to develop the information that is needed to improve maintenance and reliability. It is only necessary to introduce the people with the information to the people who need it and add enough training so that the techs understand their data and management understand their techs. This is the kind of situation that consultants live for. The money is spent, but the benefits have not been reaped.
In 2013, almost any company has some measure of this problem/opportunity. A consultant, or anyone else who wants to help and look smart in the process, can follow three steps to identify this situation and capitalize on it.
1. Find the data: See who is responsible for reliability and condition monitoring and find out what data they are generating. They’ll be somewhere in the reliability, maintenance, and production organizations, and they’ll be all too happy to let you know what is going on. After all, they’re probably spending their time developing and storing data that has great potential and that nobody is using. Develop a record of what data they have and what they think the potential for it is. Hopefully there will be multiple sources. Find them all and add to your power and information base.
{pb}2. Identify opportunities: Work with the techs you have just met and your production or production control people to identify production losses for the past year that were caused by equipment failures. See whether the failures were predictable from PdM data and what had been done to alert the production or maintenance organizations to the problems prior to the failures. Determine whether the right equipment was being monitored. If there is someone in the organization calculating and publishing overall equipment effectiveness (OEE), enlist that person immediately. OEE uses much of the data you need. It also includes capacity losses due to equipment condition. Join forces with the OEE team and use their information to supplement the failure data you have unearthed.
Chart the data flow that was supposed to have occurred when impending failures were identified. Determine what portion of the data flow failed to create corrective action. Estimate, if possible, the value to the company that having the right information in the right place would have delivered. This doesn’t have to be to three decimal places and company-wide. You just need to show that data gathering is taking place or could take place if people understood the opportunity and that the data could be used to improve reliability of production equipment. Usually that is not very hard.
3. Plan and execute the fix: Typically PdM information is gathered and analyzed by maintenance or reliability. It is then packaged by reliability and passed to maintenance and production management for action. A small group from these functions should, in one or two meetings, be able to agree on the data flow that will be required in their company. Once the players have been named and given the responsibility, technicians and, perhaps, outside vendors should be able to identify the training required to equip techs and managers to make the decisions that the data flow requires of them. If financial, engineering, or other functions should be represented, bring them into the discussion.
The resulting plan should be a relatively inexpensive shift of responsibility with enough training to equip everyone to meet new responsibilities. Add a plan for year-over-year comparison of production losses, and the process is ready to begin.
If additional condition monitoring is needed, there may be some instrumentation and training for the technical people who do the readings. This cost should be easily offset by the production losses that will be prevented.
| For additional information on PdM data management, read Lift PdM data above the information din. |
Also on the subject of cost, the decision must be made as to what mix of in-house and outside project resources to use. Here is what Jason Tranter, managing director and founder of Mobius Institute (www.mobiusinstitute.com) has to say on the subject: “I personally believe that change is best driven from within, rather than via consultants with their trademarked techniques. Too many companies have gone through multiple attempts at reliability improvement, whether it is lean, RCFA, Six Sigma, TPM, or some variant of RCM. There may therefore be a great deal of cynicism when another trademarked technique is introduced. A grassroots, teamwork-led, commonsense-driven approach to reliability improvement with realistic goals and buy-in from top to bottom cannot fail.”
Of course, none of this will happen by itself, so deputize your favorite project manager, probably yourself, to make it happen. Then add just one more task — find a better acronym than “HAH.”
Read Stanton McGroarty's monthly column, Strategic Maintenance.