Resumen
A problem of development of a method of optimal control of operation of natural gas superchargers by a criterion which both minimizes fuel gas consumption and takes into account power of nitrogen oxide emissions into atmosphere has been formulated. Technical conditions of gas pumping units, restrictions on technological parameters and the requirement to provide a planned indicator of gas pumping by a group of parallel-operated superchargers were taken into consideration.Technical condition of each unit or its assembly is assessed based on certain attributes. If observation of such attributes is made for a certain period of time, a set of attributes is obtained. Using an artificial neural network of Kohonen type, the set of attributes (images) is divided into three classes. A certain number of points is assigned to each class. This number characterizes technical condition of the class. The number of points assigned determines utilization of each supercharger. This is taken into account when limiting overall performance of the supercharger group.Formalized record of the optimal control problem contains dependences that are approximated by a polynomial of a specified order. This results in an empirical model whose structure is determined using an apparatus of genetic algorithms.For a series of reasons (errors in measuring technological parameters, errors in measurement methods, external effects, limited scope of experimental material, etc.) identification of values of empirical model parameters is based on inaccurate information. Therefore, parameters of empirical models are treated as fuzzy quantities. Based on the adopted concept, a formalized record of the problem of optimal control of operation of natural gas superchargers was obtained.Implementation of the study results will help save fuel gas and reduce nitrogen oxide emissions into environment