Chalak Qazani, Mohammad Reza, Bidabadi, Behrooz Shirani, Asadi, Houshyar, Nahavandi, Saeid, and Bidabadi, Farnoosh Shirani (2023) Multiobjective optimization of roll-forming procedure using NSGA-II and type-2 fuzzy neural network. IEEE Transactions on Automation Science and Engineering, 21 (3). pp. 3842-3851.

PDF (Published Version) - Published Version Restricted to Repository staff only

Abstract

In this research, the effective indexes in the cold roll-forming procedure that can affect the energy utilization and required maximum torque of the forming line have been investigated and optimised using NSGA-II and type-2 fuzzy neural networks. The effective parameters were strip thickness, bending angle increment, flange width, inter-distance between the rolling stands and bending radius. Recently, traditional machine-learning applications have been employed in roll-forming technology for different purposes, such as prediction of web-warping, energy efficiency, and strip breakage. A finite element model (FEM) roll-forming procedure was utilised to extract the appropriate datasets for this study. type-2 fuzzy neural network (T2FNN) is not employed in cold roll-forming technology. In this study, T2FNN is employed to imitate the dynamic model of the cold roll-forming procedure to estimate energy consumption and torque. In the following, the NSGA-II extracts the optimal cold roll-forming procedure parameters to reach the lowest energy consumption and maximum torque as the process’s most economical solution. The proposed model is designed and developed under MATLAB software. Fourteen optimal solutions are suggested based on the extracted Pareto-Front of the NSGA-II using the T2FNN of the process. Note to Practitioners—In this research, a hybrid machine-learning method is designed and developed with a combination of T2FNN and NSGA-II to extract the optimal roll-forming procedure parameters to reach the lowest usage of energy as well as the lowest requirement of maximum torque. Implementing the proposed method in cold roll-forming production lines can save millions of dollars in massive factories by reducing the usage of energy and the emission of greenhouse gases. The proposed algorithm is quite fast. Then there is no need for high graphic computers for real-time implementation of the proposed method.

Item ID:	86727
Item Type:	Article (Research - C1)
ISSN:	1558-3783
Copyright Information:	© 2023 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
Funders:	Australian Research Council (ARC)
Date Deposited:	15 Oct 2025 01:58
FoR Codes:	40 ENGINEERING > 4014 Manufacturing engineering > 401408 Manufacturing processes and technologies (excl. textiles) @ 30% 40 ENGINEERING > 4017 Mechanical engineering > 401706 Numerical modelling and mechanical characterisation @ 30% 46 INFORMATION AND COMPUTING SCIENCES > 4611 Machine learning > 461199 Machine learning not elsewhere classified @ 40%
SEO Codes:	24 MANUFACTURING > 2499 Other manufacturing > 249999 Other manufacturing not elsewhere classified @ 30% 17 ENERGY > 1701 Energy efficiency > 170102 Industrial energy efficiency @ 70%
Downloads:	Total: 1 Last 12 Months: 1
	More Statistics