The Upper Occoquan Service Authority is a modern water reclamation facility southwest of Washington D.C., which receives waste flows from five jurisdictions. The plant was constructed in the early 1970s. It is the result of the Occoquan Policy and started with a capacity of 15 million gallons per day. Subsequent construction projects brought the plant to its current capacity of 54 million gallons per day. Renewal and improvement of the plant has been an ongoing process. Improvement of operations and preventive maintenance tasks has likewise proceeded on the path of continuous improvement.
This truly is a coming of age story in the world of modern technology and is filled with missteps, epiphanies and lessons learned through constantly seeking better ways to perform maintenance.
Back in the 1970s when the facility was built, there were no personal computers on everyone’s desks. Only the most ardent practitioners of maintenance had any idea who W. Edward Deming was or how he revolutionized the world of maintenance. Taking care of equipment at that time was largely a matter of changing the oil and fixing the machines when they broke down. In essence, it was reactive maintenance.
In 1983, a computerized maintenance management system (CMMS) was introduced at the facility. This system was called the total utility maintenance management system (TUMMS). The people who launched the endeavor did their best to fit existing maintenance practices into an off-the-shelf computer program. The whole idea of using computers to manage maintenance was new and challenging. Now we can’t imagine how we’d manage without such tools.
Two experienced maintenance people were designated as maintenance planners to manage the work order queue and set up a preventive maintenance schedule. These planner positions were to evolve over the next years, as did the role of their maintenance planning office. These planners dutifully entered thousands of tasks into the CMMS based on their experiences and the recommendations of the operation and maintenance manuals.
In 2002, a new enterprise asset management (EAM) system was selected and implemented at the facility. It had a sophisticated CMMS bundled with it. The information technology department transferred all the TUMMS legacy maintenance stuff to the EAM system. As can be imagined, the transition was not perfect, but the results were workable. Everyone adapted and began to discover the new functionality of an EAM coupled with a CMMS. We could now generate reports more effectively and reveal the key performance indicators.
Bringing over the TUMMS legacy format had been advantageous in that it preserved a system that was already working. However, living with the vestiges of the original systems would soon reveal the need and the potential for improvement.
Around 2005, the original maintenance planners retired and two new maintenance planners were hired from outside of the organization. The new planners began to rethink what had been done and why it had been done that way. The new planners had to learn the plant as beginners. These guys brought with them new ideas on maintenance as well as notions of how to quantify and verify improvements.
The TUMMS had been set up originally with thousands of individual preventive maintenance tasks that defined individual steps because there was no means to associate detailed procedures. These tasks were manually checked off every two weeks in three-ring binders. The maintenance planners updated these pages. New pages were printed for the PM books and issued to the trade techs. Several trades had multiple PM books to cover designated areas.
Furthermore, these preventive maintenance tasks had trade-specific codes using three-number prefixes to denote the specific trade. For example, all electrical systems tasks began with a prefix of 300. Mechanical systems began with a prefix of 100. These prefixes had facilitated the necessary identification and sorting in TUMMS. The new EAM system had a simple way to parse out the trades which made the three-number prefixes unnecessary.
Generic Preventive Maintenance Tasks Created
In 2006, the maintenance planners agreed to create generic preventive maintenance tasks that could be used by any trade. These tasks would carry a prefix of 000 to distinguish them from legacy tasks. Generic PM tasks were developed to replace the numerous duplicate tasks and address core PM needs. The process of weeding out the old tasks began.
Prior to starting the process of PM optimization, it was not unusual for a machine to have an individual task to check the oil level, a task to check the belt tension and a general inspection task all due on the same date.
This may not sound like such a bad plan until the number of tasks is multiplied by hundreds of machines. Some machines had more than a dozen tasks due on the same date. This made for thick, cumbersome PM books which were difficult to get through in a two-week period. The size alone was daunting. The technicians were not amused.
Something had to be done to improve the PM assignments and to get more buy-in from the techs. Paring down the sheer number of preventive maintenance tasks became a priority.
Creating generic PM tasks had several purposes. First, the use of generic tasks aligns with best practices in industry. In practice, there are core tasks. For example, inspect, check oil level or verify the correct operation are some of the core tasks trade techs perform.
Generic preventive maintenance tasks would be used going forward. The CMMS also has the means of associating detailed procedures with a task when required. Associating procedures would reduce the need for hundreds of individual preventive maintenance tasks created largely to spell out procedural steps.
Second, creating generic preventive maintenance tasks initiated a review of all PM tasks, which revealed that some trades had preventive maintenance tasks with duplicate descriptions but different task codes. One trade had three different codes with the same description – exercise. Apparently, maintenance planners missed these duplications during earlier reviews.
The TUMMS had required each trade craft to have its own set of preventive maintenance task codes. As a result, each trade craft had a full set of basic tasks bearing duplicate descriptions, the only differences being the prefix. The list of PM tasks had bloomed into something large and unwieldy. That preventive maintenance task list was soon cut down to a manageable size.
The resulting review launched discussions on how PM tasks could be rewritten with reliability foremost in mind. Several approaches were considered, ranging from completely replacing the current PM assignments to changing the PM one task at a time.
A hybrid plan evolved to modify large chunks of preventive maintenance tasks and individual tasks alike. A plan emerged and was agreed upon for “scrubbing” the PM tasks. It was a bit like rebuilding a brick wall by replacing selected tiers.
The agreed plan was recorded and stored on the server for reference and then set to work executing these changes.
Getting Better and Better
As part of the process, the planners sought more training and were encouraged to attend maintenance conferences and network with peers. They began to meet more frequently with trade managers, leads and techs to get input on ways to improve the PM assignments.
