Loading...

Probabilistic Assessment of Flood Risk Using Data-Driven Flood Depth Modeling: A Case Study of Poldokhtar City

Ziya Shamami, Oveys | 2022

208 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 54969 (09)
  4. University: Sharif University of Technology
  5. Department: Civil Engineering
  6. Advisor(s): Safaie Nematollahi, Ammar
  7. Abstract:
  8. The present study aims to evaluate the flood risk of Poldokhtar city probabilistically using Monte Carlo Simulations (MCS). Two-dimensional (2D) models, which are highly accurate, have been used widely for flood modeling. However, they are not suitable for applications such as MCSs that need to be repeated many times or real-time flood forecasting applications, which require that flood inundation maps quickly be produced. In the current study, we developed a data-driven surrogate model based on the Least Squares Support Vector Machine (LS-SVM), a supervised machine learning method, to predict flood depth in order to simulate similar results to 2D hydraulic modeling. HEC-RAS was used for 2D hydraulic flood modeling, and its results were utilized to train and test the data-driven model. The results of data-driven modeling demonstrated that the model could predict flood depth accurately in a much shorter time (about 60 times faster). The model has a Root Mean Squared Error (RMSE) of 0.0063 m and a Mean Absolute Error (MAE) of 0.0145 m on the test dataset. These results indicated that the model is highly accurate and could be used in flood risk assessment. In the next step, we used this model to analyze the flood risk of Poldokhtar city for 10000 different flood scenarios. The mean annual flood damage of the city was computed as 1177034 US$, and the mean annual flood damage for the buildings which experienced at least one flood inundation in these 10000 scenarios was calculated as 352 US$. Furthermore, the results showed that the zones in the Poldokhtar city, which are in the vicinity of the Kashkan river, are at high flood risk. In conclusion, based on the results, it is recommended that to reduce flood risks in these zones, the eastern wall should be extended to the river bend, and the western wall should be extended to the end of the city
  9. Keywords:
  10. Risk Analysis ; Monte Carlo Simulation ; Machine Learning ; Least Squares Support Vector Machine (LS-SVM) ; Two-Dimensional Hydraulic Modeling ; Probability Analysis

 Digital Object List

 Bookmark

No TOC