DEVELOPMENT OF A SPECIALIZED SOFTWARE PACKAGE FOR NETWORK FORECAST OF FLOOD WATERS

• Published in: Bulletin of the Tomsk Polytechnic University Geo Assets Engineering
• Main Authors: Vladimir I. Syryamkin, Boris P. Ivanenko, Semen A. Klestov, Maria D. Khilchuk
• Format: Article in Journal/Newspaper
• Language: Russian
• Published: Tomsk Polytechnic University 2023
• Subjects: neural network modeling; forecast; flood waters; numerical modeling; accuracy characteristics; methods of constructing training samples; Engineering geology. Rock mechanics. Soil mechanics. Underground construction; TA703-712; Rus’; ob river
• Online Access: https://doi.org/10.18799/24131830/2023/5/3859; https://doaj.org/article/04904ff129b149d19ef0cbdac8329cda

Link for citation: Syryamkin V.I., Ivanenko B.P., Klestov S.A., Khilchuk M.D. Development of a specialized software package for neural network forecast of flood waters. Bulletin of the Tomsk Polytechnic University. Geo Аssets Engineering, 2023, vol. 334, no. 5, рр.205-216. In Rus. The relevance of the study is caused by the need to develop modern methods of operational monitoring of the condition of territories due to the presence of various kinds of natural phenomena, in particular, floods and flood inundations. A lot of attention is paid to solving this problem by States, regions, municipalities and settlements. Purpose: development of a specialized software package designed to solve problems of short-term and medium-term flood water level forecasting based on operational data from hydrological observations with a minimum set of input data and with the ability to work with aerospace observation data. Objects: the area where the Tom and Ob rivers merge and downstream of the Ob river at the locations of hydrological posts: villages Pobeda, Nikolskoye, Molchanovo. Methods: neural network information modeling. Results. The paper considers the method of creating a neural network simulator designed for processing the results of hydrological measurements and solving a wide range of practical problems, including prognostic ones. An original method of constructing training samples was developed, which allows obtaining results with a minimum set of initial data. The authors investigated the efficiency and accuracy characteristics of neural network algorithms in solving the problem of forecasting the flood water level in the period from April 1 to June 30, 2011–2017.