Search

The digitization of distribution power systems has revolutionized the way data are collected and analyzed. In this paper, the critical task of harnessing this information to identify irregularities and anomalies in electricity consumption is tackled. The focus is on detecting non-technical losses (NTLs) and energy theft within distribution networks. A comprehensive overview of the methodologies employed to uncover NTLs and energy theft is presented, leveraging measurements of electricity consumption. The most common scenarios and prevalent cases of anomalies and theft among consumers are identified. Additionally, statistical indicators tailored to specific anomalies are proposed. In this research paper, the practical implementation of numerous artificial intelligence (AI) algorithms, including the artificial neural network (ANN), ANFIS, autoencoder neural network, and K-mean clustering, is highlighted. These algorithms play a central role in our research, and our primary objective is to showcase their effectiveness in identifying NTLs. Real-world data sourced directly from distribution networks are utilized. Additionally, we carefully assess how well statistical methods work and compare them to AI techniques by testing them with real data. The artificial neural network (ANN) accurately identifies various consumer types, exhibiting a frequency error of 7.62%. In contrast, the K-means algorithm shows a slightly higher frequency error of 9.26%, while the adaptive neuro-fuzzy inference system (ANFIS) fails to detect the initial anomaly type, resulting in a frequency error of 11.11%. Our research suggests that AI can make finding irregularities in electricity consumption even more effective. This approach, especially when using data from smart meters, can help us discover problems and safeguard distribution networks. [ABSTRACT FROM AUTHOR]

Metal-oxide surge arresters (MOSAs) play an important role in preventing outage situations and thus in preserving the reliability of electric power system. So far, many MOSA monitoring methods have been developed. These methods propose different indicators that are more or less reliable in MOSA monitoring and diagnostics. Another important aspect of these indicators is that each one of them can be determined with a certain estimation error. None of the indicators has been proved to be supreme. Nevertheless, under different conditions, the reliability of these indicators may vary. This paper presents an analysis methodology for reliability of MOSA indicators and proposes a fuzzy logic expert system that overcomes the problem of decision-making in the case when more than one indicator is available. The analysis and verification are done by computer simulations. [ABSTRACT FROM AUTHOR]

This paper proposes a fuzzy expert system for demand-side management, management of renewable energy sources, and electrical energy storage for smart households and microgrids. The proposed fuzzy expert system is used for automatic decision making regarding energy management in smart microgrids containing renewable sources, storage systems, and controllable loads. The fuzzy expert system optimizes energy consumption and storage in order to utilize renewable energy and maximize the financial gain of a microgrid. In order to enable energy management, the fuzzy expert system uses insolation, price of electrical energy, temperature, wind speed, and power of the controllable and uncontrollable loads as input variables. These input data can be directly measured, imported from grid measurements, or predicted using any data prediction method. This paper presents fuzzification of input variables, defines a set of rules of the expert system, and presents defuzzification of outputs. The outputs of the expert system are decisions, i.e., answers to the question of how to manage energy production and consumption in a microgrid. Three outputs are defined to decide about produced energy, controllable loads, and own consumption. The first output is used to store, sell, or consume produced energy. The second output is used to manage the controllable load. The third output shows how to supply own consumption of the prosumer. The expert system is tested on hourly values of input variables in a single day in Serbia. The proposed approach is compared with other available approaches in order to validate the results. [ABSTRACT FROM AUTHOR]

This paper deals with the problem of metal-oxide surge arrester (MOSA) diagnostics. The proposed methodology is based on a genetic algorithm optimization technique. The genetic algorithm (GA) is applied to determine the parameters of MOSA equivalent model. According to the assessed parameters, which change during the lifetime of the MOSA, the MOSA condition can be determined. For this purpose, the realistic measurement data of the operating voltage and the leakage current is used. Additionally, computer simulations are conducted in order to demonstrate the influence of the operating voltage harmonics on the algorithm performance and dynamics. The proposed algorithm is tested on realistic MOSA in laboratory conditions. [ABSTRACT FROM AUTHOR]

This paper deals with the problem of optimal wind turbine micro siting on complex terrains. A realistic wind farm terrain can be very complex not only for the topography of the terrain but also for other factors. Such factors are the environmental impact of the wind farm, the costs of land, the costs of access roads, wind turbine installation and maintenance costs, the costs of electrical infrastructure, etc. A grid-based wind farm model that takes all of these factors into consideration is proposed. For the purpose of an optimization algorithm, a doublestage genetic algorithm is proposed. In the first stage, the optimal micro-areas for wind turbine positioning are determined, whereas the second stage defines precise optimal positions of wind turbine foundations. The double-stage genetic algorithm increases the accuracy of wind turbine positioning and decreases the dimensionality of the problem. The algorithm has been tested on a realistic wind farm terrain in Serbia. [ABSTRACT FROM AUTHOR]

This paper presents a software tool for evaluation of electrical energy generated by photovoltaic (PV) systems. The software was developed using the MATLAB software package and contains the elements of an expert system. The software is designed for engineers who are involved in photovoltaic system design and for students (undergraduates and Masters) of electrical engineering who, using the software, adopt theoretical assumptions and solve practical engineering problems. The structure, organization and software capabilities are illustrated within the example of the design of an actual photovoltaic system. [ABSTRACT FROM AUTHOR]

The power flow calculation considering uncertainty is common way of determining harmonic analysis of the system under uncertain conditions. Distributed generation and load are characterised by randomness. Hence, this study introduces uncertainty to the distribution network analysis. Three similar methodologies are proposed for distribution power flow, which calculate active power losses, voltage drops and voltage total harmonic distortion. The first methodology uses fuzzy logic in order to address uncertainties. Generation, active power loss variation and elasticity of power quality constraints are presented in a form of fuzzy numbers (FNs). The second methodology uses Monte Carlo simulation (MCS) to vary the values of the same variables, which are FN. The third methodology combines MCS and FN. The proposed methodologies are verified using case studies based on IEEE 33-bus system. The results of all three methodologies are compared graphically and numerically. Analysing the distribution harmonic load-flow results, one can find out the weak links of the network, which can provide references for making electrical accident premeditation and power quality problems. [ABSTRACT FROM AUTHOR]

The power flow calculation considering uncertainty is common way of determining harmonic analysis of the system under uncertain conditions. Distributed generation and load are characterised by randomness. Hence, this study introduces uncertainty to the distribution network analysis. Three similar methodologies are proposed for distribution power flow, which calculate active power losses, voltage drops and voltage total harmonic distortion. The first methodology uses fuzzy logic in order to address uncertainties. Generation, active power loss variation and elasticity of power quality constraints are presented in a form of fuzzy numbers (FNs). The second methodology uses Monte Carlo simulation (MCS) to vary the values of the same variables, which are FN. The third methodology combines MCS and FN. The proposed methodologies are verified using case studies based on IEEE 33‐bus system. The results of all three methodologies are compared graphically and numerically. Analysing the distribution harmonic load‐flow results, one can find out the weak links of the network, which can provide references for making electrical accident premeditation and power quality problems. [ABSTRACT FROM AUTHOR]