J.K. August is a plant engineer with CORE, Inc., 5915 Braun Way, Arvada, CO 80004, 303-425-7408. This is an excerpt from his October 2006 presentation at the annual SMRP conference in Birmingham, AL.
Pumps & Systems, April 2007
To quickly identify, select and perform effective PM tasks, plants must (1) identify high-risk equipment, (2) develop efficient, effective tasks to manage the most likely failures and package those tasks into work orders, and (3) implement the orders into work control systems.
While simple in principle, for any plant with multiple regulatory, production and cost obligations this is a tedious chore that must satisfy plant license technical specifications, regulatory codes and other safety requirements.
Options to develop PM plans vary from very simple one-time development to living programs. Picking the choice that best meets your needs comes down to your business case.
One way to quickly build, launch and maintain a scheduled maintenance program is to develop component-based templates. These templates provide standard work plans at a low cost by reusing developed work on identical or nearly identical components, thereby reducing the source development research time involved in vendor O&M manuals and other references.
For example, similar components or major equipment types can be grouped for common modeling, such as "Vertical Pumps." Grouping must be done carefully, however, since two similar components - check valves, for example - may not be truly identical. Lift and swing check valve construction is slightly different. They have slightly different failure modes.
Small component design variations can make a big difference in failure performance (lifetime, failure modes, etc.), which, in turn, makes all the difference in work plan tasks. Customized work plans will address these differences. Historically, maintenance work plans were customized at point of use, but in many industries that is no longer desirable or even acceptable.
When maintenance personnel are experienced and highly knowledgeable, component-based PM optimization (or PMO) allows them to reasonably customize the performance details. However, that's also a principle weakness, because this requires a skilled workforce that can perform under flexible work rules. In highly-regulated nuclear, food and drug, or aerospace industries, that's simply not feasible.
PMO has speed advantages and is easy for non-technical workers to understand. It can be implemented with traditional documents and spreadsheets. But its disadvantages are its inexact scope of work: it doesn't work well where the workforce is unskilled, lacks performance flexibility, or the performance risk is high.
PMO is fast, but inexact for custom equipment installations.
Another way is systems-based PM development, a risk-based approach that reflects the work in its most current maintenance form. This approach requires more cross-organizational support and sophisticated software tools to keep it straight.
Developing standard equipment templates speeds up the systems analysis by classifying and applying risk that, in turn, customizes the work tasks, intervals and content exactly. This systems-based approach provides the best work basis, since ultimately not all equipment is created equally. Focusing by risk maximizes the maintenance expense return.
Developing many system components concurrently accelerates the creation of systems-based PM. Using templates and symmetry maps will speed up the process by identifying all of the common system component tags. Combined with application software and knowledgeable staff reviews, this provides risk-based work plans that are the most cost-effective over plant operating lifetimes.
Another advantage of the systems approach is that risk-based understanding is better suited for instrumentation and controls. Operations and engineering personnel already assess failing equipment daily by using the systems approach, so risk-based formalization helps them.
Greater effort in developing the component plans is the primary disadvantage. Without database software, the component plans are very hard to use. Analysis is repetitive (and boring) for mechanics and electricians as task users, but those who maintain the tasks find it useful.
Which is Right?
Statistically, certain equipment types will lead to different plan strategies. Equipment wear-out supports time-based PM, though condition assessment helps determine when to perform PM effectively. How can you tell you have the right approach?
Developing work processes will identify performance indicators (PIs) that are important to your organization. Software applications also have predefined indicators, depending on their sophistication. All of these will open discussions that are useful in identifying key performance indicators (KPI). A substantial body of information already exists that can suggest what indicators can be useful.
Be aware that information collection costs can make certain KPIs impractical. Many plants operate several different reporting requirements, such as NERC, FERC, 10CFR50.65, the Maintenance Rule in custom software. However, operating many different reporting systems independently like this has huge embedded costs that aren't always apparent.