Condition monitoring is defined as the collection and interpretation of relevant parameters that are characteristic for the equipment condition. These characteristics, such as vibration and temperature, usually remain stable as long as the equipment is healthy. However, an abnormality in these characteristics may indicate the occurrence of a functional failure. Information on an incipient equipment failure can be obtained from the monitoring of these parameters. When the monitored equipment parameters exceed the alarm level, alarms will be activated, indicating that a certain condition threshold has occurred. Then a diagnostic system will be activated to look for a similar situation and automatically pick up the maintenance remedy.
CBM assumes that indicative parameters can be detected and used to identify possible failure before it actually occurs. Prognostic parameters provide the indicators of potential problems and incipient faults. Trending deterioration can be identified through analysis of data. Maintenance actions are than executed based on measured abnormalities.
A decision support system could enhance decision making by providing: fault recognition, fault structure, statistical tools and knowledge base. Such a system should enable easier and faster generation of alternatives, and may increase the decision-making process. It also can help to make effective and efficient decisions in complex situations.
Parts of such a support system should be at least contain a knowledge base, Petri networks, neural networks, fuzzy logic, and Bayesian models and AI (Artificial Intelligence). The fundamental issue in building such a support system involves linking the knowledge of experts with analytical decision techniques.
When a support system is in use, a fault diagnosis is triggered by the detection of a condition that is recognized as a deviation from the expected level. Than an automated process is starting that detects abnormal problems and faults, recognizes and analyses the symptomatic information, identifies and locates the root causes of a failure, obtains the fault development trend, and predicts the remaining lifetime of the equipment. The intelligence used in this support system can be divided into three categories:
- Rule-based diagnostic systems. 2. Case-based diagnostic systems. 3. Model-based diagnostic systems.
Rule-based diagnostic systems detect and identify incipient faults in accordance with the pre-defined rules of each possible fault with the monitored condition. Case-based diagnostic systems use historical records of maintenance cases to provide an interpretation for the actual monitored conditions. The case library of contains records all previous similar incidents. If a previous fault occurs again, a best practice, remedy will be presented. A model-based diagnostic system uses logical methods to improve diagnostic reasoning. The monitored condition is compared with a model of the object in order to predict the fault behavior.