In my experience, there are three types of maintenance philosophy regarding vibration programs:
- No philosophy at all
- Utmost attention with zeal
- Attention to particularly problematic units
The first group simply does not know any better. Such plants may have pumps repaired every 12 to 18 months, and consider that normal. Some of their pumps may have been pulled to alignment by comealongs that wreck couplings, make leaking seals and failed bearings. I have seen vertical pumps sitting on soleplates with no anchor bolts, and no holes drilled in the soleplate for the anchor bolts to match the support head bolts (see Figure 1).
Figure 1. This pump needs help, and so does the plant.
Explaining the benefits of vibration analysis to a maintenance manager at such an organization is usually futile. These organizations do not ask how a vibration analysis program can stretch a MTBR from 18 months to 18 years, but which repair shop can repair the failed pumps less expensively every 18 months. The only way to make a difference with such an organization is to convince management to at least review the cumulative costs of the repair work orders and compare the cost of the proposed implementation, with clear correlation to the payback time.
The second group is the enthusiasts. They typically have high energy, high speed rotating equipment, and the remnants of the previously strong reliability department/group. The plant's Vibration Specialist Bob is a few years from retirement and has an ample database of field success stories-which is ready to be buried, along with Bob's computer, after his retirement. The management accounting system continues to decimate such groups in response to cost cutting initiatives, regardless of the needs of the machinery and vanishing expertise.
Pleasant exceptions exist. Vibration analysis is thriving at many companies, with a full spectral analysis of critical or problematic units supplemented with periodic overall vibrations (rms) routes to monitor and remove problems.
The third group, similar to the second, does understand (or has heard of) the benefits and savings of vibration analysis, but, due to limited or reduced budgets, can only implement it in rare cases, when a particular piece of equipment begins to fail too frequently and touch-and-feel troubleshooting no longer identifies the root cause.
All three types are united by the ever-increasing costs of operation, and a need to justify any investment, in any technology, with a clear and simple, easy-to-understand program. Unfortunately, the technology experts, including vibration performers, cannot often condense their impressive knowledge of Fast Fourier Transform functions and phase angles into a clear and understandable message that relates the cost of the program to the real savings it offers to the plant.
Implementing Vibration Analysis
When a plant considers implementing vibration analysis, the questions that arise are:
- Why do it?
- Which equipment to monitor?
- How often to monitor?
- What are the limits criteria?
- What to do when limits are exceeded?
The answer to "Why" is to save money. "Which equipment" should initially be the most critical 5 percent of the pump population. This is simpler to implement and easier to measure the results, which builds confidence in the overall program. "How often" is once per quarter, or, if budgets are truly skimpy, twice a year. The limits: set the value of 0.30 in/sec (vibration velocity rms, overall) as a warning limit, and over 0.30 in/sec as an alarm. Do not initially consider the specific pump type's height to the measurement position (for vertical turbine pumps, for example); keep it as simple as possible. When the limits are exceeded-act. This may mean a more detailed spectral analysis, involving operations, engineering and maintenance groups, and ultimately deciding on a specific corrective action.
The most important things to watch are trends, as they are often even more important then the vibration level per se. Consider the two examples in Figures 2 and 3.
Figure 2. Vibration trend steadily increases with time
Figure 3. Vibration changes dramatically, from low to highly unacceptable, from one visit to another
In the first example, vibration steadily increases and crosses the warning and (eventually) alarm values. Continuation of the trend will likely cause catastrophic failure of the unit, and corrective action is critically important. In the second example, however, vibration fluctuates from one visit to another, ranging from a low acceptable value to exceeding the alarm value. In this case, different products are pumped by the pump, and critical frequencies are excited by insufficient damping and stiffness characteristics of a certain product on the clearances. Identification of this issue clearly could not be made in only one visit.
Make sure to involve maintenance and operations in the solution. Taking ownership in the project is often the key to its success.
In the next issue, we will discuss the reasons for the fluctuating trends for the pump in Figure 3. If you think you know already, drop me a line. If you can guess correctly, you win a free ticket for the next Pump School!
Pumps and Systems, June 2009
