Forecasting electricity consumption is one of the most important scientific and practical tasks in the field of electric power engineering. The forecast accuracy directly impacts the operational efficiency of the entire power system and the performance of electricity markets. This paper proposes algorithms for source data preprocessing and tuning XGBoost models to obtain the most accurate forecast profiles. The initial data included hourly electricity consumption volumes and meteorological conditions in the power system of the Republic of Tatarstan for the period from 2013 to 2025. The novelty of the study lies in defining and justifying the optimal model training period and developing a new evaluation metric for assessing model efficiency—financial losses in Balancing Energy Market operations. It was shown that the optimal depth of the training dataset is 10 years. It was also demonstrated that the use of traditional metrics (MAE, MAPE, MSE, etc.) as loss functions during training does not always yield the most effective model for market conditions. The MAPE, MAE, and financial loss values for the most accurate model, evaluated on validation data from the first 5 months of 2025, were 1.411%, 38.487 MWh, and 16,726,062 RUR, respectively. Meanwhile, the metrics for the most commercially effective model were 1.464%, 39.912 MWh, and 15,961,596 RUR, respectively.

A solid oxide fuel cell (SOFC) is an electrochemical energy conversion device that provides higher thermoelectric efficiency than traditional cogeneration systems. Current research in this field highlights a variety of mathematical models. These models are based on complex physicochemical and electrochemical reactions, enabling accurate simulation and optimal control of fuel cells. However, these models require substantial computational resources, leading to high processing times. White box and gray box models are unable to achieve real-time optimization of control parameters. A potential solution involves using data-driven machine learning (ML) black-box models. This study examines three ML models: artificial neural network (ANN), random forest (RF), and extreme gradient boosting (XGB). The training dataset consisted of experimental results from SOFC laboratory experiments, comprising 32,843 records with 47 control parameters. The study evaluated the effectiveness of input matrix dimensionality reduction using the following feature importance evaluation methods: mean decrease in impurity (MDI), permutation importance (PI), principal component analysis (PCA), and Shapley additive explanations (SHAP). The application of ML models revealed a complex nonlinear relationship between the SOFC output voltage and the control parameters of the system. The default XGB model achieved the optimal balance between accuracy (MSE = 0.9940) and training speed (τ = 0.173 s/it), with performance capabilities that enable real-time enhancement of SOFC thermoelectric characteristics during system operation.

The innovative baromembrane technologies for water demineralization were introduced at Russian TPPs more than 25 years ago. While being used in the power engineering industry of Russia, these technologies demonstrated certain advantages over the traditional ion-exchange and thermal technologies of makeup water treatment for steam boilers. Water treatment units based on the baromembrane technology are compact, easy to operate, and highly automated. The experience gained from the use of these units shows that their reliability depends directly on preliminary water treatment. The popular water pretreatment technology with coagulation by aluminum oxychloride proved to be inefficient during the seasonal changes of source water quality that occurs at some stations. The use of aluminum coagulant at pH 8 and higher does not ensure the stable and qualitative pretreatment regime: soluble aluminum forms slip on membranes of the ultrafiltration unit, thereby causing pollution and intoxication as well as leading to structural damages or worsening of mechanical properties of the membranes. The problem of increased pH and seasonal changes of the source water quality can be solved by substitution of the traditional coagulant into a new one. To find the most successful coagulant for water pretreatment, experiments have been performed on both qualitative and quantitative analysis of the content of natural organic matters in the Volga water and their structure. We have developed a software program and measured the concentrations of soluble aluminum and iron salts at different pH values of the source water. The analysis of the obtained results has indicated that iron sulfate at pH 6.0−10.2, in contrast to aluminum oxychloride, is not characterized by increased solubility. Thus, the basic process diagrams of water pretreatment based on baromembrane technologies with pretreatment through coagulation by iron salts and wastewater amount reducing from 60–40 to 5–2% have been introduced for thermal power stations.

Вторым товаром на рынках электрической энергии является электрическая мощность. Для потребителей тарифицируемый объем «генераторной» мощности определяется как среднее из часовых объемов потребления в рабочие дни в часы пиковой нагрузки в регионе. Стоимость мощности в отдельных регионах может достигать 40 % от конечного тарифа, поэтому снижение нагрузки в пиковые часы на 10 % может привести к уменьшению ежемесячных платежей на 3 %. Однако такой способ экономии для потребителя недоступен – коммерческий оператор оптового рынка электрической энергии и мощности публикует часы пиковой нагрузки регионов после 10-го числа следующего месяца, когда данная информация уже не актуальна. Своевременное прогнозирование часов пиковой нагрузки позволит, с одной стороны, снизить издержки потребителей на платежах за электрическую мощность, с другой – сгладить суточный график электрической нагрузки энергосистемы, оптимизировав тем самым работу генерирующего оборудования станций и сетей системного оператора. В статье приводится исследование эффективности методов машинного обучения в контексте прогнозирования пикового часа региональной энергосистемы. Исследование затрагивает временной период с ноября 2011-го по октябрь 2023 г., охватывает 76 регионов Российской Федерации, включая субъекты ценовых (1-й и 2-й) и неценовых зон и насчитывает 10 методов машинного обучения. Результаты исследования показали, что статистически метод кластеризации K-ближайших соседей оказывается наиболее точным, хоть и не универсальным. Высокую эффективность (с точки зрения точности и быстродействия) продемонстрировали методы опорных векторов и классификация деревьями. В ходе исследований также было опровергнуто предположение о том, что наибольшую ценность при прогнозировании пикового часа оказывают наиболее близкие, с позиции временного ряда, данные.

The introduction of baromembrane water treatment technologies for water desalination at Russian thermal power plants was beganed more than 25 years ago. These technologies have demonstrated their definite advantage over the traditional technologies of additional water treatment for steam boilers. However, there are problems associated with the reliability and economy of their work. The first problem is a large volume of waste water (up to 60% of the initial water). The second problem a expensive and unique chemical reagents complex (biocides, antiscalants, washing compositions) is required for units stable and troublefree operation. Each manufacturer develops his own chemical composition for a certain membrane type. This leads to a significant increase in reagents cost, as well as creates dependence of the technology consumer on the certain supplier. The third problem is that the reliability of the baromembrane units depends directly on the water preliminary treatment. The popular pre-cleaning technology with coagulation of aluminum oxychloride proves to be unacceptable during seasonal changes in the quality of the source water at a number of stations. As a result, pollution, poisoning and lesion of the membrane structure or deterioration of their mechanical properties are observed. The report presents ways to solve these problems.

The correct prediction of heating network pipeline failure rates can increase the reliability of the heat supply to consumers in the cold season. However, due to the large number of factors affecting the corrosion of underground steel pipelines, it is difficult to achieve high prediction accuracy. The purpose of this study is to identify connections between the failure rate of heating network pipelines and factors not taken into account in traditional methods, such as residual pipeline wall thickness, soil corrosion activity, previous incidents on the pipeline section, flooding (traces of flooding) of the channel, and intersections with communications. To achieve this goal, the following machine learning algorithms were used: random forest, gradient boosting, support vector machines, and artificial neural networks (multilayer perceptron). The data were collected on incidents related to the breakdown of heating network pipelines in the cities of Kazan and Ulyanovsk. Based on these data, four intelligent models have been developed. The accuracy of the models was compared. The best result was obtained for the gradient boosting regression tree, as follows: MSE = 0.00719, MAE = 0.0682, and MAPE = 0.06069. The feature «Previous incidents on the pipeline section» was excluded from the training set as the least significant.