The process of addressing the identified problems began with a careful review of all existing PM tasks. The goal was to reveal duplication of preventive maintenance tasks between trades and to determine what tasks should be deleted.
A fair amount of overlap was discovered. For example, electrical and mechanical trades might both have a task to inspect the same equipment. In practice, an electrician can perform a visual inspection just as well as a mechanic and vice versa. One of these two tasks was deleted. A procedure was added when merited.
It was also discovered by researching PM backlog records that a sizable number of tasks had been created but never used or assigned. The never used tasks were deleted from the system.
One of the first steps the planners took was to create a PM change log. A log entry was made for each PM change including which planner made the change, the date and why. Keeping this log may seem like extra and unnecessary work, but it proved its worth on more than one occasion when historical reference was needed. There were numerous cases when trade personnel asked for a change one week and then asked to reverse the same change a few weeks later. This log also included codes to identify changes as deletion, additions or modification. These codes were used to characterize how the overall PM tasks list was changing in terms of reduction and improvement.
Another effort was to work with trade managers and lead techs to rebalance and better allocate the preventive maintenance task load among their staff. The planners spent many hours communicating and making changes to meet this purpose. This provided great relief to the techs, making preventive maintenance more agreeable and manageable. The buy-in from the techs increased.
Another impediment to rational preventive maintenance was that planners often created PM tasks that reflected as accurately as possible what the manufacturer put in their operation and maintenance manuals. Manufacturers often go a bit overboard and suggest a preventive maintenance plan that is unnecessary or at too frequent an interval. It would be hard to fault these planners, as they were simply doing what seemed right. By careful review of maintenance histories, the new planners were able to right size PM assignments to meet operational circumstances. The results have been enhanced equipment reliability and savings in maintenance costs. This has been revealed through analysis of corrective work orders.
An existing oil analysis program was also improved and expanded, resulting in a reduction of the number of oil changes. This saved both labor and material costs. Additional machine health monitoring technologies are being deployed to further increase equipment reliability.
The maintenance planners continue to meet regularly with section leaders to discuss the PM process and maintenance problems in general. The initial efforts for PM optimization would include harvesting the low-hanging fruit. This involved deleting preventive maintenance tasks with no value and tasks that were redundant between trades. Due to the sheer number of total PM tasks (more than 32,000) entered as legacy PMs, this required months to accomplish.
By definition, PM tasks should be short activities consuming a few minutes each. It was realized that better instruction was needed on how to perform a task, such as greasing, inspection and operational tests. A major advancement was associating detailed procedures with preventive maintenance tasks. Procedures can call out the amount of grease, calibration specifications, torque values and other vital information. Procedures also function as a training tool for new hires because they capture expert knowledge and make it readily available. Implementing procedures fit nicely with the goal to improve PM precision.
For example, a motor may need a specified amount of grease, to be running during lubrication and to have a grease relief port opened. Procedures were created and associated with a PM task to meet this purpose for specific motors. These procedures are included with the PM books as an appendix for the techs to reference.
There was some pushback to incorporating procedures. Certain individuals viewed it as an unnecessary change. The trades began to appreciate these procedures when they discovered that they saved time and spared uncertainty. It also helped them plan their work. Getting the trade techs involved with writing the procedures also assured buy-in.
In the future, these well-vetted tasks and procedures will be used on handheld devices, with check off and update done by computer. Procedures are also being employed as part of operator care preventive maintenance tasks to make certain operators know what is expected of them.
The next steps in PM optimization will be parsing out the most important equipment to examine their failure modes. As predictive maintenance is implemented into the overall maintenance approach, understanding how machines fail becomes more intimate. When the root causes of failure are known, steps can be taken to mitigate recurrence. These are steps on the path to condition-based maintenance.
Another initiative that has been committed to is scoring criticality, probability of failure and current condition for these important assets. The calculated scores help to determine how best to allocate maintenance resources based on the consequence of failure.
PM is also being standardized for like assets where practicable, based on past successes. While new construction is still in the planning stages, the engineers are trying to build in ways to make the equipment easier to maintain. We’ve learned from experience that it is much easier to build in maintainability than to retro-fit later.
The CMMS includes a useful feature which allows planners to create standard work orders in template forms that can be generated based on time intervals or meters (hours, miles, etc.) input. This template form is referred to as “model work orders” and is used when maintenance evolutions require significant material cost, outside labor cost, produce inspection documents and other requirements. Being able to store all the pertinent information in one record helps manage maintenance activities and facilitates in analyzing improvement steps. We have steadily expanded our use of these model work order templates.
There are no simple answers to maintaining a large and complicated plant. We started our PM optimization process by cleaning up individual tasks to make them relevant. Along the way, our eyes opened to many additional steps for improvement. Our efforts led to changes in maintenance perceptions and behaviors. The maintenance planning personnel have worked to foster an environment of trust, customer service and improved reliability.
We now have fewer high priority work orders than in the past. Plus, we are maintaining additional equipment with no additional staff. Maintenance planning personnel have collaborated on reference documents that explain what has worked well with maintenance at our facility. These living documents provide a framework for future planners to follow. These documents explain how success was achieved and measured.
In looking to the future, anticipated changes in available resources make a compelling reason to discover, achieve and adopt best management practice. We work every day to get better at what we do because whatever got us to these levels of success will not be enough to keep us there. PM optimization is an ongoing process.
This article was previously published in the Reliable Plant 2017 Conference Proceedings.
By Saul Cizek, Upper Occoquan Service Authority