Search academic texts

Information × Registration Number 0214U005051, 0111U001230 , R & D reports Title The pilot neural network system for instant synfhesis of optimal control of melting operational modes in electric arc furnace popup.stage_title Head Lozynsky O. Y., Registration Date 30-01-2014 Organization Lviv Polytechnic National University popup.description2 Expediency of multicriterion optimal control strategies implementation for nonlinear nonstationary objects on the basis of fuzzy sets theory is substantiated. A system for electric arc furnace arcs length control based on fuzzy regulator has been implemented and method of regulator synthesis has been worked out. Obtained results showed improvement in dynamic regulation of arcs lengths during the use of fuzzy correction. The structure of three-contour fast-acting electromechanical system for positioning and dynamic stabilization of coordinates based on pulse-width converter and implementing fuzzy correction of control signal is proposed. The theoretical basis of fuzzy regulators based multicriterion optimal control systems synthesis and analysis have been developed. Additive functional model for optimal control utilizing fuzzy controller for changing weights of partial criteria has been worked through. The influence of membership functions parameters on dynamic indices has been studied. The structure of arc lengths control system based on neural controllers NARMA-L2 Controller, NN Predictive Controller have been and Model Reference Controller developed. Neural controllers synthesis for electric arc furnace arc lengths control system have been done and numeric simulations on created mathematical and digital models have been carried out. Obtained results showed improved accuracy of electric arc furnace electric mode coordinates dynamic stabilization while using neural controllers. A method of functional based multicriterion control signal operative synthesis is proposed, as an additive linear combination of partial criteria with variable in time weights through the use of fuzzy controllers. An approach for the formation of fuzzy controller belonging function form, which provides not only the optimum transition to the desired level of performance, but also the desired behavior of the system under the influence of external disturbances. The method of neural network based dynamograms recognition for rod deep oil well pumping installation is proposed. The structure of the optimal control of rod deep pumping plant based on neural network is proposed. A mathematical and numerical model of rod deep pumping installation has been developed and its dynamics has been explored. To create a sensorless control system of switched reluctance motors (SRM) with a more simple structure of the neural network than existing ones, a feedforward artificial neural network (ANN) and Kohonen network, that uses currents as the input signals, were synthesized. Also based on the feedforward network a neuro-estimator, that provides a flat speed-torque characteristics of SRM, was modeled. The work of the sensorless control system using created ANNs was researched by computer simulation . To verify the simulation results, an experimental apparatus was built. The artificial neural network for majority voting block operation of fractional-fold redundancy systems is developed. Such network provides correct control signal selection and non-operational block of system identifying. The reliability mathematical models of systems with fractional-fold redundancy are developed. Such models adequately taking into account the critical and non-critical failures of structural components, non-exponential character life of components, as well as load-sharing impact. Product Description popup.authors Андреїшин А.С. Бобечко Ю.О. Головач І.Р. Демків Л.І. Лозинський А.О. Лозинський О.Ю. Маляр А.В. Марущак Я.Ю. Мацигін А.Б. Михайлович Т.І. Мороз В.І. Паранчук З.Л. Паранчук Р.Я. Паранчук Я.С. Цяпа В.Б. Щербовських С.В. popup.nrat_date 2020-04-02 Close