Kozlov L. Scientific foundations for designing the systems of manipulator hydraulic drives with an adaptive neural network-based controllers for mobile working machines

The thesis deals with statement, scientific substantiation and solution of the problem of improving static, dynamic and power characteristics of drives for mobile working machines through the development of new circuits based on variable pumps, proportional hydraulic equipment and adaptive neural networks. Scientific foundations of their development are created with the application of nonlinear mathematical models based on inertia tensors with fuzzy components. Correction of the adaptive controller parameters is performed using fuzzy dependencies. Peculiarities of the processes occurring in the mechatronic drives during simultaneous operation of two hydraulic engines and during the working member spatial motion have been determined. The ways for improving dynamic characteristics of hydraulic drives during adaptive regulated delays of the feedback signals are proposed as well as the ways for limiting their duration. Peculiarities of the operation of two regulators working simultaneously for one controlled object are determined, combinations of their parameters are found. An algorithm and a program for controlling the system of hydraulic drives are proposed, which provide operative commutation of hydraulic engines with variable drives in the working cycle of the machine and the respective reduction of unproductive power losses. Circuits and designs of the pump regulator and sectional directional control valve are developed for implementing the system of hydraulic drives for mobile working machines.