Translation. Region: Russian Federal
Source: Peter the Great St. Petersburg Polytechnic University –
An important disclaimer is at the bottom of this article.
Scientists from the Polytechnic University have developed an innovative hybrid algorithm to improve the operation of power equipment. The new system combines digital twin technologies with dynamic self-parameterization and AI. This allows predicting changes in the operation of complex power equipment, preventing emergency situations, and increasing the marginal income of the generating enterprise.
Reliable monitoring and forecasting of the state of complex power equipment is one of the key tasks for the Russian energy sector. This is directly related to ensuring national energy security and meets the goals Energy Strategy of the Russian Federation until 2050, which involves the implementation of digital twins and predictive analytics systems based on AI. Despite the widespread development of global research in this area, foreign solutions poorly cover the specifics of Russian thermal power plants, which are distinguished by the diversity of equipment, the complexity and variety of modes of combined production of thermal and electrical energy, etc.
Currently, Russian predictive analytics systems are based primarily on the analysis of trends in key parameter changes and use classic neural models built on statistical information from automated process control system (APCS) devices. The widespread implementation of this approach in the domestic energy sector is limited by a number of reasons. These include the low degree of automation of power equipment by APCS devices, the unreliability of some signals, and the introduction of new domestic energy equipment for which a pool of necessary statistical information on operation in various modes has not yet been collected.
The development of the SPbPU research team in the field of predictive analytics is intended to forecast degradation and defects of power equipment. At the first stage, a digital model of the station is created, data on the operation of the thermal power plant from standard devices is loaded into specialized software. Then, information from standard devices of the automated process control system is received in real time. After checking for adequacy, the model issues recommendations on the optimal management of the mode.
Using machine learning methods, our development automatically takes into account changes in the physical characteristics of key equipment units, occurring, for example, due to natural wear. The model is self-parameterized and can fill in gaps in the data obtained, for example, about those station units where it is impossible to install a monitoring sensor, and eliminate inaccuracies in existing measurements. Having received a reliable digital copy of the most complex power equipment, we can conduct an in-depth analysis of the station's operation and predict the occurrence of defects in the future, as well as study data on the complex influence of many factors on technical processes. Until now, it was impossible to obtain such information either theoretically or practically, – noted the project manager, associate professor of the Higher School of Nuclear and Thermal Energy of SPbPU Irina Anikina.
This task is especially relevant for new domestic gas turbine units, for which a large array of statistical information has not yet been collected. A pilot prototype of the system has been tested at some stations in the North-West region. Scientists believe that it will ultimately be possible to increase the marginal income of the thermal power plant by 7-8% by selecting optimal operating modes taking into account the actual state of the equipment.
In addition, new hybrid algorithms will reduce the number of unplanned repairs due to abnormal equipment behavior and optimize the repair schedule. This is important, since losses in case of accidents can vary from several million to billions depending on the capacity, cost of generating equipment and complexity of repairs, features of the sales activities of the thermal power plant, etc.
The team’s plans include further development of the system, its adaptation for other types of generating equipment and scaling to other energy industry enterprises.
The research work is carried out with the support of the SPbPU Development Program for 2025–2036 as part of the implementation of the Priority 2030 program (the national project Youth and Children).
Please note: This information is raw content obtained directly from the source of the information. It is an accurate report of what the source claims and does not necessarily reflect the position of MIL-OSI or its clients.
.
