The reliable operation of the electrical-network plays an important role in the present and the asset management assures the principle of it. That means the frequently condition measurements, repairments and replacements of the equipments. These should be scheduled effectively so the quality of the service does not decay. To do so there are decision support expert systems which make the registering and processing of the information more faster and secure. The recognition of the possible potential contingency is the most important during the operation so the prognosis of the failure of the equipments is in the focus.
In my master of science thesis, I have developed a statistical model for an existing expert system based on the calculation of health index, which model will take part in the paramater tuning of the health index based system. First of all I have demonstrated statements based on my databank made by myself so the data can be easily graphically reviewed. The results I have achieved were tested in different ways comparing the test-methods with each other about their limits.
Proceeding from the results I have created my remaining life estimating model, which can be used in general via parametering. The model is not bounded to any specific equipment(s).
The system is not recommended to be used as a stand-alone application so I have fit it in different ways to an expert system which I am familiar with. The utility of the cooperation is not only the more precise results also via data-sharing the algorythm can be implemented in such a form that it will be able to improve itself alone.
Some further developing opportunities can come up during the beta tests, which tests I consider the most important following step after planing, because in that way can be the validating and the verificating of the software assured certainly after the realization of it. My feature aims include the integration of the model to variable exsisting expert systems and the examination of the operation at different populations of equipments.