Subscribe to our

Pumps & Systems, May 2008

There are no easy solutions to the high costs of maintenance. A substantial amount of time and effort is required to select predictive methods with the most cost-effective means to evaluate the operating condition of critical plant systems; establish a program plan; create a viable database and establish a baseline value is substantial. The actual time and manpower required varies depending on plant size and the complexity of process systems. For a small company, the time required to develop a viable program is about 3 man months. For large, integrated process plants, this initial effort may be as much as 15 man years. Are the benefits worth this level of effort? In almost every instance, the answer is an absolute yes.

Here are ten steps that can help you implement a successful total plant predictive maintenance program:

1. Determine Existing Maintenance Costs

The most difficult step in the initial justification of a predictive maintenance program is the determination of actual maintenance costs. Most plants do not track all controllable costs that are directly driven by the maintenance operation. In most cases, the cost-accounting function limits cost tracking to actual labor and material used to maintain plant equipment. They do not include the impact of maintenance on availability, production capacity, operating costs, product quality and the myriad of other factors that limit plant effectiveness.

In addition to maintenance labor and material costs, an evaluation should include all maintenance-related costs associated with delays, reduced capacity operation, overtime premiums and product quality. Safety and environmental compliance should be included.

In some cases, the accounting department can help develop a close approximation of the true costs of maintenance. Explain the reason for the request and let accounting experts help quantify the historical plant costs. The cost history developed at this time is extremely important. Initially, it will be used to develop a cost-benefit analysis and justification for the predictive maintenance program. Later, this data set will become the baseline for quantifying the actual benefits derived from the program.

Plants should not shortcut this part of the program implementation. Accuracy and completeness of this data set is critical to the long-term success of the program. The majority of programs that failed in the first two years following implementation can be directly attributed to the lack of quantified results.

2. Select Predictive Systems and Vendors

Another major contributor to program mortality is the selection of either the wrong predictive technologies or a vendor who cannot provide long-term program support. Extreme care must be used during this selection process.

A total plant predictive maintenance program must use a combination of monitoring and diagnostic techniques to achieve maximum benefits. None of the individual technologies, such as thermal imaging and vibration, provide all capabilities required to evaluate critical plant processes and systems. What combination of technologies is best for your plant?

Unfortunately, there is no easy answer. The predictive requirements of each plant are different. As a minimum, the program should include (1) key operations processes analysis, (2) thermal imaging, (3) process parameters, (4) visual inspection and (5) vibration. Lubricating oil and wear particle analysis (tribology) should be used only where the added information derived will justify the costs.

Care should be exercised when selecting predictive systems and vendors. The following should be considered the minimum when selecting predictive maintenance systems:

a. Adequacy to specific needs-No predictive maintenance systems are perfect. Each has its unique strengths and weaknesses. For example, many of the vibration monitoring systems cannot handle machine speeds below 600-rpm or lack the ability to use a variety of transducers. Either or both of these limitations will reduce the benefits that can be derived from your program. Define the specific requirements for your systems and make sure that the selected systems will fulfill all requirements.

b. Stability of system and vendor-It is essential that the systems selected for the plant remain viable for an extended time period. Competition within the predictive maintenance arena is fierce and many early players have gone out of business, merged with other companies or changed system structure. All of these factors affect the long-term status of the program. The evaluation should include financial strength of the vendor; history of product development; technical support and existing client base.

3. Training Requirements and Support

Most predictive maintenance vendors offer some level of training. However, most of these training programs are directed toward the use of a specific system, i.e., software and instrumentation, rather than comprehensive use of the technology. As a reference, I have used all of the predictive maintenance technologies for more than 30 years and still learn something new every day. There are a number of vendors who offer technical training that can support the predictive maintenance program. However, carefully evaluate the merit of their courses before electing to use them as training support. In general, independent training companies, with no association with equipment manufacturers, can provide high quality training with an unbiased approach.