English (Greek will also follow below)
This paper deals with the design of a Hybrid Power Source Management System, consisted of (but not limited to) three different power producing sources surrounded by the appropriate control logic modules. The RES model is developed around a Wind Turbine, an array of Photovoltaic cells and a Fuel Cell stack, all interconnected by the means of the paper’s proposed architecture. The system converts wind, solar and chemical power (inputs) into electrical power and supplies it in a way that it can be fed either on a grid or an autonomous worksite (outputs). The unpredictable (stochastic) behavior of both the power supplement and the power consumption sides introduce a considerable control challenge. The system must be able to process and deliver the power needed by the facilities without any interruptions or wastes, and it must always deliver in the most efficient and conditioned way possible (regardless of all the changes that occur on instant power and load values).
Special care was given to purge any instability issues that are introduced as the result of various uncertainties, such as intensity and direction of wind, rain, temperature and fuel composition heterogeneity. These phenomena could interfere to the proper function of the system by giving random input signals, while not excluding the possibility of a hardware failure mishap.
For all of the above mentioned reasons involved, the paper proposes the introduction (and study) of decentralized Fuzzy Logic Control, the use of Artificial Neural Network mechanisms and the improvement of the P&O algorithm (by using steps of variable length). All the comparisons show that for the majority of the scenarios developed, the performance of the proposed system is found to be superior by outperforming any other conventional non-linear control techniques.
Η παρούσα εργασία ασχολείται με την σχεδίαση ενός υβριδικού συστήματος διαχείρισης των πόρων ήπιων μορφών ενέργειας αποτελούμενο