A Risk Estimation Methodology Based on Machine Learning Algorithms in Underground Metro Structures

  • Authors

    • Muhammad Fayaz
    • Do-Hyeun Kim
    2018-12-13
    https://doi.org/10.14419/ijet.v7i4.39.24373
  • Neural Network, Metro Structure, Risk Assessment, Underground Risk
  • Background/Objectives: Underground risk index assessment is a very challenging task due to the unavailability of underground features information. A lot of factors normally contribute in underground failures. Underground failures occur in a random manner, but a proficient underground risk assessment method can avoid underground failures. Metro risk is a serious threat to underground structures.Methods/Statistical analysis: In this paper, we have proposed a risk index assessment methodology for underground metro structure. The proposed methodology consisted of three stages, namely the data layer, the risk index estimation layer, and performance evaluation layer. Two parameters, namely year of burial, and degree of depression have been used in the proposed work. These parameters are then further used as inputs to risk index estimation layer. The feed-forward neural network (FFNN) and classification and regression tree (CART) have been used in the risk index estimation layer for metro structure risk index estimation. The output of the neural network is further evaluated in performance evaluation layer, where root means square error (RMSE), mean absoluter error (MAE) and mean absolute percentage error (MAPE) have been used.Findings: It is very difficult to develop a methodology to asses underground risk index taken into all parameters. Underground risk index analysis is very complicated due to its complex nature. The only one way is to assess one by one. The proposed method estimates the risk index of metro structure risk index. The proposed method can be efficiently used to estimate risk index of underground structure.Improvements/Applications: We have used machine learning algorithms such as FFNN and CART for metro structure risk index estimation which is a novel idea. The results indicate that the performance of CART is better as compared to FFNN. 

  • References

    1. [1] Sturk R, Olsson L, Johansson J. Risk and decision analysis for large underground projects, as applied to the Stockholm ring road tunnels. Tunnelling and Underground Space Technology. 1996 Apr 1;11(2):157-64.

      [2] Fayaz M, Ullah I, Park DH, Kim K, Kim D. An integrated risk index model based on hierarchical fuzzy logic for underground risk assessment. Applied Sciences. 2017 Oct 11;7(10):1037.

      [3] Paolucci R, Pitilakis K. Seismic risk assessment of underground structures under transient ground deformations. InEarthquake Geotechnical Engineering 2007 (pp. 433-459). Springer, Dordrecht.

      [4] Fayaz M, DH, Kim D. An Estimation Method of Geo-environmental Risk Index Based on Feed Forward Neural Network for Surveilling Underground Facilities. International Journal of Advanced Science and Technology, 113,11-20,2018

      [5] Fayaz M, Ahmad S, Ullah I, Kim D. A Blended Risk Index Modeling and Visualization Based on Hierarchical Fuzzy Logic for Water Supply Pipelines Assessment and Management. Processes. 2018 May 22;6(5):61.

      [6] Wang YM, Elhag TM. An adaptive neuro-fuzzy inference system for bridge risk assessment. Expert Systems with Applications. 2008 May 1;34(4):3099-106.

      [7] Fayaz M, Ullah I, Kim DH. Underground Risk Index Assessment and Prediction Using a Simplified Hierarchical Fuzzy Logic Model and Kalman Filter. Processes. 2018 Jul 29;6(8):103.

      [8] Wahid F, Ghazali R, Fayaz M, Shah AS. Statistical features based approach (SFBA) for hourly energy consumption prediction using neural network. Networks. 2017;8:9.

      [9] Ullah I, Ahmad R, Kim D. A Prediction Mechanism of Energy Consumption in Residential Buildings Using Hidden Markov Model. Energies. 2018 Feb 3;11(2):358.

      [10] Abhaya Kumar Samal, Pradeep Kumar Mallick, Jitendra Pramanik and Subhendu Kumar Pani, “Functional Link Artificial Neural Network (FLANN) based Design of a Conditional Branch Predictorâ€, Cognitive Informatics and Soft Computing, (Part of AISC Springer), Vol: 768, ISBN: 978-981-13-0617-4, pp 131-152, 2018. (https://link.springer.com/chapter/10.1007/978-981-13-0617-4_14).

  • Downloads

  • How to Cite

    Fayaz, M., & Kim, D.-H. (2018). A Risk Estimation Methodology Based on Machine Learning Algorithms in Underground Metro Structures. International Journal of Engineering & Technology, 7(4.39), 545-549. https://doi.org/10.14419/ijet.v7i4.39.24373