The Prediction Of The Centrifugal Compressor Based On The Theory Of ANFIS

The Centrifugal compressor has been widely used in refrigeration industry, petroleum chemical industry, metallurgical industry, power industry, natural gas transport and other industrial sectors and it plays an important role in the national economy. It is also the core equipment in the oil gas injection, ammonia, urea synthesis, sintering fan, supercharged engine and other vital areas. As the condition of the production practice has an important influence on the performance of the compressor, it is quite important to build the mathematical model which is based on the actual operation of the machine. Originally, the modeling work of the compressor is always based on empirical formulas or the theory of fluid mechanics. But as the performance of the centrifugal compressor is influenced by many factors, the accuracy of the traditional modeling methods can’t meet actual requirements. As a result, a new method of modeling should be used to build the mathematical model of the compressor.In this thesis, the object of study is the Ningxia Petrochemical Company’s urea production plant which is a carbon dioxide centrifugal compressor. By analyzing the principle of centrifugal compressors, the characteristic variables which have a close relationship with the performance are chosen and the inputs and outputs of the model are chosen form the characteristic variables which are mentioned above. Based on a large number of actual measured data, the method which name is ANFIS is used to build the model of the compressor and the method is also used to predict the key parameters of the compressor.By analyzing the centrifugal compressor mathematical model simulation, the controller parameters are tuning under the stable system and it provides the basis for setting the parameters for the actual controller. As it’s showed in the simulation results, the compressor’s model which is based on the theory of ANFIS can actually reflect actual operating conditions of the centrifugal compressor and it can do better in learning, mapping and predicting than the traditional method. All in all, there is a certain reference value in the method which is described above.