No abstract

No abstract

No abstract

This paper studies neural networks in order to identify and control the traditional ball-plate problem. Firstly, a nonlinear model of ball and plate system consisting of two parts is established. Secondly, a multilayer perceptron neural network is employed to identify the plant. Next, a feedback controller is designed based on neural network method to control the system. Eventually, simulations are accomplished via Matlab/Simulink and results show the remarkable ability of identifier and effectiveness of the proposed neural networkbased controller.

Electric train system is a very large load for the power network. This load consumes a large amount of reactive power. In addition, it causes a massive unbalance to the network, which results in many problems such as voltage drops, high transmission losses, reduction in the transformer output ability, negative sequence current, mal-operation of protective relays, etc. In this paper, a novel real-time optimization approach is presented to adjust the static VAR compensator (SVC) for the traction system to realize two objectives; current unbalance reduction and reactive power compensation. A multi-objective optimization technique entitled non-dominated sorting genetic algorithm (NSGA-II) is used to fulfill the regarded objectives simultaneously. A comprehensive simulator has been designed for electric train network modeling that is able to adjust the parameters of SVC in an optimum manner at any time and under any circumstances. The results illustrate that the provided method can efficiently reduce the unbalancing in current as well as supply the demanded reactive power with acceptable precision.