Optimization of spring back in U-die bending process of sheet metal using ANN and ICA

Mojtaba Sheikhi Azqandi,Navid Nooredin,Ali Ghoddosian 국제구조공학회 2018 Structural Engineering and Mechanics, An Int'l Jou Vol.65 No.4

The controlling and prediction of spring back is one of the most important factors in sheet metal forming processes which require high dimensional precision. The relationship between effective parameters and spring back phenomenon is highly nonlinear and complicated. Moreover, the objective function is implicit with regard to the design variables. In this paper, first the influence of some effective factors on spring back in U-die bending process was studied through some experiments and then regarding the robustness of artificial neural network (ANN) approach in predicting objectives in mentioned kind of problems, ANN was used to estimate a prediction model of spring back. Eventually, the spring back angle was optimized using the Imperialist Competitive Algorithm (ICA). The results showed that the employment of ANN provides us with less complicated and time-consuming analytical calculations as well as good results with reasonable accuracy.

A hybrid imperialist competitive ant colony algorithm for optimum geometry design of frame structures

Sheikhi, Mojtaba,Ghoddosian, Ali Techno-Press 2013 Structural Engineering and Mechanics, An Int'l Jou Vol.46 No.3

This paper describes new optimization strategy that offers significant improvements in performance over existing methods for geometry design of frame structures. In this study, an imperialist competitive algorithm (ICA) and ant colony optimization (ACO) are combined to reach to an efficient algorithm, called Imperialist Competitive Ant Colony Optimization (ICACO). The ICACO applies the ICA for global optimization and the ACO for local search. The results of optimal geometry for three benchmark examples of frame structures, demonstrate the effectiveness and robustness of the new method presented in this work. The results indicate that the new technique has a powerful search strategies due to the modifications made in search module of ICACO. Higher rate of convergence is the superiority of the presented algorithm in comparison with the conventional mathematical methods and non hybrid heuristic methods such as ICA and particle swarm optimization (PSO).

A hybrid imperialist competitive ant colony algorithm for optimum geometry design of frame structures

Mojtaba Sheikhi,Ali Ghoddosian 국제구조공학회 2013 Structural Engineering and Mechanics, An Int'l Jou Vol.46 No.3

This paper describes new optimization strategy that offers significant improvements in performance over existing methods for geometry design of frame structures. In this study, an imperialist competitive algorithm (ICA) and ant colony optimization (ACO) are combined to reach to an efficient algorithm, called Imperialist Competitive Ant Colony Optimization (ICACO). The ICACO applies the ICA for global optimization and the ACO for local search. The results of optimal geometry for three benchmark examples of frame structures, demonstrate the effectiveness and robustness of the new method presented in this work. The results indicate that the new technique has a powerful search strategies due to the modifications made in search module of ICACO. Higher rate of convergence is the superiority of the presented algorithm in comparison with the conventional mathematical methods and non hybrid heuristic methods such as ICA and particle swarm optimization (PSO).

Sensitivity analysis based on complex variables in FEM for linear structures

Azqandi, Mojtaba Sheikhi,Hassanzadeh, Mahdi,Arjmand, Mohammad Techno-Press 2019 Advances in computational design Vol.4 No.1

One of the efficient and useful tools to achieve the optimal design of structures is employing the sensitivity analysis in the finite element model. In the numerical optimization process, often the semi-analytical method is used for estimation of derivatives of the objective function with respect to design variables. Numerical methods for calculation of sensitivities are susceptible to the step size in design parameters perturbation and this is one of the great disadvantages of these methods. This article uses complex variables method to calculate the sensitivity analysis and combine it with discrete sensitivity analysis. Finally, it provides a new method to obtain the sensitivity analysis for linear structures. The use of complex variables method for sensitivity analysis has several advantages compared to other numerical methods. Implementing the finite element to calculate first derivatives of sensitivity using this method has no complexity and only requires the change in finite element meshing in the imaginary axis. This means that the real value of coordinates does not change. Second, this method has the lower dependency on the step size. In this research, the process of sensitivity analysis calculation using a finite element model based on complex variables is explained for linear problems, and some examples that have known analytical solution are solved. Results obtained by using the presented method in comparison with exact solution and also finite difference method indicate the excellent efficiency of the proposed method, and it can predict the sustainable and accurate results with the several different step sizes, despite low dependence on step size.

Optimization of spring back in U-die bending process of sheet metal using ANN and ICA

Azqandi, Mojtaba Sheikhi,Nooredin, Navid,Ghoddosian, Ali Techno-Press 2018 Structural Engineering and Mechanics, An Int'l Jou Vol.65 No.4

The controlling and prediction of spring back is one of the most important factors in sheet metal forming processes which require high dimensional precision. The relationship between effective parameters and spring back phenomenon is highly nonlinear and complicated. Moreover, the objective function is implicit with regard to the design variables. In this paper, first the influence of some effective factors on spring back in U-die bending process was studied through some experiments and then regarding the robustness of artificial neural network (ANN) approach in predicting objectives in mentioned kind of problems, ANN was used to estimate a prediction model of spring back. Eventually, the spring back angle was optimized using the Imperialist Competitive Algorithm (ICA). The results showed that the employment of ANN provides us with less complicated and time-consuming analytical calculations as well as good results with reasonable accuracy.

