Marinko Barukčić

Hybrid Evolutionary-Heuristic Algorithm for Capacitor Banks Allocation

Hybrid Evolutionary-Heuristic Algorithm for Capacitor Banks Allocation The issue of optimal allocation of capacitor banks concerning power losses minimization in distribution networks are considered in this paper. This optimization problem has been recently tackled by application of contemporary soft computing methods such as: genetic algorithms, neural networks, fuzzy logic, simulated annealing, ant colony methods, and hybrid methods. An evolutionaryheuristic method has been proposed for optimal capacitor allocation in radial distribution networks. An evolutionary method based on genetic algorithm is developed. The proposed method has a reduced number of parameters compared to the usual genetic algorithm. A heuristic stage is used for improving the optimal solution given by the evolutionary stage. A new cost-voltage node index is used in the heuristic stage in order to improve the quality of solution. The efficiency of the proposed two-stage method has been tested on different test networks. The quality of solution has been verified by comparison tests with other methods on the same test networks. The proposed method has given significantly better solutions for time dependent load in the 69-bus network than found in references.

Multi objective optimization of energy production of distributed generation in distribution feeder

Distributed generation (DG) based on renewable sources becomes more and more implemented in distribution networks. Some of these sources have non constant (wind power plants, photovoltaic planes) and some have a constant power production (biogas plants). Also, loads on a feeder change during the day. There is a need for different optimizations because a number of variable changes over time. The total daily active energy losses, daily financial profit, and total daily active energy production of DG are considered for optimization in the paper. The load changes are considered on day level and for each load separately (not on the feeder level, but on node level). The multiobjective approach is applied in the paper for optimizing. The evolutionary strategy is utilized as the optimization method. The two and three objective optimization problems are presented and solved in the paper. The IEEE 13 node unbalanced distribution test feeder is used for presentation of multiobjective optimization. The proposed procedure is performed using OpenDSS and MATLAB software.

Battery Pack Design Problems – Influence of the Transverse Magnetic Field on Internal Battery Resistance

The battery pack represents an essential part of electric vehicles (EV), so its design significantly influences the drive characteristics. The system of energy flow control in EV use highfrequency power electronic components that cause the appearance of high harmonics super-imposed on DC component of the battery current. This paper examines the influence of the distribution of the transversal magnetic field inside of the battery pack depending on battery pack design and its influence on battery parameters. DC component of the transversal magnetic field inside of battery cells cause Hall effect ions path allocation that is reflected on internal battery resistance negative change. AC component of the transversal magnetic field in electrically closed loops within battery pack induced e.m.s. that also cause negative phenomena on battery pack parameters.

The evolutionary optimization approach for voltage profile estimation in a radial distribution network with a decreased number of measurements

The procedure for estimation of voltage profile in a radial distribution network considering decreased number of measurements is presented in the text. The proposed procedure is based on hypothesis that feeder voltage profile in case of the limited number of feeder measurements can be successfully estimated using evolutionary optimization approach. The procedure consists of two main steps. In the first step, the location of the voltage measurements are determined with aim to have as small as possible the number of the locations. This is done by applying the analytical mathematical procedure. After the measurement locations are defined, in the second procedure step the optimization aims to estimate whole voltage profile is performed employing the evolutionary algorithm. The proposed procedure is tested on two test radial distribution feeders to verify the proposed procedure.

Output characteristics analysis of gensets for modeling of an expert system for fault detection purposes

In complex systems, such as military organizations, a large number of technical resources requiring regular maintenance is used. These resources are often exposed to difficult working conditions increasing the probability of a fault occurrence. Fault diagnosis as a part of the maintenance process is often a complex process, which is why expert systems for fault diagnosis are developed. One of the possible systems developed for fault diagnosis on gensets uses the Bayesian network to decide on a fault with the highest probability. For proper system operation, it is important to have a reliable “knowledge base” encompassing results of diagnostic tests. Apart from direct measurements, simulation results are also used in order to obtain diagnostic results. Fault simulations are performed on a model developed using the Matlab Simulink program package.

Evolutionary optimization approach for performing interval power flow considering uncertainties in electric power systems

The paper deals with problem of power flow calculation if there are uncertainties present in power distribution network variables or parameters. Load, network parameters and supply uncertainties can be represented by ranges of values. From mathematical point of view these ranges can be defined as intervals, fuzzy numbers or a probability density function. Since the input data are interval instead crisp numbers, the question is what bounds of the power flow output variables are. Literature suggests various methods such as probabilistic power flow, bound power flow and interval and fuzzy arithmetic for calculating power flow considering uncertainties. The proposed approach is based on optimization problem formulation. The presence of uncertainties is defined in optimization problem form with minimization and maximization of power flow output quantities as objective functions of the optimization. The uncertainties in input quantity are taken into account through optimization problem inequality constraints. The evolutionary algorithm is used to solve the optimization problem. The standard power flow calculation for crisp numbers is used instead power flow based on interval or fuzzy arithmetic. The proposed procedure is realized through co-simulation between PSAT and MATLAB computer programs. The proposed approach is tested on IEEE 14 node test distribution feeder. The simulation results indicate that proposed procedure has capability to solve considered problem.

Applying an evolutionary strategy for multiobjective optimization of capacitor banks allocation in distribution feeders

This paper aims at applying the Evolutionary Strategy (ES) to capacitor banks optimal allocation problem in the case of multiobjective optimization. The multiobjective optimization includes several objectives such as power losses, capacitor costs and voltage deviation. The multiobjective optimization problem consists of conflicting objectives which need to be simultaneously optimized. Some other constraints are also considered. In the paper the following is proposed: the ES mutation operator adapted to the optimization problem, individual coding and generating of the initial population and finally, the mutation operator adapted to the optimization problem with discrete variables. The proposed approach has been evaluated on the test radial feeder and simulation results have been compared with those existing in the literature.