Vol 33, No 3 (2025)

The global energy transition and the drive to decarbonize the economy have set the scientific community the task of developing and implementing effective hydrogen production technologies.

Methane pyrolysis is one of the promising technologies for producing low-carbon hydrogen, but the diversity of its technological implementations makes it difficult to conduct an objective comparative analysis and select the most effective solutions for industrial implementation.

Each implementation option is characterized by its own set of key features and indicators, differing in physicochemical mechanisms, reactor and plant design, process conditions, and the carbon product obtained. This diversity precludes simplified evaluation and necessitates a systems approach that captures interdependencies among process subsystems and factors.

The developed systems analysis methodology is based on a structural-functional approach with hierarchical decomposition and an iterative feedback loop. It allows us to consider the pyrolysis process as a holistic system with multiple subsystems, taking into account its connection with the external environment and the possibility of forming input and output parameters for further use in Data Envelopment Analysis (DEA). The methodology for the systematic analysis of methane pyrolysis technology developed on the basis of this approach includes 12 stages grouped into three interrelated blocks: 1) description and analysis of subsystems, 2) assessment of technical, economic, and environmental characteristics, 3) data integration and a multi-factor DEA-based efficiency assessment, followed by managerial decision support.

The developed methodology provides a comprehensive treatment of the subject, explicitly integrates statistical analysis and DEA, accounts for catalyst deactivation dynamics, incorporates an iterative analysis loop, and is oriented toward scale-up and the identification of optimization opportunities.

In the tasks of synthesizing automatic control systems (ACS) for objects with parametric uncertainty, one of the problems is the need to take into account possible deviations of the values of uncertain parameters from their nominal values. The problem of choosing a method for solving the problem of synthesizing the control law is complicated not only by their diversity, but also by the lack of awareness of the design engineer directly in the theory of control. The use of automatic solutions for non-procedural tasks, i.e. without specifying the method of their solution, it is not used not only because of the need to create and use a knowledge model of the problem domain of the theory of automatic control. But also because of the difficulties of planning actions to solve them using iterative methods. Because the problem of automatic synthesis of cyclic programs does not have a universal approach to planning actions to solve declarative set tasks. However, for a specific class of tasks, a certain generalized structure of an action plan for solving problems of this class can be defined. The growing need to build new self-propelled guns for new control facilities requires solving this problem. One of the directions of building control systems for objects with uncertain parameters is the methods of building robust control systems that involve iterative processes of performing a number of sub-tasks. This article is devoted to the study of ways to solve declarative set tasks of robust management. An approach to adjusting the knowledge model is proposed that allows the planning subsystem, based on planning artificial neural networks, to generate action plans for declarative set tasks of robust stabilization, allowing their cyclical execution until acceptable indicators of the requirements for the desired result of solving the problem are achieved. We were developing the scheme for the presentation of cyclic action plans that implement iterative methods for synthesizing control laws. An example of solving a declaratively set task of robust stabilization of a two-mass system is considered.

The paper discusses a methodology for analyzing, predicting the condition, and identifying latent defects in steam turbines using parametric diagnostics based on data measured during the operation of a turbine at a thermal power plant. The article presents a diagnostic analysis of a T-110/120-130 turbine. The assessment and prediction of the T‑110/120-130 turbine's condition are based on the results of analyzing measured parameters and calculating a technical condition index using artificial intelligence methods and neural network algorithms. The main emphasis is placed on the principles of building parametric diagnostic systems for auxiliary work as assistants at power stations. The proposed methodology allows for the development of a diagnostic system for the T-110/120-130 turbine with the capability to detect and identify latent defects and predict the turbine's condition over specific time intervals.

Market and technological possibilities of solving the problem of providing EV with electric filling service by installing electric charging stations in already built and operated apartment buildings and public buildings are discussed. The method of determining the available charging capacity is analyzed depending on contractual restrictions on the retail electricity (capacity) market and circuit conditions at the point of connection of the station to the power supply system. Strategies and principles of technological control of electric vehicle charging in conditions of power shortage have been developed. It is concluded that it is advisable to make a forecast of the available charging capacity using one of the algorithms based on the theory of artificial neural networks, with training information in the form of retrospective time series of the load of the power supply system, external weather conditions and the presence (absence) of centralized heating and hot water supply. Heuristic forecasting methods also have good prospects.