Elsevier

Measurement

Volume 87, June 2016, Pages 87-98
Measurement

Application of a genetic algorithm in predicting the percentage of shear force carried by walls in smooth rectangular channels

https://doi.org/10.1016/j.measurement.2016.03.018Get rights and content

Highlights

  • Percentage of shear force (%SFw) carried by walls in smooth rectangular channels was predicted.

  • Genetic Algorithm Artificial (GAA) neural network and Genetic Programming (GP) was used.

  • Eight data series with a total of 69 different data were used.

  • A program and an equation for the GP and GAA respectively, are presented.

  • Obtained results indicate the superiority of GAA in predicting %SFw.

Abstract

Shear stress comprises basic information for predicting average depth velocity and discharge in channels. With knowledge of the percentage of shear force carried by walls (%SFw) it is possible to more accurately estimate shear stress values. The %SFw in smooth rectangular channels was predicted by extending two soft computing methods: Genetic Algorithm Artificial (GAA) neural network and Genetic Programming (GP). In order to investigate the percentage of shear force, 8 data series with a total of 69 different data were used. The outcomes of the GAA model (an equation) and the GP model (a program) were presented. In order to detect these models’ ability to predict %SFw, the obtained results were compared with several equations derived by other researchers. The GAA model with RMSE of 2.5454 and the GP model with RMSE of 3.0559 performed better than other equations with mean RMSE of about 9.630.

Introduction

Numerous engineering problems such as erosion of channels, flow resistance, sediment pollutant transport and designing stable channels require knowledge of boundary shear stress distribution values. Many researchers have tried to measured shear stress distribution experimentally, as presented by Bonakdari et al. [1] in detail. Since the experimental estimation of boundary shear stress is a difficult and time consuming procedure, other researchers have attempted to calculate boundary shear stress by analytical and numerical methods [2], [3], [4], [5], [6], [7], [8], [9], [10], [11]. The use of soft computing technique in modeling complicated problems is recently expanding. Khatibi et al. [12] used the GP model to predict the outflow of a river reach from its given inflows, and by comparing the results of three methods they found the GP model made better predictions. Kisi et al. [13] estimated daily suspended sediment load using a GP model. They also compared this method with several soft computing techniques and deducted that the GP model operates better than the others. Genetic programming through bi-objective genetic algorithms was applied to model different phenomena [14], [15]. Huai et al. [16] used the ANN model to estimate the apparent shear stress acting at the vertical line between subareas of a compound channel’s cross section. The shear stress is a very important parameter in channel problems but only few studies have dealt with estimating the percentage of shear force in channels. Cobaner et al. [17] utilized an ANN model in order to estimate the shear force carried by walls in rectangular channels and ducts. Sheikh Khozani et al. [18] predicted percentage of shear force carried by walls in rectangular channels with rough boundaries by using GP and GAA methods.

In this study, two soft computing techniques, namely the GAA and GP model are extended for computing the percentage of shear force carried by walls in smooth rectangular channels. These models are explained step-by-step and the best output of models, a program and an equation for the GP and GAA respectively, are presented. The models are also compared in order to determine which has greater prediction ability. Finally, a comparison between the proposed models and the equations obtained by former researchers is carried out.

Section snippets

GAA processing

A Multi-Layer Perceptron Artificial Neural Network (MLP-ANN) consists of three types of layers: input, hidden and output layers. Each layer has some neurons that collect the weighted summation of the previous layer’s neurons and transfer the results to the next layer by a nonlinear transfer function. The number of neurons in the input layer is equal to the input variables while the number of neurons in the output layer is equal to the output variables. In this study, both the input and output

Optimum determination of the GAA and GP models

In modeling with the GAA and GP techniques, the optimum one was selected first by investigating some fitness functions. Therefore, to select the appropriate fitness functions in the GAA and GP models, the MSE and MAE functions were examined. The comparison results for the two different functions using statistical parameters are presented in Table 3. In the first step of the GAA process, the logarithmic function for input layers, purelin function for output layers and aspect ratio as input data

Conclusion

In this research, the ability of GAA and GP models to estimate the percentage of shear force carried by walls (%SFw) in smooth rectangular channels was investigated. This study demonstrated these methods step-by-step. Different fitness functions were examined in order to select the best for both models. Several transfer functions were investigated for the GAA model and finally, the best equation for the GAA model was presented by MSE fitness function and hyperbolic tangent in the hidden layer

References (36)

  • K. Yang et al.

    Modeling depth-averaged velocity and boundary shear stress in rectangular compound channels with secondary flows

    J. Hydraul. Eng.

    (2013)
  • A. Kabiri-Samani et al.

