Simulation modelling of the filling oil products process in the arctic zone using neural network forecasting

Abstract

The intensification of the extraction of natural resources in the Arctic leads to an increase in the anthropogenic load on the environment. Due to wear and tear of equipment, corrosion of metals, mechanical damage, due to lack of proper maintenance and prolongation of the operating time in difficult climatic conditions, the operation of equipment used at enterprises leads to an increase in the risk of accidents. On the territory of the Russian Arctic, there are a number of large potentially dangerous objects, which have over 60 tanks for storing petroleum products. In 2020, there were technological accidents related to the bottling of petroleum products, which actualized the problem of taking prompt measures to prevent such emergencies. The methods for assessing the area of bottling of petroleum products currently used, especially in the Arctic zone, have a number of limitations. In the work with the use of the software product Toxy + risk, the risk of an emergency at one of the enterprises of the Russian Arctic was calculated. Neural network prediction of the area of pollution of the surface of the Kheta River was carried out using the NeuroPro neural network simulator. Comparison of the simulation results with the data obtained earlier in the analysis of the accident in the tank farm located beyond the Arctic Circle in the city of Norilsk.