Lift PdM data above the information din
In brief:
- Valuable predictive maintenance or condition-based maintenance data can sometimes add to the noise surrounding managers today.
- When dealing with PdM or CBM output, the decisions aren’t usually very complex, but the data can be baffling.
- Maintenance data needs to be distilled into actionable information and presented in an understandable and convincing manner to the proper organizations.
We’re all living in the Data Deluge of 2012. Broadcasting more raw data, even valuable predictive maintenance (PdM) or condition-based maintenance (CBM) data, just adds to the noise surrounding managers today. Most of them are trying to do their jobs while reading hundreds of emails per day and sorting through input from all the employees and contractors around them. Maintenance data needs to be distilled into actionable information and presented in an understandable and convincing manner to the proper organizations. This kind of information will help them make the right decisions in an efficient timeframe. In other words, it will make maintenance and reliability effective.
Data, data, everywhere
One organization that needs timely, well-organized information to maintain the asset health of its equipment is the Des Moines Metropolitan Wastewater Reclamation Authority (DMMWRA, www.dmmwra.org) in Iowa, which enlisted the help of a full-line PdM service provider for Web-based data and information support for maintenance decision making.
“After performing our monthly vibration, oil analysis, and annual air leak detection and infrared thermography inspections, our service providers post the results on the website within 24 hours,” says Bill Miller, CMM, MMC, MRO, ASE-EAM, ops support maintenance administrator, facilities management, City of Des Moines WRF. “They also post any results that need to be addressed within a specific time constraint on our alert board. We have also implemented the Infor10 EAM Asset Sustainability Edition (ASE) condition-based monitoring program, which is monitoring the energy consumption of our critical assets around the clock. Both of these programs are integrated into the ASE email alert system and EAM auto work-request (WR) generation function. Email alerts are sent out to schedulers, planners, and management, while WR are directed to the maintenance technicians. Through CBM we run a tight ship, should critical asset trending run outside of plus-or-minus two standard deviations from a normal curve, email alerts and WR are automatically generated and sent out to respective personnel. These two programs are set up and configured around a maintenance best practice and MRO storeroom environment.”
[pullquote]DMMWRA’s scale of operations and annual budget-driven operating approach dictate a different decision-making approach, as well. “Best-in-class advanced analytic reports we have developed pull accumulative data and upload it into our newly developed "quad" advanced analytic reports — the only one of its kind,” reports Miller. “The quad reports pull data from four major inspection and trending programs and incorporate this information into the quad report. The four areas from which data are pulled are facility conditions (FCAs), PdM inspections, EAM work-order historical records, and ASE energy consumption trending reports. Then the pertinent fields from each quadrant are uploaded into an advanced real-time physical condition lifecycle report used by engineering and finance for budgetary and CIP replacement scheduling purposes.”
The use of real-time data gathering and information generation have paid some exciting dividends for DMMWRA. “The biggest improvements have come through these key areas,” says Miller. “Through continuous energy consumption monitoring with the Infor10 EAM-ASE program, we have experienced annual energy savings in the area of $200,000. With the implementation of our advanced PdM and CBM programs, we have experienced extended equipment lifecycles ranging between 10 and 20%. The development of our advanced analytic reports has helped eliminate the run-to-failure maintenance approach, thereby taking the guesswork out of equipment replacement for engineering and finance. And all the information listed here assisted the Des Moines Metropolitan Wastewater Reclamation Authority in becoming one of the first industries in the nation to achieve PAS 55 certification.”
A good deal of training went along with the implementation of DMMWRA’s system. Even though most information gathering and processing is performed by contractors, several groups received training in the system. These groups included reliability and sustainability engineers, managers, supervisors, and maintenance technicians.
Train the users
When dealing with PdM or CBM output, the decisions aren’t usually very complex, but the data can be baffling. Usually the decisions take the form of when to issue maintenance work orders for repairs or periodic maintenance. The data, on the other hand, take the form of strangely named parameters and trends with mystically derived limits. For these to be useful, the data users must be trained. Staff turnover also requires that the training be renewed frequently in most companies.
