The article is devoted to increasing energy saving and energy efficiency of public sector institutions by automation of the process of energy-saving measures fitting. A classifier based on an artificial neural network is proposed as a technology for selecting measures. The base numeric dataset was supplemented by categorical data on buildings and structures. The urgency of the work is justified by the need to develop solutions aimed at introducing measures to ensure energy saving and energy efficiency in the public sector. The structure and operating principle of a module of auto-selecting energy saving measures as a part of the Energy Resources Management System (SUER) are described. The work studies the quality of fitting energy-saving measures with or without using data on categorical features. An analysis of the selected categorical features, as well as a comparative analysis of their encoding methods, is carried out. An accuracy evaluation method of the classifier in the context of the work is proposed. A series of experiments were conducted by enumerating combinations of numeric and categorical data converting methods, in order to determine the significance of categorical features in general, as well as to determine the most effective combination of encoding methods. A comparative analysis of the results of the experiments was performed and the most successful model was determined, with an additionnal assessment of the quality of the model was made based on the metrics of precision, recall and F1-score. Conclusions were made on the advisability of supplementing the original dataset with categorical data to improve the performance of the system.

In digital relay protection devices, fundamental harmonic signals are used as useful information. To isolate them from complex input currents and voltages of the protected object, non-recursive digital Fourier filters are used. The signals obtained as a result of filtering are the basis for determining the information parameters of the quantities controlled by digital organs. At the nominal frequency in the power system, the information parameters are calculated without additional error. In the event of a frequency deviation from the nominal value due to a number
of reasons, an additional error component appears in the information parameters of the controlled values. When there are minor frequency fluctuations, the specified component is insignificant and can be neglected in practice. In case of significant frequency deviations in the power system, the additional component of error in the information parameters can negatively affect the operation of the functional algorithms of the relay protection. One of the ways to solve this problem is to correct the specified component when the frequency deviates from the nominal one. To implement this, it is necessary to have the results of frequency assessment. The most rational way to correct the frequency component of the error is to use the current frequency. Determining the specified frequency of alternating current consists of measuring the instantaneous frequency at different points in time and averaging these values over a certain interval. The instantaneous frequency is calculated from the instantaneous value of the dynamic cosine of the sampling angle, for the determination of which two functional algorithms have been developed and studied. The first of these is based on the use of three consecutive values of one of a pair of orthogonal signals. The second algorithm is implemented using two adjacent values of each of the mutually orthogonal signals. The results of the conducted studies showed that the developed algorithms for determining the current frequency in digital relay protection devices ensure its receipt with acceptable reliability in the range of 47–51 Hz at a change rate of up to 4 Hz/s. 

The ionization of negative ions in columnar voids is determined by temperature, as well as by the externally applied electric field strength. Due to the ionization of negative ions in the conduction band of the semiconductor base, free electrons are formed without their movement, forming a conduction current. These free electrons together create an internal electric field, which changes due to the processes of ionization and recombination. The changing electric field oscillates in the infrared region of the spectrum, which causes the displacement current to arise. The magnitude of the bias current is determined by the applied voltage between the emitter and collector, but in a rather complex way. The optimal applied voltage is ~6 V. At this voltage, the calcula-ted electric current in the emitter reaches milliamperes, and in the base – microamperes, which corresponds to experimental data. Continuous oscillation of the electric field due to ionization and recombination of negative ions in the columnar void leads to heating of the transistor by infrared radiation. Optimal operating conditions for the transistor are achieved when an additional voltage of 2.5–3.0 V is applied to the base. To reduce the heating temperature of the transistor, a massive collector is created, which is grounded.

