This article highlights the industry experience of the development and practical implementation of a short-term photovoltaic forecasting system based on machine learning methods for a real industry-scale photovoltaic power plant implemented in a Russian power system using remote data acquisition. One of the goals of the study is to improve photovoltaic power plants generation forecasting accuracy based on open-source meteorological data, which is provided in regular weather forecasts. In order to improve the robustness of the system in terms of the forecasting accuracy, we apply newly derived feature introduction, a factor obtained as a result of feature engineering procedure, characterizing the relationship between photovoltaic power plant energy production and solar irradiation on a horizontal surface, thus taking into account the impacts of atmospheric and electrical nature. The article scrutinizes the application of different machine learning algorithms, including Random Forest regressor, Gradient Boosting Regressor, Linear Regression and Decision Trees regression, to the remotely obtained data. As a result of the application of the aforementioned approaches together with hyperparameters, tuning and pipelining of the algorithms, the optimal structure, parameters and the application sphere of different regressors were identified for various testing samples. The mathematical model developed within the framework of the study gave us the opportunity to provide robust photovoltaic energy forecasting results with mean accuracy over 92% for mostly-sunny sample days and over 83% for mostly cloudy days with different types of precipitation.

The paper addresses the problem of operated power transformers actual technical state identification by using existing technical diagnostics retrospective data. The initial data set was composed of results of transformer oil analysis, loading conditions, infrared snapshots and aggregated characteristics of the technical state of bushing, surge arrester and cooling system. Retrospective data of technical diagnostics was used. Technical state estimation process was divided on several steps and was performed by Python. First step was feature selection, where the features with low meaning in terms of the problem formulation are ejected. The second step is the processing of missing data to increase the dataset. KNN algorithm was used to restore missing values. The final step is learning transformer technical state classifier based on random forest tree approach. The results of power transformer state's classification demonstrated relatively high accuracy of identification.

The paper presents an alternative approach to voltage stability boundaries evaluation. The proposed method is direct, but has advantages over conventional direct methods. It is based on optimization technique and allows to evaluate the distance to power system voltage stability boundary. The paper demonstrates the concept of the approach based on simple power system models. Also implementation of the approach is demonstrated with interconnected network.

The problem of stability boundary evaluation for the current power system state is of paramount importance for power system operation. In addition, maintaining stability after large scale disturbances has come to the fore in recent years. This paper presents the algorithm, which allows us to both evaluate the stability boundary of a power system and to calculate emergency control actions for maintaining stability in the case of blackout. The algorithm is based on Newton’s method for solving optimization problems. There are a number of emergency actions algorithms in literature, but most of them use heuristic rules. On the contrary, the proposed method has a reasonable analytical background. Thus, having an adequate power system model, the proposed method is able to calculate more accurate control actions. The paper demonstrates the very basic idea of the approach with the simplest example. Keywords: feasible boundary evaluation, power system stability, Newton’s method, Jacobian matrix.

At present, the power industry pays much attention to the development of digital substation technologies, implemented in accordance with IEC 61850 standard [1], [2]. Based on the industrial developments and achievements of the 80's, in 2005 and 2010 the standard was subjected to significant updates, which improved its efficiency in terms of secondary substation systems design: relay protection and automation systems, data acquisition systems, technological process automated control systems [3].