Optimization of Energy Consumption for Light Rail Electric Transport with Supercapacitor System Using Neural Network Controllers

In this paper the solution of energy saving problem is proposed. To achieve useful utilisation of regenerative energy and reduce the overall energy consumption, the braking energy should be temporarily saved in an energy storage system ESS, based on supercapcitor, until another power consumer is connected to the overhead line. Such a storage system is able to cope with the common task of peak power reduction and overhead voltage stabilization. ESSs could be installed stationary at substations, weak spots of network or on-board vehicle. The purpose of this paper is to develop model for transport control system is to coordinate energy consumption of multiple various participants of traffic. The transport system is a cooperative system, where behaviour of one participant depends on other. Therefore, the negotiation process between vehicles is necessary. Each vehicle has electronic device which controls his own object sending data to the control centre, which is responsible for optimization and coordination of negotiation process. Controller for speed prediction of electric transport is based on nonlinear autoregressive neural network with exogenous inputs. Inputs of controller are current tram speed, distance from the beginning of the route, type of next speed change point, light upon arrival to the traffic light on the way, time interval between current time and directive time. Training set for the controller is received from the simulation model of T3A tram, moving on the part of route containing two passenger stops and traffic light. Neural network is trained and results of its workability are proposed. Usage of neural network gives possibility to predict actions of all participants of electric transport flow and allow using ESS more efficiently