Background. With the increase in observability and controllability of regimes, the development of methods for managing distributed objects of the electrical network is becoming more and more important. The main research directions in smart grids are based on the theory of fuzzy sets, genetic algorithms, neural networks, stochastic control, spectral graph, bilinear matrix inequality constraints. They are aimed at solving multicriterion optimization problems of electric networks with distributed objects and are computationally-demanding and time-consuming. Meanwhile, the methods of multi-agent control of the power supply system based on the parallelization of information flows and coordination of the operation of distributed linear regulators are becoming more common. The purpose of this study is to develop methods for controlling the operating modes of smart distribution electric networks of railways using an agent-based approach for stabilizing voltages within specified limits and reducing electric power losses. This goal can be achieved by solving the problems of developing an algorithm for managing power flows based on the coordinated work of active and reactive power sources and principles of demand management of active consumers.
Materials and methods. The multi-agent power flow control was realized in the AnyLogic program, the simulation modeling of the electrical network modes was performed in Matlab Simulink with assumptions of linear characteristics of voltage loads.
Results. A method has been developed to control the operation modes of smart distribution electric networks of railways based on the presented power flow control algorithm, the hallmarks of which are the use of linearized equations for determining control actions in small increments, which allows high speed data analysis in real time without calculating steady-state modes with disturbances.
Conclusions. The obtained simulation results prove the validity of power flow control methods for voltage stabilization based on multi-agent control and the possibility of their practical implementation on modern equipment in smart distribution networks of railways.