Predictive Maintenance Technologies - Effective Predictive Maintenance

Elements of an Effective Predictive Maintenance Program

An effective Predictive Maintenance Program (PdM) is a process of five continually repeated steps:

1. Data Collection - Vibration data is collected from machines with

portable data collectors, online acquisitions or a combination of

both.

2. Data Processing - The data is loaded into a computer and

analyzed with the aid of Predictive Maintenance Software. The

software provides access to the vibration data and data history.

Alarm functions are provided that check the data and report any

readings that are over predetermined alarm levels. Generally

the PDM software turns the raw data into information about the

vibration characteristics of the machine.

3. Data Analysis - At this stage, the information must be interpreted.

The interpretation requires knowledge about vibration characteristics

and is preformed by trained specialists who are able to look at the

vibration data and understand the overall readings and spectra

graphs.

4. Repair Recommendation - Once the vibration patterns are

understood, knowledge of the affected equipment must be

considered. For instance, can an individual bearing be replaced,

or will the entire piece of equipment have to be removed for repair.

This step also compares the cost of a minor rebuild to that of a major

rebuild.

5. Implementation - The final task in the process is to consider the

repair advice and other relevant factors such as cost of repairs,

production requirements, budgets, ect… and make a decision. It

should be implicitly understood that when a decision is made it be

implemented. The appropriate decision may be inaction rather than

action. The "do nothing" alternative is often appropriate. In such

cases the work leading up the this decision has not been wasted, at

least you know why you are doing nothing and are aware of the key

factors that may change this strategy.