Development of an intelligent complex for stabilization column parameters adaptive control of the catalytic reforming stabilization unit

The solution of control problems for complex technical systems or objects is based on the development of their mathematical models, which further determine the control algorithm. For auto-correction of the control algorithm, it is advisable to use adaptive approaches in control theory – artificial neural networks. The use of a neural network controller, which combines adaptive self-learning approaches and the experience of an expert – adjuster of an automated process control system, seems appropriate when introducing control of complex multifunctional objects. Purpose of the research is development of an adaptive control system by creating neural networks based on the data of the catalytic reforming stabilization process in the Matlab Simulink environment and their comparison. Neural networks are used in industrial units in the presence of disturbing influences on the control object, when traditional solutions in control systems are not effective enough. Electric drive complexes control using neural networks is advisable when the speed or precision of traditional control systems with linear controller is insufficient. The result of the research is development of an adaptive control systems based on neural networks in the Matlab system. Their comparative analysis was carried out to control the stabilization process of catalytic reforming.