Diagnostika průmyslových systémů

Abstract

The thesis deals with the reliability of the vacuum pump diagnostics. Its goal is to provide information about test effectiveness and recommend new test limits. The thesis describes very complex field, so it was necessary to grasp the theoretical knowledge from several fields - neural networks, reliability, diagnostics, programming, sensors, technical characteristics and applications of the tested product. The theoretical part is divided into logical units. First, the general criteria and reasons why industrial diagnostics are important are described. Then, the theoretical characteristics are specific to the test process. This process consists of: vacuum pump, test station, sensor. These elements are described and analyzed in detail. The practical part of the thesis deals with the actual testing method and the boundary limits, which classify the same and the non-conforming product. Based on a large volume of test data, it was created using neural networks using a model that specifies the classification. The output of the work is a window application that makes it possible to compare existing test results with newly created and recommended tests. A diagnostic program has been developed to analyze the test limits settings, which will be updated on a daily basis for each product produced for maximum transmission.