“All rotating or reciprocating machinery emits a unique pattern of vibration characteristics. The pattern of vibration, or vibration signature, Represent the current mechanical condition of the individual machine. As time passes, the equipment mechanical condition will change due to internal wearing, unbalance, misalignment, looseness, and related problems. These changes in machinery condition will also affect the machines vibration signature. The purpose of a vibration monitoring and analysis program is to detect changes in equipment signatures and use this information to pinpoint equipment degradation and thus schedule corrective maintenance or overhaul. The statistical methods outlined in this paper are effective tools for monitoring machinery condition, improving production processes and reducing unscheduled failures.

“Statistics:

Webster defines statistics as, “”the mathematics of the collection, organization, and interpretation of numerical data, esp. the analysis of population characteristics by inference from sampling. As Webster’s definition points out, data collection by sampling, interpretation of numerical data, and analysis of population characteristics are all tools used in a vibration program. The statistical tools outlined in this paper are more tools to help monitor machinery condition.

The following FACTORS are presented as an aid to help avoid problems in the statistical methods presented in this paper.

“Factor 1: Data Collection:

Regardless of type of data collection equipment used, CONSISTENCY is of the utmost importance in data collection. Consistency in the, data collection techniques, location of data collection points, equipment operating parameters and, in some cases, ambient conditions.

It is highly desirable that the same personnel remain in the program. It is especially important to have the same personnel collect data. Consistency in personnel helps to ensure accurate and consistent data collection and analyses.

“Factor 2: Documentation:

Once data is collected, various statistical methods may be used for analysis so that the data becomes a meaningful source of information. The data recording methods will vary greatly but the following points need to be considered no matter what method is used.

First, the origin of the data must be clearly recorded. Data whose origin is not clearly known becomes dead data. Quite often, little useful information is obtained despite the fact that months were spent collecting data, because the date it was collected or which machines the data represented was omitted.

Secondly, data should be recorded in such a way that it can be used easily. Since data is often used later to calculate the mean, standard deviation, and days to alarm etc, it is better to record the data in a manner which will facilitate these computations.”