Sensitivity Analysis of Parameters Affecting Scour Depth at Downstream of Twin Bridge Piers Using Extreme Learning Machine

Subject Areas : Farm water management with the aim of improving irrigation management indicators

siamak amiri ¹ , mohammad ali izadbakhsh ^{2 *} , saeid shabanlou ³

1 - Ph.D. Candidate, Department of Water Engineering, Kermanshah Branch, Islamic Azad University, Kermanshah, Iran
2 - Department of Water Engineering, Kermanshah Branch, Islamic Azad University, Kermanshah, Iran
3 - Department of Water Engineering, Kermanshah Branch, Islamic Azad University, Kermanshah, Iran

Received: 2020-09-28 Accepted : 2021-12-18 Published : 2022-04-21

Keywords: Twin bridge piers, Uncertainty Analysis, Scour depth, Extreme Learning Machine (ELM), Partial derivative sensitivity analysis (PDSA),

Abstract :

Background and Aim: Local scouring has been identified as one of the important factors that cause the structure of bridges, breakwaters, and piers to rupture. The complexity of the scouring mechanism has made this one of the most important fields of civil engineering studies. In recent years, many studies have been performed on local scouring around bridge piers. Due to the great importance of predicting and estimating the scour pattern in the vicinity of bridge piers, many studies have been done on this type of structure.Method: In this study, for the first time, using a new extreme learning machine (ELM) method, the scour depth near the foundations of the twin bridges was simulated. First, effective parameters were identified and four ELM models were developed. Then, numerical results were validated using Monte Carlo simulation and the cross-validation method. Then the sin activation function was determined as the best activation function. In addition, ELM results were compared with artificial neural network (ANN) models that ELM models estimated scour values more accurately. Uncertainty analysis was performed for the superior ELM and ANN models and a relationship was proposed for the superior model. Partial derivative sensitivity analysis (PDSA) was also performed for all input parameters.Results: Among the existing activation functions, the sin function had the optimal performance compared to other activation functions. According to the analysis of modeling results, ELM 1 model was introduced as the superior model. This model was a function of all input parameters. Also, by removing the landing number, the accuracy of the numerical model was significantly reduced, so the mentioned parameter was identified as the most effective parameter in scouring modeling around the bases of the twin bridges by the model of Strength training machine.Conclusion: By analyzing the modeling results the superior ELM model was introduced. The results of ELM models were also compared with ANN models, which showed that ELM models simulate scour values more accurately. For the superior ELM model, a relation was proposed to calculate the scour hole depth, and further uncertainty analysis showed that this model had a higher performance than the actual value. In addition, the relative derivative sensitivity analysis for the input parameters showed that with increasing the landing number, the value of the objective function (scour depth) increases.

References:

Azamathulla, H.M. 2012. Gene-expression programming to predict scour at a bridge abutment. Journal of Hydroinformatics, 14(2): 324-331.

Azimi, H., Bonakdari, H. and Ebtehaj, I. 2017. Sensitivity Analysis of the Factors Affecting the Discharge Capacity of Side Weirs in Trapezoidal Channels using Extreme Learning Machines. Flow Measurement and Instrumentation, 54: 216-223.

Azimi, H., Bonakdari, H., and Ebtehaj, I. 2019. Gene expression programming-based approach for predicting the roller length of a hydraulic jump on a rough bed. ISH Journal of Hydraulic Engineering, 1-11.

Azimi, H., Bonakdari, H., Ebtehaj, I., Shabanlou, S., Talesh, S. H. A., and Jamali, A. 2019. A Pareto design of evolutionary hybrid optimization of ANFIS model in prediction abutment scour depth. Sādhanā, 44(7), 1-14.

Azimi, H., Bonakdari, H., Ebtehaj, I., Talesh, S.H.A., Michelson, D.G. and Jamali, A. 2016. Evolutionary Pareto optimization of an ANFIS network for modeling scour at pile groups in clear water condition. Fuzzy Sets and Systems, 319: 50-69.

Bonakdari, H., Moradi, F., Ebtehaj, I., Gharabaghi, B., Sattar, A. A., Azimi, A. H., and Radecki-Pawlik, A. 2020. A non-tuned machine learning technique for abutment scour depth in clear water condition. Water, 12(1), 301.

Ebtehaj, I., Bonakdari, H., Moradi, F., Gharabaghi, B., and Khozani, Z. S. 2018. An integrated framework of Extreme Learning Machines for predicting scour at pile groups in clear water condition. Coastal Engineering, 135, 1-15.

Ebtehaj, I., Sattar, A. M., Bonakdari, H., and Zaji, A. H. 2017. Prediction of scour depth around bridge piers using self-adaptive extreme learning machine. Journal of Hydroinformatics, 19(2), 207-224.

Firat, M. and Gungor, M. 2009. Generalized regression neural networks and feed forward neural networks for prediction of scour depth around bridge piers. Advances in Engineering Software, 40(8): 731-737.

Huang, G.B., Zhu, Q.Y. and Siew, C.K. 2004. Extreme learning Machine: a new learning scheme of feedforward neural networks. International Joint Conference on Neural Networks, 2: 985-90.

Liriano, S.L. and Day, R.A. 2001. Prediction of scour depth at culvert outlets using neural networks. Journal of Hydroinformatics, 3(4): 231-238.

Sharafi, H., Ebtehaj, I., Bonakdari, H. and Zaji, A.H. 2016. Design of a support vector machine with different kernel functions to predict scour depth around bridge piers. Natural Hazards, 84(3): 2145-2162.

Trent, R., Gagarin, N. and Rhodes, J. 1993. Estimating pier scour with artificial neural networks. In Hydraulic Engineering, (1043-1048). ASCE.

Wang, H., Tang, H., Liu, Q. and Wang, Y. 2016. Local scouring around twin bridge piers in open-channel flows. Journal of Hydraulic Engineering, 142(9): 06016008.

Yua, L., Danninga, Z. and Hongbinga C. 2015. Prediction of length-of-day using extreme learning machine. Geodesy and Geodynamics 16(2): 151–159.