Diagnosis of abnormal operation modes of a machine utilizing self organizing map

A diagnosis apparatus for diagnosing abnormal machine operation modes being performed by a machine, comprising:

• detecting means for detecting a plurality of data sets that each correspond to a different one of a corresponding plurality of different machine operation modes performed by the machine, with each of the plurality of data sets including values of n parameters that each vary in accordance with an operation of the machine in each one of the different corresponding plurality of machine operation modes;
• self-organizing map creating means for creating self-organizing maps serving as separation models corresponding to each one of the plurality of different machine operation modes using a plurality of the data sets detected by said detecting means while the machine performs a preliminary operation of a normal state relative to each one of the different machine operation modes;
• abnormal data creating means for creating a multiplicity of abnormal data sets for each one of the plurality of data sets, the multiplicity of abnormal data sets representing virtual abnormal states in which each one of the plurality of different machine operation modes has abnormality, created by modifying the values of the n parameters of the plurality of data sets detected during the preliminary operation of each normal state with a number of different deviation vectors each of which has n deviation components;
• and
• operation mode proportion calculating means for calculating an operation mode proportion vectors for each of the deviation vectors, which abnormal operation mode proportion vector includes a plurality of proportions, one representing each of the plural operation modes, by selecting a self-organizing map which has the highest similarity degree to each of data sets which have been input among the self-organizing maps created by said self-organizing map creating means;
• and
• judging means for judging abnormality of the machine while the machine performs actual practice by obtaining a number of actual practice data sets detected by said detecting means, selecting a self-organizing map from the self-organizing maps created by said self-organizing map creating means which has the highest similarity degree relative to each of the actual practice data sets corresponding to each of the plurality of different machine operation modes, obtaining an actual-practice operation mode proportion vector including a plurality of proportions representing each one of the plurality of different machine operation modes, selecting an abnormal operation mode proportion vector which has the highest similarity degree to the actual-practice operation mode proportion vector from the abnormal operation proportion vectors calculated by said operation mode proportion calculating means, and obtaining each deviation component values of the abnormal operation proportion vector which has been selected.