+7 (495) 987 43 74 ext. 3304
Join us -              
Рус   |   Eng


Kirillova Elena A.

Cand. Sci. (Econ.), Associate Professor, Information Technologies in Economics and Management Department, Branch of the National Research University "MPEI" in Smolensk
Smolensk, Russia

Intelligent control algorithm for autonomous integrated power plants for Arctic regions

The article considers the information and technical aspects of an innovative autonomous integrated power plant management, including alternative energy sources and a diesel generator set, as well as controlled inverters that provide energy supply to consumers of various priority categories, which can be used in the Arctic regions of the Russian Federation. The main aspects of creating innovative systems are considered and it is determined that the creation of integrated energy systems requires a significant deepening of cooperation between national producers in order to ensure the scalability of integrated energy systems by ensuring the unity of information means of data exchange between individual modules and the control system. It is shown that a specific requirement for control systems of complex power plants is the requirement of high autonomy, including the ability to provide consumers with electricity under variable environmental conditions without direct intervention of operational personnel. The article substantiates the division of the information and algorithmic support of the control system of a complex power plant into two modules – analytical and control. For the analytical module, an algorithm is proposed that ensures the development of control solutions in a complex energy system, ensuring the stability of energy supply to the most important consumers. At the same time, the algorithm provides an increase in the reliability of the energy storage device based on Li-Ion batteries used in the system not only by eliminating excessive charge and deep discharge, but also by reducing the number of charge/discharge cycles. The solution of system autonomy problem is provided by a multivariate algorithm for predicting weather conditions using statistical data and methods for analyzing fuzzy time series. The intelligent control algorithm was implemented in C++, the weather forecasting algorithms were implemented in Python using the ANFIS library. Read more...

Neural network model to support decision-making on managing cooperative relations in innovative ecosystems

Currently, the specifics of external conditions and peculiarities of innovation activity main subjects development determine not only the need for close, long-term scientific and technical cooperation with the state for the sustainable development of territories, but also the need to develop and substantiate proposals for managing the development of innovation processes in such a system as a whole. The article proposes a model for the representation of scientific and industrial interaction in the implementation of regional innovation processes in the form of a three-dimensional "slice" of the triple helix as a resource VRIO-profile of cooperative formation, which allows to clearly demonstrate the system of relations, identify in which direction the problem area is, influencing which it will be possible to return the system to an equilibrium state of sustainable development in a strategic perspective. The analysis of modern scientific works shows the relevance, necessity and effectiveness of using methods based on neural networks to predict changes in the state of complex socio-economic systems, such as regional innovation systems. Existing approaches, as a rule, demonstrate a narrow focus and belonging to a separate enterprise or organization, and therefore do not meet all the requirements from both the implementation of the innovation process itself and the modification of the external environment. In this connection, the authors proposed an information and analytical solution for using the described model to support decision-making on the management of cooperative formations. The developed program is based on predicting the future state (position in a three-dimensional coordinate system) of the system using deep neural networks, namely recurrent. The described practical approbation of the model can in the future serve as a basis for decision-making on the choice of forms and directions of interaction of cooperative formations in the strategic perspective. Read more...