Research of the Underground Water Level Prediction Model in Deep Foundation Pit Engineering

Abstract:

The effect such as ion exchange, precipitation, corrosion and consolidation can occur between groundwater and rock mass, it will cause a variety of adverse effects on deep foundation pit engineering. Prediction of the underground water level and take corresponding precipitation control measures is very important. Underground water level deformation is a complicated ,nonlinear and stochastic problem, it is unable to establish accurate mathematical model. An underground water level deformation prediction model based on BP neural network was constructed in this paper. Five closely related factors in underground water level deformation are river flow, temperature, saturation deficit, rainfall and evaporation, they were selected as input vector of BP neural network, underground water level measured value as a model target output. In Matlab 2011b simulation software, 24 groups observation data for underground water level and five closely related factors of a underground parking lot deep foundation pit engineering in Jilin as the sample set,19 groups were randomly selected as the training sample set , other 5 groups were used as the testing sample set .The simulation result shows that testing value is very close to the true value in this method and the average relative error was 2.9708%.The method in this paper can achieve higher accuracy of groundwater level prediction in deep foundation pit engineering.

• Periodical: Applied Mechanics and Materials (Volumes 675-677)
• Pages: 901-904
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.675-677.901
• Online since: October 2014
• Authors: Hao Peng Li*
• Keywords: BP Neural Network (BPNN), Deep Foundation Pit, Prediction, Underground Water Level

References

[1] Yu Ying-wu, Li Zhan-hua and Wei Mao-jie: submitted to Journal of Ground water (2007).

[2] Zhang Yu-xin and Liu Yu: submitted to Journal of Applied Mechanics and Materials ( 2013).

[3] Liu Yu, Lu yu and Zhang Yu-xin: submitted to Journal of Metallurgical Industry Automation(2012).

[4] Chen Ming, in: MATLAB neural network principle and essence of instance, edtied by Qinghua University Publising, Beijing(2013), in press, , in Chinese.

[5] Wang Xiao-chuan, Shi Feng, Yu Lei, etl. in: MATLAB neural network 43 case analysis , edtied by Beihang University Publising, Beijing(2013), in press, in Chinese.