Oil products export control based on artificial neural network

The operation principle of the parametric leak detection system in oil pipeline systems is considered. The structure of this leak detection system is described. Factors affecting the accuracy of flowmeters measurement used in leak detection are analyzed. Characteristics and features of operation and installation of ultrasonic flowmeters are considered. The analysis of works which investigate creating diagnostic systems of flowmeters is carried out. Their advantages and disadvantages are described. The aim of the study is to develop a flow measurement system using indirect parameters, which could diagnose the state of its testimony and double measuring tools during their systematic maintenance or repair. A virtual sensor acts as a system and is implemented on the basis of software and is capable of duplicating the flowmeter. The structure of the neural network is given on the basis of which this system is built. A possible example of application of this measurement system has been considered. Application of neural network for solving this problem will allow to create a mathematical model of measurement that is not dependent on uncontrolled parameters. Implementation of this neural network solves the problem of artificial generation of signals from sensors used in the calculation of the system of detecting leaks at the time of their output. The preliminary results of the research are presented. The results obtained in the process of model realization are compared with experimental data.