The stochastic nature of renewable energy sources, especially solar PV output, has created uncertainties for the power sector. It threatens the stability of the power system and results in an inability to match power consumption and production. This paper presents a Convolutional Neural Network (CNN) approach consisting of different architectures, such as the regular CNN, multi-headed CNN, and CNN-LSTM (CNN-Long Short-Term Memory), which utilizes a sliding window data-level approach and other data pre-processing techniques to make accurate forecasts. The output of the solar panels is linked to input parameters such as irradiation, module temperature, ambient temperature, and windspeed. The benchmarking and accuracy metrics are calculated for 1 h, 1 day, and 1 week for the CNN based methods which are then compared with the results from the autoregressive moving average and multiple linear regression models in order to demonstrate its efficacy in making short-term and medium-term forecasts.

Modern frequency power converters generate a wide spectrum of harmonic components. Large converters systems can also generate noncharacteristic harmonics and interharmonics. Standard tools of harmonic analysis based on the Fourier transform assume that only harmonics are present and the periodicity intervals are fixed, while periodicity intervals in the presence of interharmonics are variable and very long. A novel approach to harmonic and interharmonic analysis, based on the "subspace" methods, is proposed. Min-norm harmonic retrieval method is an example of high-resolution eigenstructure-based methods. The Prony method as applied for signal analysis was also tested for this purpose. Both high-resolution methods do not show the disadvantages of the traditional tools and allow exact estimation of the interharmonics frequencies. To investigate the methods several experiments were performed using simulated signals, current waveforms at the output of a simulated frequency converter, and current waveforms at the output of an industrial frequency converter. For comparison, similar experiments were repeated using the fast Fourier transform (FFT). The comparison proved the superiority of the new methods. However, their computation is much more complex than FFT.

The spectrum-estimation methods based on the Fourier transform suffer from the major problem of resolution. The methods were developed and are mostly applied for periodic signals under the assumption that only harmonics are present and the periodicity intervals are fixed, while periodicity intervals in the presence of interharmonics are variable and very long. A novel approach to harmonic and interharmonic analysis based on the "subspace" methods is proposed. Min-norm and music harmonic retrieval methods are examples of high-resolution eigenstructure-based methods. Their resolution is theoretically independent of the signal-to-noise ratio (SNR). The Prony method as applied for parameter estimation of signal components was also tested in the paper. Both the high-resolution methods do not show the disadvantages of the traditional tools and allow exact estimation of the interharmonic frequencies. To investigate the methods, several experiments were carried out using simulated signals, current waveforms at the output of an industrial frequency converter, and current waveforms during out-of-step operation of a synchronous generator. For comparison, similar experiments were repeated using the fast Fourier transform (FFT). The comparison proved the superiority of the new methods.

The article presents calculations and power flow of a real virtual power plant (VPP), containing a fragment of low and medium voltage distribution network. The VPP contains a hydropower plant (HPP), a photovoltaic system (PV) and energy storage system (ESS). The purpose of this article is to summarize the requirements for connection of generating units to the grid. Paper discusses the impact of the requirements on the maximum installed capacity of distributed energy resource (DER) systems and on the parameters of the energy storage unit. Firstly, a comprehensive review of VPP definitions, aims, as well as the characteristics of the investigated case study of the VPP project is presented. Then, requirements related to the regulation, protection and integration of DER and ESS with power systems are discussed. Finally, investigations related to influence of DER and ESS on power network condition are presented. One of the outcomes of the paper is the method of identifying the maximum power capacity of DER and ESS in accordance with technical network requirements. The applied method uses analytic calculations, as well as simulations using Matlab environment, combined with real measurement data. The obtained results allow the influence of the operating conditions of particular DER and ESS on power flow and voltage condition to be identified, the maximum power capacity of ESS intended for the planed VPP to be determined, as well as the influence of power control strategies implemented in a PV power plant on resources available for the planning and control of a VPP to be specified. Technical limitations of the DER and ESS are used as input conditions for the economic simulations presented in the accompanying paper, which is focused on investigations of economic efficiency.

The International Electrotechnical Commission (IEC) standards characterize the waveform distortions in power systems with the amplitudes of harmonic and interharmonic groups and subgroups. These groups/subgroups utilize the waveform spectral components obtained from a fixed frequency-resolution discrete Fourier transform (DFT). Using the IEC standards allows for a compromise among the different goals, such as the needs for accuracy, simplification, and unification. In some cases, however, the power-system waveforms are characterized by spectral components that the DFT cannot capture with enough accuracy due to the fixed frequency resolution and/or the spectral leakage phenomenon. This paper investigates the possibility of a group/subgroup evaluation using the following advanced spectrum estimation methods: adaptive Prony, estimation of signal parameters via rotational invariance techniques, and root multiple-signal classification (MUSIC). These adaptive methods use variable lengths of time windows of analysis to ensure the best fit of the waveforms; they are not characterized by the fixed frequency resolution and do not suffer from the spectral leakage phenomenon. This paper also presents the results of the applications of these methods to three test waveforms, to current and voltage waveforms obtained from simulations of a real DC arc-furnace plant, and to waveforms measured at the point of common coupling of the low-voltage network supplying a high-performance laser printer.