Condition based maintenance provides many well established benefits, however, in order to perform condition based maintenance, the condition of the machine must be established. Traditional approaches of trending and comparison with baseline conditions help detect that something has changed. However, a vital step is to determine what is wrong or changing with the machine so that appropriate actions can be taken.Traditionally this assessment of the problem or diagnosis, is performed manually by a machinery expert. He uses his knowledge of the machine and its behaviour to interpret the measurements available from the condition monitoring system. Knowledge based systems are a way of capturing that expertise and making it available in a computer program. Such a program provides a way to gain an expert evaluation of the data, even when the expert is not available. This could be used for diagnostics 24 hours a day, providing not only continuous monitoring of the plant, but also continuous diagnosis. In practice, condition monitoring means checking and evaluating the condition of the system on a regular basis. The primary goal is to determine when the machine is working properly and no action needs to be taken, and more importantly, detecting as early as possible the first indications of a developing problem.