    Boundary shear stress in smooth trapezoidal open channel flows

    J. Hydraul. Eng.

    (2013)
  • H. Bonakdari et al.

    Study of shear stress in narrow channels: application to sewers

    Urban Water

    (2008)
  • Z. Sheikh, H. Bonakdari, Prediction of boundary shear stress in circular and trapezoidal channels with entropy concept....
  • A.R. Zarrati et al.

    Semianalytical model for shear stress distribution in simple and compound open channels

    J. Hydraul. Eng.

    (2008)
  • B.K. Giri et al.

    Genetic programming evolved through bi-objective genetic algorithms applied to a blast furnace

    Mater. Manuf. Process.

    (2013)
  • W. Huai et al.

    Predicting apparent shear stress in prismatic compound channels using artificial neural network

    J. Hydroinformatics

    (2013)
  • M. Cobaner et al.

    Boundary shear stress analysis in smooth rectangular channels and ducts using neural networks

    Water Environ. J.

    (2010)
  • Cited by (16)

    • Unequal limit cuckoo optimization algorithm applied for optimal design of nonlinear field calibration problem of a triaxial accelerometer

      2020, Measurement: Journal of the International Measurement Confederation
      Citation Excerpt :

      The frequently used intelligence population-based optimization methods are the Genetic Algorithm (GA) [31] and the Particle Swarm Optimization (PSO) [32]. Nowadays, these algorithms have been promising techniques for the MEMS accelerometer calibration and other real-world optimization problems, some of which can be found in [33–39]. In [33], the GA was used for calibration of the MEMS accelerometer.

    • Estimation of the Darcy–Weisbach friction factor for ungauged streams using Gene Expression Programming and Extreme Learning Machines

      2019, Journal of Hydrology
      Citation Excerpt :

      Artificial intelligence (AI) techniques, such as Gene Expression Programming (GEP), are useful for accurate prediction of hydraulic phenomena. Techniques such as GEP have reasonable calculation times and have become popular with researchers (e.g., Guven and Gunal 2008; Ghani and Azamathulla 2011; Mohammadpour et al. 2013; Ebtehaj et al. 2015a, Sheikh et al. 2016; Azimi et al. 2017; Bonakdari et al., 2018; Sattar et al., 2018) as GEP facilitates the precise prediction of hydraulic phenomena. In addition, GEP has the capacity to create an explicit equation for predicting variables; a powerful advantage of this method for practical applications.

    • An expert system for predicting shear stress distribution in circular open channels using gene expression programming

      2018, Water Science and Engineering
      Citation Excerpt :

      Utilizing soft computing (SC) methods to solve different problems is another progressing concept. SC methods, including artificial neural networks (ANN), particle swarm optimization (PSO), adaptive neuro-fuzzy inference systems (ANFIS), the group method of data handling (GMDH), genetic algorithms (GA), and genetic programming (GP), are used in diverse engineering fields (Melin et al., 2013; Kaydani et al., 2014; Najafzadeh et al., 2014; Najafzadeh and Azamathulla, 2015; Sheikh Khozani et al., 2016a, 2017). One of the newest and most powerful SC methods is gene expression programming (GEP), which is an extension of GP and GA (Azamathulla and Ab Ghani, 2010; Sattar and Gharabaghi, 2015), and was first introduced by Ferreira (2001).

    • An analysis of shear stress distribution in circular channels with sediment deposition based on Gene Expression Programming

      2017, International Journal of Sediment Research
      Citation Excerpt :

      Cobaner et al. (2010) simulated the boundary shear stress in a smooth rectangular channel using Artificial Neural Networks (ANN). Later, Sheikh Khozani et al. (2016a, 2016b) used genetic programing (GP) and genetic artificial algorithms (GAA) to estimate the mean shear force carried by walls in rectangular channels with smooth and rough boundaries. Genc et al. (2015) modeled the shear stress in small streams using ANNs as well.

    • Estimating the shear stress distribution in circular channels based on the randomized neural network technique

      2017, Applied Soft Computing Journal
      Citation Excerpt :

      With the increasing use of soft computing methods for solving hydraulic problems and predicting different phenomena, Cobaner et al. [19] employed artificial neural network (ANN) modeling to estimate the percentage of shear force carried by walls. Sheikh Khozani et al. [20,21] predicted the percentage of shear force carried by walls in rough and smooth rectangular channels using two soft computing techniques. Nikolic et al. [22] utilized the RNN method for sensorless estimation of wind speed based on wind turbine parameters.

    View all citing articles on Scopus
    View full text