The second part of the article presents a study and modeling of the electrophysical characteristics of parts with various purposes featuring a specific laser-induced configuration of microand nanostructural surface features on the product. The relevance of this study is due to the fact that during the operation of any power plant, the values of its electrophysical cha-racteristics are an integral part of it. The conducted research enables control over these chara-cteristics both in the input platform for power supply from corresponding generators to the installation and in the output energy used to ensure the operation of respective units, whose power supply is provided by energy installations of different classes during their operation in required modes, including extreme dynamic operating modes. This study specifically addresses demonstration circuits with prototypes of systems using microscrolls of 1D titanium dioxide structures in metal-carbon compounds (carbon – gold) under C–Au chain doping conditions. In this case, a three-stage process was implemented using laser ablation from a titanium target, synthesizing a thin porous titanium dioxide film and depositing it onto a demonstration quartz glass substrate. Subsequently, linear carbon chains, stabilized by gold nanoparticles at their edges, were introduced into the porous titanium dioxide film matrix by jet spraying, forming an array of microscrolls through mechanical action. Mathematical and computer modeling of topological microand nanostructures on the surfaces of metal complexes with laser-controlled configurations was performed. The conducted analysis allows us to draw a conclusion, based on the proposed procedures and processes regulating the topological structure of the considered surface objects during their laser synthesis, about the prospects of this direction, which is associated with the possibility of controlling the functional surface characteristics in the required direction, particularly regarding their electrophysical para-meters for various products used in energy devices.

A modern Nuclear Power Plant (NPP) is equipped with localizing safety systems to contain radioactive substances and ionizing radiation within the boundaries specified by the design in the event of an accident. To protect the border, a hydrogen removal system is used, including passive autocatalytic hydrogen recombiners. The system prevents the formation of flammable and explosive concentrations of hydrogen by converting hydrogen into water during a reaction with atmospheric oxygen on a catalyst. The main catalyst material is usually platinum with a proportion of palladium. Along with hydrogen, the emergency environment contains specific substances known as catalyst poisons. Poisons reduce catalyst activity and decrease the performance of recombiners. Electron pair donor substances, such as tellurium, are dangerous for the platinum catalyst. The amount of released catalytic poisons at the stage of active zone melting is sufficient to reduce the activity of the catalyst. The level of reduction is estimated by calculations. Aerosol poisons are dangerous in the zone of separated flows near the surface of the catalytic element. Poisons in atomic (molecular) form are dangerous to the catalyst along the entire length of the element. The poison causes a gradual decrease in the productivity of recombiners as the amount of reacted hydrogen increases. The rate of poisoning depends on the type and concentration of the poison, and the mass of active platinum per unit surface area of the catalyst. The calculation shows the possibility of a decrease in the productivity of a passive catalytic hydrogen recombiner due to catalyst poisoning under accident conditions at a Nuclear Power Plant with Water-Moderated Water-Cooled Reactors. Quantitative data on the poisoning of recombiners of the FRand RVK-types are presented. The poisoning effect should be taken into account when selecting the performance of the hydrogen removal equipment of the localizing safety system for a power unit at a Nuclear Power Plant with Water-Moderated Water-Cooled Reactors.

Heat demand forecasting is necessary to achieve optimal management of building energy consumption. The purpose of this article is to identify the most important factors influencing the accuracy of forecasting heat consumption of buildings using neural networks, which is in line with the national strategy for the development of artificial intelligence of the Russian Federation. The article studies the dependence of modeling accuracy on various combinations of environmental parameters, as well as on the application of different activation functions of neural networks, widely used in the practice of creating artificial intelligence systems. It is demonstrated that machine learning models based on a large number of data on thermal consumption have great possibilities in predicting real patterns and trends of consumption, and the value of the average absolute percentage error of the best prediction model is comparable to the value of the maximum limit of the tolerable relative error of thermal energy measurements by the measuring channel of the heat meter. On the basis of data obtained using the developed system of remote monitoring of individual heating points of buildings, a comparison of actual values of heat consumption and values of heat consumption obtained using the prediction model was demonstrated. Savings of energy, heat carrier and other things at the object cannot be measured directly, because the savings represent the absence of consumption, so a universal approach using artificial intelligence for a technically sound and economically feasible method of predicting the results of the application of energy-saving solutions to compare the measured energy consumption before and after the implementation of energy-efficient measures may allow to improve the efficiency of decision-making in the field of saving energy resources.