Tomas Vantuch

A power quality forecasting model as an integrate part of active demand side management using Artificial Intelligence Technique - Multilayer Neural Network with Backpropagation Learning Algorithm

This paper presents a power quality forecasting model with using Artificial Intelligence Technique, more precisely the Multilayer Neural Network with Backpropagation Learning Algorithm. This forecasting model is used as a supporting tool for a keeping of power quality parameters within the limits in the Off-Grid systems with renewables sources connected via AC By-Pass topology. Results of the most important power quality parameters forecasting are introduced in this paper. The developed algorithm of this model will be implemented into system for controlling the power flows inside the Off-Grid systems operated under Active Demand Side Management.

Dynamical properties of partial-discharge patterns

Abstract In this paper, Partial Discharge pattern as an indicator of the fault state of insulation systems of medium voltage overhead lines with covered conductors are described, analyzed, and their dynamical properties are researched. Application of data obtained in natural environment with huge variety of noise interferences, affected by various weather conditions, location and time of the day lead to questioning whether the PD-activity can be considered as a system with emergent-like behaviour. The complexity of obtained data and several signal types are examined and described in this contribution. As a main result, a complexity of signals is researched using approximate entropy, sample entropy, and correlation dimension. Finally, 0-1 test for chaos is used to show chaos of almost all signals and for one signal randomness is detected using newly applied stress test. In this paper, Partial Discharge pattern as an indicator of the fault state of insulation systems of medium voltage overhead lines with covered conductors are described, analyzed, and their dynamical properties are researched. Graphical Abstract In this paper, Partial Discharge pattern as an indicator of the fault state of insulation systems of medium voltage overhead lines with covered conductors are described, analyzed, and their dynamical properties are researched. As a main result, a complexity of signals is researched using approximate entropy, sample entropy, and correlation dimension. Finally, 0-1 test for chaos is used to show chaos of almost all signals and for one signal randomness is detected using newly applied stress test.

The Power Quality Forecasting Model for Off-Grid System Supported by Multiobjective Optimization

Measurement and control of electric power quality (PQ) parameters in off-grid systems has played an important role in recent years. The purpose is to detect or forecast the presence of PQ parameter disturbances to be able to suppress or to avoid their negative effects on the power grid and appliances. This paper focuses on several PQ parameters in off-grid systems and it defines three evaluation criteria that are supposed to estimate the performance of a new forecasting model combining all the involved PQ parameters. These criteria are based on common statistical evaluations of computational models from the machine learning field of study. The studied PQ parameters are voltage, power frequency, total harmonic distortion, and flicker severity. The approach presented in this paper also applies a machine learning based model of random decision forest for PQ forecasting. The database applied in this task contains real off-grid data from long-term one-minute measurements. The hyperparameters of the model are optimized by multiobjective optimization toward the defined evaluation criteria.

A novel approach of partial discharges detection in a real environment

The presence of partial discharges (PD) in medium voltage overhead lines with covered conductors (CC) may indicate insulation fault, rupture or downfall of the line. In a real environment (e.g. forested terrain), presence of PD usually means that branch or tree have direct contact with CC. The detection of the PD in a real environment is an important task because it helps technicians to accurately identify fault and therefore it has economical benefits. However, the automatic detection of PD in a real environment has to deal with high interference of background noise. A novel approach of PD detection based on Singular Value Decomposition (SVD) and Particle Swarm Intelligence (PSO) is proposed in this paper. Its performance is compared to Support Vector Machines (SVM), Back Propagation (BP) and Extreme Learning Machine (ELM) methods. These methods use standard features like number of peaks, fractal dimension, etc. The quality of the detection was successfully increased by the proposed approach. So the faults in a real environment can be more accurately detected and thus more economically repaired.

Dimensionality Reduction Method's Comparison Based on Statistical Dependencies

Abstract The field of machine learning deals with a huge amount of various algorithms, which are able to transform the observed data into many forms and dimensionality reduction (DR) is one of such transformations. There are many high quality papers which compares some of the DRs approaches and of course there other experiments which applies them with success. Not everyone is focused on information lost, increase of relevance or decrease of uncertainty during the transformation, which is hard to estimate and only few studies remark it briefly. This study aims to explain these inner features of four different DRs algorithms. These algorithms were not chosen randomly, but in purpose. It is chosen some representative from all of the major DRs groups. The comparison criteria are based on statistical dependencies, such as Correlation Coefficient, Euclidean Distance, Mutual Information and Granger causality. The winning algorithm should reasonably transform the input dataset with keeping the most of the inner dependencies.

A Novel Method for Detection of Covered Conductor Faults by PD-Pattern Evaluation

The partial discharge activity as a side effect of the current conductor’s disorder was taken into analysis to develop the correct classification of its behavior. This derived knowledge can decrease the risk of the possible damage on the environment caused by uncontrolled conductor’s failure. The preprocessing part of the experiment was the synthesis of non-linear features by genetic programming (GP). The inputs for GP were obtained by discrete wavelet transformation (DWT) of the signal data. This preprocessing phase was aimed to create the input values for the classification algorithm which was based on the artificial neural network (ANN).

Phase transitions in swarm optimization algorithms

Natural systems often exhibit chaotic behavior in their space-time evolution. Systems transiting between chaos and order manifest a potential to compute, as shown with cellular automata and artificial neural networks. We demonstrate that swarms optimisation algorithms also exhibit transitions from chaos, analogous to motion of gas molecules, when particles explore solution space disorderly, to order, when particles follow a leader, similar to molecules propagating along diffusion gradients in liquid solutions of reagents. We analyse these ‘phase-like’ transitions in swarm optimization algorithms using recurrence quantification analysis and Lempel-Ziv complexity estimation. We demonstrate that converging and non-converging iterations of the optimization algorithms are statistically different in a view of applied chaos, complexity and predictability estimating indicators.

Partial Discharge Pattern Classification Based on Fuzzy Signatures

The presence of partial discharge pattern in medium voltage overhead lines with covered conductors may indicate insulation fault, rupture or downfall of the line. These failures can cause problems in the electrical energy distribution to customers. This paper focuses on the detecting and classification of the partial discharge patterns. The presented method transform a captured input data into fuzzy signatures. This allows us to deal with the captured data as regular text documents. The obtained fuzzy signatures are used in the classification phase using k-NN. The proposed method can correctly classify the captured data up to 76% of accuracy.

An Examination of an Entropy Based Features on Partial Discharge Pattern

Presence of partial discharge pattern implies the fault behavior on insulation system of medium voltage overhead lines, especially with covered conductors (CC). This paper covers the examination of Approximation and Sample entropy as a signal complexity measures on partial discharge patterns of several kinds of faults. These features are calculated on multiple different adjustments such as the different applied denoising schemes and varying embedding dimensions. The final results reveal the splitting ability of these complexity measures on applied data.

Market Prices Trend Forecasting Supported By Elliott Wave’s Theory

The forecasting of the stock markets’ trends is one of the most frequently applied point of interests in machine learning (ML) in-dustry from its beginning. The theory of Elliott waves’ (EW) patterns based on Fibonacci’s ratios is also heavily applied in several trading strategies and tools which are available on the market and also there are many studies based on analysis and application of those patterns. This paper covers market’s trend prediction by ML algorithms such as Random Forest and Support Vector Machine. The trend prediction is supported by application of recognized Elliot waves which was performed by custom developed algorithm based on available knowledge about the patterns. The combination of ML algorithms and EW pattern detector achieved significantly higher performance compare to the ML algorithms only.