In a plant with a sizable technical staff and a maintenance organization that can handle work orders reliably and in priority order, it may be adequate to train reliability and supervisory maintenance people and to publish data directly to them. This is the most economical approach to data use; however, training must be constant and up to date. Also, the maintenance work order system must be working reliably, or predictive work orders will constantly be displaced by breakdown work. Of course, the predictive orders eventually become breakdown work if they are ignored, but this will erase the safety, efficiency, and profitability improvements that the PdM system is designed to deliver.
One organization that has the necessary technical and maintenance staff to make use of data directly is Motorola. It also has the advantage of being a producer of information management hardware that simplifies the real-time distribution of data to factory staff.
“Our manufacturing customers are very concerned with enabling the most timely, accurate data collection possible,” says June Ruby, manufacturing principal at Motorola Systems (www.motorola.com). “That way, data can be analyzed and presented for decision support where and when it is needed. Mobile devices that include data capture technologies such as barcode readers or RFID interrogators capture the asset tag, date/time, and operator information for each station on the operator’s rounds without requiring data entry. The application tests the data entered by the operator for reasonableness. With the real-time connectivity of an industrial wireless LAN, the data is transformed into the decision support information promptly — for example, trends of the data with appropriate control limits can be displayed and a work order can immediately be scheduled to prevent a dangerous situation.”
In this situation, data are converted to information by the addition of specification data and graphical templates, and the new information is broadcast to trained personnel who will use it in day-to-day decision making. Even Motorola doesn’t simply pass data along, but converts it to information first.
“The needs to reduce unscheduled downtime and avoid more extensive repairs are the most frequently cited goals that drive investment in technology to collect accurate and timely condition-based data and manage the technician work order process,” says Ruby. “Providing the operators with a mobile view of the production system HMI allows them to collect the data and perform inspections while maintaining visibility of their process parameters. This provides context to the situational data and allows the operators to make informed decisions.”
Motorola is a company with a unique cost advantage in personal data management equipment. This kind of advantage can provide instant data access, and, when coupled with appropriate software, it can avoid some of the cost of distributing costly enterprise software to every employee on the floor. Instead, a great deal of data management can occur on versatile personal equipment with its own software. Maintenance work order generation can be performed at a more limited number of locations after the work requests have been generated on the distributed equipment.
“Mobile devices are becoming necessary tools to increase the efficiency and productivity of the workforce,” says Ruby. “Manufacturers are carefully selecting these devices to perform in their environment in order to reduce the effects of device failure and breakage. The younger, less experienced workforce can benefit from standardized workflows built by leveraging the knowledge of the more experienced workers. These workflows alert the operators or technicians when their task is next in the critical path of a line or unit shutdown or startup, for example. Business rules may be included that require supervisor or safety inspector signature capture on the mobile device before the task is considered complete.”
Think of the savings in time and money that could become available if this approach were applied to tasks like permitting and completed work approval in a large plant setting. In considering the fit of this kind of automation to the PdM system in a plant, it is important to ask a set of questions that match practitioners and decision makers to the use of data and information. It is essential for the organization to distribute decision-making authority to the places where information is available if the organization is to use real-time data at the sites where it is generated.
It usually makes sense to identify which decision requires support from the data. Usually the answer is straightforward. Common answers are things like whether to issue a maintenance work order or whether to continue running the equipment. Answers to these questions can usually be built into the performance limits that are part of the information presentation that is driven by data capture.
“Who will make the decision?” is a common next question. If the organization is willing to distribute decision making authority, as well as the necessary training, then it is positioned to capture the advantages of real-time, on-site decision support.
Part of the real-time, decision-support discussion should also usually address how fast an answer is needed. If time is not an issue, then the perceived safety of decision making further up in the organization may be an attractive alternative. But when is speed not important when the issue is including condition-monitoring data in operating decisions?
If the objective is to build a system to bring questions and answers together for the right person in the right time frame, speed is typically a prime consideration. So is cost. For some companies, this may be a more complex discussion.