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:

  1. 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.
  2. 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.

4. Get Management Support

Lack of a total commitment from plant or corporate management to provide the resources required to implement and maintain a program is the single largest reason for failure of predictive maintenance programs. There are a number of reasons for lack of long-term commitment. However, in most cases, it stems from the lack of planning and justification in the pre-program effort. Management must know the true cost and potential benefits of the program before it begins. After implementation, they must be continually informed of the progress and actual program benefits. Therefore, a viable means of quantifying the actual results of the program must be developed and the ongoing status of the program communicated to all key management staff.

Management support should include implementation of a formal maintenance planning function, a viable information management program and craftsman skill training to gain maximum benefits from predictive maintenance. The predictive program provides the trigger for maintenance activities, but without proper planning and repair skills, full benefits cannot be obtained. The information management program has two functions: (1) maintain equipment histories and (2) track program benefits.

5. Develop a Program Plan

A definite program plan that includes all activities required by a total plant predictive maintenance program must be developed before implementing the program. The program plan should include:

  • Specific scope of the program
  • Goals and objectives

Methods that will be used to implement, maintain and evaluate the program. The plan should also include specific return-on-investment (ROI) milestones that can be used to measure the success of the program.

6. Dedicated Personnel

A key part of a successful program is a full-time, dedicated staff. The program cannot be implemented or maintained with part-time personnel. Regardless of the predictive maintenance techniques used for the program, regular, periodic monitoring of critical plant parameters is an absolute necessity. Most programs implemented with part-time staff have failed because activities required to maintain the program have been delayed or ignored because of other pressing demands on staff time.

7. Establish Accountability

The predictive maintenance team must understand the reason for implementing the program and be accountable for its success or failure. Staff commitment is an absolute requirement for a successful program. Without this total commitment, the program will probably fail. Division or area managers must also accept responsibility for program success. In most plants, these managers control the resources, both financial and personnel, within their departments. Without their full support and commitment to the program, little can be accomplished.

8. Develop a Viable Database

The actual benefits derived from a program will depend on the accuracy and completeness of the database developed. All predictive maintenance technologies depend on a clear, detailed definition of the critical equipment included in the program.

Database development requires a tremendous effort in both manpower and time. A typical microprocessor-based predictive maintenance program may require as much as 10 man years to develop in a large, integrated process plant. Even small plants must invest an average of 1 to 3 man years in this startup effort. However, the time is well spent. The initial investment will greatly reduce the manpower and time required to maintain your program and will greatly improve the benefits.

Many program failures result from shortcutting the database development step. In part, this is driven by the absence of accurate machine data and by the restrictions of many predictive maintenance systems. To achieve maximum benefits from your program, invest the time and manpower required to establish a complete database.

9. Maintain the Program

Do not quit after the implementation phase is complete. Many programs fail because the plant staff did not follow through after the development stage. Follow the program plan. Meet each of the schedules and milestones developed in the program plan. Constantly evaluate the program's progress and correct any errors or problems that may exist. A successful predictive maintenance program must be dynamic. Follow through.

10. Communicate

Communication is necessary for long-term success. All successful programs have a well-defined communications plan that includes transmittal of corrective actions identified by the program, feedback from manufacturing and a regular program status report circulated throughout the plant and corporate management team.

Program justification is a never-ending process. Management and other plant team members must be continually informed of the program's status and benefits. Failure to communicate will severely reduce the potential for success.

The Payoff

Although the effort required to implement and maintain a total plant predictive maintenance program is great, so are the benefits. Properly implemented and maintained, predictive maintenance, as part of a total plant performance management program, can reduce the negative impact of maintenance on availability, product quality and operating profit.

Predictive maintenance can transform the maintenance operation from an expensive support function to a full member of the profit generating team in your plant. Do not expect an easy quick fix. Like all things of value, a certain amount of effort is required to gain positive results. Follow these steps and establish a total plant predictive maintenance program that will provide maximum benefits.

Pumps & Systems, May 2008