Approach to the selection of ship compressor equipment using statistical data processing methods

Abstract

Compressors are necessary to provide consumers of ship power systems and the ship as a whole with compressed air of various pressures and flow rates. The correct functioning of water transport infrastructure depends on the correct selection of compressor equipment. This makes it necessary to consider this issue in detail. If all the variables, the compressor selection parameters under consideration, were measured in the same scales and units of measurement, it would be possible to add up all the values, but this approach is very rough. The output is to normalize the values of variables, and then calculate the final criterion based on them. The article deals with the selection of compressor equipment for the ship's starting and working air systems using exponential and linear normalization methods. For statistical processing, certain parameters of compressors from domestic and foreign manufacturers are proposed. The result of the study is to obtain one formal criterion the final rating, instead of several qualitative parameters. Based on the obtained value, the optimal compressor is determined by the combination of its characteristics. The next stages of data processing are the proposal of maximum likelihood and logistic regression methods. The probability of assigning a compressor with a certain set of characteristics to a positive or negative class is simulated. The model was trained based on the available training data for selecting compressors.