Natural Rubber Bearing Incorporated with Steel Ring Damper (NRB-SRD)

Javad Sheikhi,Mojtaba Fathi 한국강구조학회 2020 International Journal of Steel Structures Vol.20 No.1

The present paper is aimed to investigate the behavior of Natural Rubber Bearing incorporated with steel ring damper (NRBSRD). These types of dampers are integrated of several steel rings which are considered with two confi gurations namely, continual steel ring damper and separate steel ring damper and are inserted between top and bottom plates. The performance characteristics of the system such as eff ective horizontal stiff ness, energy dissipation, equivalent viscous damping and residual deformation are calculated and then compared with the results of high damping rubber bearings and also shape memory alloy (SMA)-lead core rubber bearing (SMA-LRB). The results show that the energy dissipation in steel rings are mainly based on plastic deformation due to fl exural behavior of the rings. NRB-SRD shows better performance in energy dissipation comparing to SMA-LRB and HDRB. These additional dampers show higher stability and energy dissipation in low shear strains due to developing of link between structure and substructure having desirable initial stiff ness under weak earthquakes and wind loads and also in higher shear strains due to creation of higher energy dissipation, stability and secondary stiff ening.

The effects of cluster pruning and the K:N ratio on greenhouse tomato yield and quality

Majid Fallah,Mojtaba Delshad,Hossein Sheikhi 한국원예학회 2021 Horticulture, Environment, and Biotechnology Vol.62 No.5

Both the number of fruit in a cluster and the K:N ratio in the nutrient solution can aff ect the yield and quality of greenhousegrowntomatoes ( Solanum lycopersicum L.). To prevent the loss of photoassimilates to fruit at the end of clusters, which rarelymature and ripen, it is necessary to know the optimal number of fruit in a cluster under diff erent nutritional conditions. In thisstudy, we investigated the eff ects of cluster pruning and the K:N ratio in the nutrient solution on yield, yield components, andquality of tomato fruit grown in a greenhouse. The treatments consisted of fi ve levels of cluster pruning (control or unpruned,and retaining 4, 6, 8, and 10 fruit per cluster), and two levels of the K:N ratio in the nutrient solution (2:1 and 4:1). Underboth K:N ratios, the unpruned and ten-fruit cluster treatments had the highest yield. These treatments also had the highestunripe fruit weight. The highest ripe fruit dimensions and weight were found in four-fruit clusters fed by the solution with aK:N ratio of 2:1. Fruit fi rmness was higher than in other treatments in four-fruit clusters when the K:N ratio was 4:1. Fruitharvested from the four-fruit cluster treatment had the highest titratable acidity compared to the other treatments, while theK:N ratio did not have a signifi cant eff ect on this trait. The highest total soluble solids were found in fruit obtained fromfour-fruit clusters grown with the 2:1 K:N solution. In general, we concluded that under our growing conditions, keepingten fruit in a cluster and adjusting the K:N ratio to 2:1 in the nutrient solution is advisable for the cultivar of cherry tomatoused in this study, although some fruit quality factors may be improved when the number of fruit in clusters is manipulated.

CO₂ emissions optimization of reinforced concrete ribbed slab by hybrid metaheuristic optimization algorithm (IDEACO)

Shima Bijari,Mojtaba Sheikhi Azqandi Techno-Press 2023 Advances in computational design Vol.8 No.4

This paper presents an optimization of the reinforced concrete ribbed slab in terms of minimum CO₂ emissions and an economic justification of the final optimal design. The design variables are six geometry variables including the slab thickness, the ribs spacing, the rib width at the lower and toper end, the depth of the rib and the bar diameter of the reinforcement, and the seventh variable defines the concrete strength. The objective function is considered to be the minimum amount of carbon dioxide gas (CO₂) emission and at the same time, the optimal design is economical. Seven significant design constraints of American Concrete Institute's Standard were considered. A robust metaheuristic optimization method called improved dolphin echolocation and ant colony optimization (IDEACO) has been used to obtain the best possible answer. At optimal design, the three most important sources of CO₂ emissions include concrete, steel reinforcement, and formwork that the contribution of them are 63.72, 32.17, and 4.11 percent respectively. Formwork, concrete, steel reinforcement, and CO₂ are the four most important sources of cost with contributions of 67.56, 19.49, 12.44, and 0.51 percent respectively. Results obtained by IDEACO show that cost and CO₂ emissions are closely related, so the presented method is a practical solution that was able to reduce the cost and CO₂ emissions simultaneously.