St Petersburg University students optimise the work of industrial companies using artificial intelligence
St Petersburg University students are developing algorithms based on artificial intelligence for real power production facilities. With the help of software developed by students of St Petersburg University, the companies can optimise production processes, forecast failures in the work of the equipment and reduce costs.
Students work on the projects in the sphere of artificial intelligence within the framework of open international competitions organised by large-scale power production companies on an annual basis. A team of students from various fields of study is gathered for every competition. Supervised by Ovanes Petrosian, Associate Professor in the Department of Mathematical Modelling of Energetic Systems at St Petersburg University, the students solve the problems set by the production companies.
In the course of competitions, students mostly develop projects on optimising the processes of computer control, prediction and automatic detection of anomalies or failures in various energy systems. Within the framework of the projects, we develop software for the companies. This is a well-designed and packed algorithm that enables the equipment at the facility to take most efficient decisions without human participation.
Ovanes Petrosian, Associate Professor in the Department of Mathematical Modelling of Energetic Systems at St Petersburg University
Ovanes Petrosian explained that modern large-scale electric power companies usually provide for automatic control of equipment. In other words, the performance of all the electrical grid is supervised and controlled by special technical systems, for example, controllers, as opposed to people. They protect electric equipment from overloads, unscheduled outage and other failures leading to the breakdown of technical equipment at the facility.
However, according to Ovanes Petrosian, these systems are still incapable of optimising their actions independently. For example, they cannot take decisions reducing production costs or increasing energy production. Modern algorithms of artificial intelligence developed by the students of St Petersburg University help to overcome it.
More information about the results of work performed by the students of St Petersburg University can be found in the articles published in the following journals: Communications in Computer and Information Science, Mathematics and Control Processes and Sustainability.
Thus, for example, the students developed software to improve the quality of technical maintenance for electric grids for Réseau de Transport d’Électricité (RTE), a French energy company. ‘We had to develop an algorithm enabling the companies to calculate an optimal schedule of inspecting electric power network. This schedule had to include a limited number of support personnel and minimise possible risks: failures, defects and relevant maintenance costs,’ noted Ovanes Petrosian. ‘To solve this problem, our students used modern algorithms from the field of global optimisation, which enabled them to cope with the proposed case successfully.’
In the course of another competition, St Petersburg University students developed algorithms to control an electric system made of electric batteries and solar panels connected to them for Schneider Electric, a French power machine producer. According to Ovanes Petrosian, the students had to develop software that would minimise power production costs for the company.
As a result, the students developed a special algorithm for the controller of the battery charge and discharge at the production facility. The resulting software can teach the control system to efficiently decrease power production costs with the help of optimisation and prediction.
Ovanes Petrosian, Associate Professor in the Department of Mathematical Modelling of Energetic Systems at St Petersburg University
‘In other words, our algorithm enables the controller to independently determine its actions in any given moment: charge the battery or, on the contrary, discharge it depending on the current charge level predicted demand and electricity output with the help of solar panels,’ added Ovanes Petrosian.
The University students also developed software for automatic search of failures in electric power grids for Schneider Electric. The software enables the controllers to analyse the data on electric load from the electric grid sensors and determine the location of the failure. It significantly improves the detection of production issues and their further solution at the facility.
From the end of 2020 till the beginning of 2022, St Petersburg University students developed six projects in the field of artificial intelligence for industrial companies. Currently, the developments of the University students are available to all business companies. The developed software has been uploaded on the Internet and anyone can use it in their work at facilities.
Researchers at St Petersburg University also work on the projects in the field of artificial intelligence on the basis of the Centre for Artificial Intelligence and Big Data Science. It was created at the University in May 2020 to train specialists, organise thematic research and hold expert and analytical work. In 2022, at the premises of the centre, St Petersburg University researchers are implementing projects in 16 fields including machine learning, neural networks, distributed ledgers and blockchain.
According to Ovanes Petrosian, the work on the projects in the field of artificial intelligence is beneficial not only for businesses, but also for students. ‘During competitions, students write research articles based on their projects to be further published in major indexed research journals. Students also make presentations about the developed algorithms in domestic and international conferences significantly improving their research skills. Most importantly, their portfolio has ready-made projects with links to open-source software and video presentations,’ emphasised Ovanes Petrosian.
Ovanes Petrosian added that currently the specialists in the field of artificial intelligence are especially much-in-demand in large companies in the field of power supply, oil production and IT. These production facilities already experience the demand for specialists skilled in working with machine learning algorithms. The researcher believeds that the demand for such experts will grow in the future when small and medium size companies have the resources to integrate artificial intelligence in their work.
St Petersburg University gives a possibility to receive education in the field of artificial intelligence. The following programmes are currently available: the bachelor’s programme in Applied Mathematics, Computer Programming and Artificial Intelligence; and master’s programmes Artificial Intelligence and Data Science, Mathematical Modelling, Programming and Artificial Intelligence, Artificial Intelligence and International Security, and Engineering Oriented Computer Science and Artificial Intelligence. Starting this year, the following programmes are also open for applicants: Mathematical Robotics and Artificial Intelligence, Artificial Intelligence in Speech Technologies and Artificial Intelligence and Big Data Technologies.