Novosibirsk State Technical University (NETI) is developing a program for calculating and designing electrical networks based on the use of evolutionary algorithms and artificial intelligence methods.
According to Nikita Sergeev, an assistant at the Department of Electric Power Plants at NSTU-NETI, when designing new electric networks, they used to follow existing standards and had a well-established design approach. Now there is a tendency to switch to distributed energy, and the optimal design of power grids is becoming more difficult. When designing electrical networks, it is increasingly necessary to take into account the presence of several low-power energy sources, and to sort out the options for arranging electrical equipment. There can be a large number of such options, and this is labor-intensive.
"I suggested using genetic and swarm algorithms, which are essentially mathematical methods based on the principles of natural processes such as evolution and the collective behavior of living organisms. Swarm intelligence focuses on the behavior of packs (birds, wolves). Suppose, birds move in a flock, they try to find the optimal flight path. In the conditions of choosing a balance between the optimal trajectory for each individual and for the pack as a whole, the direction of movement is constantly adjusted. This can be expressed mathematically and used to find optimal solutions to various engineering problems. For example, you can select the optimal one from a variety of electrical network options, evaluating each option in terms of reliability and economic efficiency. Artificial intelligence methods are used to find such optimal options," Nikita Sergeev said.
According to him, when designing an electric grid, it is necessary to place switches between the source of electricity and consumers. This equipment allows you to automatically disconnect the damaged network element in case of an accident and quickly put it into operation after troubleshooting. Since switches are expensive equipment, it is important to determine their optimal number and location in order to both reduce the cost of the network and ensure a sufficiently reliable power supply. The program goes through and evaluates various options for the arrangement of switches and offers several solutions at the output.
"The results of the study can be used in the design of autonomous energy complexes in remote areas of Russia, such as the Far East and the Far North," Nikita adds.
Currently, a certificate has been received for the registration of the computer program "Program for finding the optimal location and number of points for partitioning the distribution electrical network using evolutionary and swarm algorithms." Based on the results of this study, a hardware and software complex can be created for use by specialists when comparing several options for the configuration of an electrical network.