L. Dusonchet

The aim of the paper is to evaluate the technical feasibility and aspects related to the use, in an uninterruptible power supply (UPS), of fuel cells instead of traditional electrochemical batteries. The proposed system, named UPS-FCB (fuel cell based) presents a modular and versatile system configuration; particularly it is possible to integrate it with traditional UPS. Moreover the system features can be modified in terms of power, reliability, back-up time, etc., by changing the power devices. A detailed experimental analysis has been carried out on a passive stand-by UPS integrated with a FCB; experimental results can be easily extended to the other possible system architectures.

This paper presents a new formulation including robustness of solution of constrained multiobjective design or reactive power compensation. The algorithm used for optimization is the NSGA-II (non dominated sorting genetic algorithm II) with a special crowded comparison operator for constraints handling. The need for including the issue of robustness of solutions derives from the simple observation that loads are uncertain in distribution systems and their estimation is often affected by errors. In design problems it is desirable to consider the loads with a certain range of variation. In this paper the NSGA-II algorithm is applied to efficiently solve the issue and the solutions attained considering stochastic loads are compared to those considering precise loads

In this paper, a methodology to treat constrained scheduling problems based on the repeated application of a fuzzy-logic-based multiobjective algorithm is presented. The application domain is that of automated distribution systems management. In particular, the problem of voltage regulation and power loses minimization is here considered. The classical formulation of the problem of optimal control of shunt capacitor banks and under load tap changers, ULTC, located at high voltage/medium voltage (HV/MV) substations has been coupled with the optimal control of tie-switches and capacitor banks on the feeders of a large radially operated meshed distribution system with the aim of attaining minimum power losses and the flattening of the voltage profile. The considered formulation requires the optimisation of two objectives; therefore the use of adequate multiobjective heuristic optimisation methods is needed. The heuristic strategy used for the optimisation is an evolutionary algorithm in which the objectives formulation is based on fuzzy sets theory. After a brief description of the general problem of multiobjective optimisation in automated distribution networks management problems, the most recent papers on the topic are reported and commented. Then, the problem if formulation and the solution algorithm are described in detail. Finally, test results on a large distribution system are reported. They demonstrate that the proposed formulation and approach are effective and feasible for finding a good dispatching schedule for the considered system.