The Rosenergoatom Joint-Stock Company (a part of the Electric power division of the Rosatom State Corporation) has recently released the interim results of the pilot project aimed at establishing a predictive analytics system for the NPP generating equipment using the 6th power unit of the Novovoronezh NPP as an example.

The pilot implementation is done under the auspices of the ‘Single technical policy – guaranteed electric supply’ of the Ministry of Energy. The project is scheduled for years 2019 – 2021. In addition to Rosenergoatom and a number of external suppliers, its affiliates and regional offices are involved, too, including the Novovoronezh NPP, AO VNIIAES and AO Consist-OS.

The goal of the project is to set up a prototype of a predictive analytics system that would timely identify any latent defects in the equipment. It requires both a practical way to process the mathematical methods of predictive data analysis for equipment operation and a way to gather and transfer the data in the real time from the equipment control and diagnostics systems (which is a challenge for all energy companies), as well as to involve seasoned specialists to review the predictive analytics results.

The initial stage of the project was to conduct an open technical tender, the participants of which had to analyze the information of NPP power units’ turbine generators operation. 19 companies took part in the competition: IT companies of different scales, R&D facilities and institutes, organizations specializing in equipment monitoring. As a result, the participants presented a technical review of the information analysis results using predictive analytics along with a description of their solution. The submissions were evaluated based on several criteria. Following that, a test range for the predictive analytics system was set up and launched using the capabilities of AO VNIIAES and the Rosenergoatom Joint-Stock Company’s data processing center.

‘Following the first stage of the project, we have managed to develop and verify the algorithms and mathematical models that help identify emerging and developing latent defects in the NPP equipment. For some defects that can happen primarily in the major power unit equipment, we have managed to detect the most important parameters. The changes in all of those parameters, not in separate ones, help to identify any deviations in equipment operation. Some of these deviations can remain within the normal mode, which, however, does not guarantee that the equipment has no latent defects. In that case, we need to understand possible defect causes, how quickly they evolve and how much time will it take before the equipment parameters stop being normal. The introduction of such a system will help us reduce fines for unscheduled, urgent and emergency stops of NPP power units, which, obviously, can sometimes be unavoidable’, Liubov Andreeva, the director of the Digital energy and commercial dispatching department at the Rosenergoatom Joint-Stock Company, said.