Robust design of artificial neural network for roll force prediction in hot strip mill

2002 ◽  
Author(s):  
Young-Sang Kim ◽  
Bong-Jin Yum ◽  
Min Kim
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Robust Design ◽  
Hot Strip Mill ◽  
Roll Force ◽  
Force Prediction ◽  
Hot Strip ◽  
Artificial Neural

MATHEMATICAL-ARTIFICIAL NEURAL NETWORK HYBRID MODEL TO PREDICT ROLL FORCE DURING HOT ROLLING OF STEEL

2013 ◽  
Vol 02 (01) ◽  
pp. 1350004 ◽  
Author(s):  
S. RATH ◽  
P. P. SENGUPTA ◽  
A. P. SINGH ◽  
A. K. MARIK ◽  
P. TALUKDAR
Keyword(s):  
Neural Network ◽  
Mathematical Model ◽  
Artificial Neural Network ◽  
Flow Stress ◽  
Mathematical Models ◽  
Hybrid Model ◽  
Data Driven ◽  
Roll Force ◽  
Hot Strip ◽  
Artificial Neural

Accurate prediction of roll force during hot strip rolling is essential for model based operation of hot strip mills. Traditionally, mathematical models based on theory of plastic deformation have been used for prediction of roll force. In the last decade, data driven models like artificial neural network have been tried for prediction of roll force. Pure mathematical models have accuracy limitations whereas data driven models have difficulty in convergence when applied to industrial conditions. Hybrid models by integrating the traditional mathematical formulations and data driven methods are being developed in different parts of world. This paper discusses the methodology of development of an innovative hybrid mathematical-artificial neural network model. In mathematical model, the most important factor influencing accuracy is flow stress of steel. Coefficients of standard flow stress equation, calculated by parameter estimation technique, have been used in the model. The hybrid model has been trained and validated with input and output data collected from finishing stands of Hot Strip Mill, Bokaro Steel Plant, India. It has been found that the model accuracy has been improved with use of hybrid model, over the traditional mathematical model.

An Application of Physics-Based and Artificial Neural Networks-Based Hybrid Temperature Prediction Schemes in a Hot Strip Mill

2008 ◽  
Vol 130 (1) ◽  
Author(s):  
Wouter M. Geerdes ◽  
Miguel Ángel Torres Alvarado ◽  
Mauricio Cabrera-Ríos ◽  
Alberto Cavazos
Keyword(s):  
Neural Network ◽  
Heat Transfer ◽  
Artificial Neural Network ◽  
Heat Transfer Model ◽  
Hot Strip Mill ◽  
Transfer Model ◽  
Temperature Prediction ◽  
Hybrid Schemes ◽  
Hot Strip ◽  
Artificial Neural

In hot strip mills, estimation of rolling variables is of crucial importance to setting up the finishing mill and meeting dimensional control requirements. although the use of physics-based models is preferred by the specialists to keep the fundamental knowledge of the underlying phenomena, many times a purely empirical model, such as an artificial neural network, will provide better predictions although at the cost of losing such fundamental knowledge. This paper presents the application of physics-based and artificial neural networks-based hybrid models for scale breaker entry temperature prediction in a real hot strip mill. The idea behind combining these two types of models is to capitalize in what are often portrayed as their main advantages: (i) keeping the physics knowledge of the process and (ii) providing better predictions. Temperature prediction schemes with different hybrid levels between a pure heat transfer model and an artificial neural network alone were evaluated and compared showing promising results in this case study. Using an artificial neural network together with the heat transfer model helped to achieve better temperature predictions than using the heat transfer model alone in every instance, thereby proving the hybrid schemes attractive to the industry. In this work, three different hybrid schemes combining the knowledge imbedded in a heat transfer model and the prediction capabilities of an artificial neural network in temperature prediction in a hot strip mill were tried. The hybrid models came out quite competitive in this case study. The results support the use of empirical models to foster the prediction ability of physics-based models; that is, they make the case for their joint use as opposed to their exclusive use.

Robust Design of Artificial Neural Network Methodology to Solve the Inverse Kinematics of a Manipulator of 6 DOF

2021 ◽  
pp. 171-210
Author(s):  
Ma. del Rosario Martínez-Blanco ◽  
Teodoro Ibarra-Pérez ◽  
Fernando Olivera-Domingo ◽  
José Manuel Ortiz-Rodríguez
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Robust Design ◽  
Inverse Kinematics ◽  
Artificial Neural

scholarly journals An Evaluation of Hand-Force Prediction Using Artificial Neural-Network Regression Models of Surface EMG Signals for Handwear Devices

2017 ◽  
Vol 2017 ◽  
pp. 1-12 ◽  
Author(s):  
Masayuki Yokoyama ◽  
Ryohei Koyama ◽  
Masao Yanagisawa
Keyword(s):  
Neural Network ◽  
Artificial Neural Network ◽  
Regression Models ◽  
Maximum Voluntary Contraction ◽  
Surface Emg ◽  
Voluntary Contraction ◽  
Force Prediction ◽  
Hand Force ◽  
Artificial Neural ◽  
Artificial Neural Network Ann

Hand-force prediction is an important technology for hand-oriented user interface systems. Specifically, surface electromyography (sEMG) is a promising technique for hand-force prediction, which requires a sensor with a small design space and low hardware costs. In this study, we applied several artificial neural-network (ANN) regression models with different numbers of neurons and hidden layers and evaluated handgrip forces by using a dynamometer. A handwear with dry electrodes on the dorsal interosseous muscles was used for our evaluation. Eleven healthy subjects participated in our experiments. sEMG signals with six different levels of forces from 0 N to 200 N and maximum voluntary contraction (MVC) are measured to train and test our ANN regression models. We evaluated three different methods (intrasession, intrasubject, and intersubject evaluation), and our experimental results show a high correlation (0.840, 0.770, and 0.789 each) between the predicted forces and observed forces, which are normalized by the MVC for each subject. Our results also reveal that ANNs with deeper layers of up to four hidden layers show fewer errors in intrasession and intrasubject evaluations.

Sign in / Sign up

Export Citation Format

Share Document