Marinko Barukcic earned his BsC, MsC, and PhD degrees in 1998, 2008, and 2012, respectively, from the Faculty of Electrical Engineering, Computer Science, and Information Technology in Osijek, Croatia. From 1998 to 2007, he worked at the Croatian Electricity Distribution System Operator. From 2007 to present, he has been working at the Faculty of Electrical Engineering, Computer Science and Information Technology in Osijek. Marinko is involved in several scientific projects as a project leader and member of a project team. At the Faculty he lectures in various courses for Bachelor, Master and PhD degree programs. He is the author of more than 80 scientific articles in journals and conferences. His scientific interests include modeling, analysis and simulation of power systems, electrical machines and computation of electromagnetic fields. He is particularly interested in the application of computational intelligence techniques in electrical engineering.
Title: Computational Intelligence and Electrical Engineering
Abstract: Computational Intelligence (CI) is a term that describes algorithms and techniques that mimic various phenomena from nature by describing them through mathematical procedures and models. The three main areas of CI techniques are artificial neural networks, fuzzy inference systems, and evolutionary computation. Thanks to the development of computer hardware and CI methods and techniques, CI is now increasingly used in areas of electrical engineering, especially in new concepts such as Smart Grid and Industry 4.0, and CI tools and also provide a new (improved) approach to solving some "classical" problems. CI is now used in all areas of electrical engineering and for a variety of purposes, including modeling, control, and estimation of electrical machines, power grids, electronic devices, electromagnetic calculations, and more. The application of computational intelligence methods in various segments of electrical engineering enables the successful handling of a large amount of data, uncertainties, and missing information present in modern electrical systems and devices. CI procedures enable the solution of problems with a limited amount of measured data, more realistic modeling of physical systems, the application of new control and estimation methods, and problems that cannot be solved in a classical, analytical way. In addition, computational intelligence methods enable the simultaneous use of several different computer tools in the so-called co-simulation setup to perform simulations and calculations of complex modeled systems. It is expected that the applications of CI processes will be increasingly used thanks to the advancement of CI techniques and their adaptation for use in the field of electrical engineering.
Attila Magyar obtained his MSc and his PhD from the University of Pannonia in 2004 and 2008, habilitated at the University of Pannonia in 2017. Presently he is an associate professor at the University of Pannonia, Department of Electrical Engineering and Information Systems, his teaching activity includes systems and control theory and robotics for Bachelor, Master and PhD programs. He is the secretary of IEEE Control Systems Society, Hungarian Section and the head of the Electrical Energy Research Lab at tha University of Pannonia. He is the author of more than 60 scientific publications, his main research interest is the use of nonlinear control theory and optimization in the field of electrical energy systems.
Title: Direct Optimization in Electrical Energy Systems
Abstract: In recent decades, several countries have changed their laws regulating power supply to allow for grid-tie inverter systems to provide spare power to local low voltage electrical grids. This power is utilized locally, decreasing electrical power loss due to transmission. In addition, grid-tie inverters are suitable for conditioning power lines, correcting accurate voltage forms, and repairing reactive power in the mains. This
decreases losses further, given that nonlinear distortion in the mains induces losses in both the phase and the neutral conductor. This additional functionality does not require expensive changes to existing technology. Only the control methods and regulators need to be modified in order to allow for line conditioning. The cost of changing the controlling processor and control software in this system is negligible when compared
to the cost of changing equipment.
Although model based methods are promising tools for the optimal operation of energy systems, in some cases the properties of system to be controlled disables the development of a computationally effective and precise model. The expectation of optimal operation nonetheless exists for such systems. Electrical networks with stochastically changing loads and generators are difficult to describe
with dynamical models. On the other hand, power quality problems such as total harmonic distortion due to the increasing number of nonlinear capacitive loads, or voltage unbalance problems due to the non-uniform distribution of prosumers on the phases provide several optimization problems to be solved in a model free manner.
With the growing popularity of distributed electricity generation, injecting current (mostly from renewable sources) into the electrical grid is getting widespread. One of the problems investigated is to find an optimal way of current injection (single phase) to the grid so that the total harmonic distortion decrease. There supposed to be no information available from the grid that is why a model-free compensation method has to be developed. On the other hand, in a three-phase network, voltage unbalance is an important indicator of power quality. The above problem of optimal current injection to the grid isinvestigated for the three phase network case where the optimality criterion is a measure of voltage unbalance.