Determination of hot and cold rolling textures of steels: Combined Bayesian neural network model

Abstract

The work reported in this paper outlines the use of a combined artificial neural network model capable of fast online prediction of textures in low and extra low carbon steels. We approach the problem by a Bayesian framework neural network model that take into account as input to the model the influence of 23 parameters describing chemical composition, and thermomechanical processes such as austenite and ferrite rolling, coiling, cold working and subsequent annealing involved on the production route of low and extra low carbon steels. The output of the model is in the form of fiber texture data. The predictions of the network provide an excellent match to the experimentally measured data. The results presented in this paper demonstrate that this model can help on optimizing the normal anisotropy (r m) of steel products.

Probabilistic Force Prediction in Cold Sheet Rolling by Bayesian Inference

“…Karandikar and Zapata applied Bayesian updating to investigate spindle speed selection for optimum tool life and cost in milling [10,11], In rolling, Ettler discussed an implementation overview of Bayes' theorem when including several models to control strip thickness, but gave no details of the models, their parameters, or the results [12], Capdevila describes a Bayesian/neural-network application for deter mining the textures in hot and cold-rolled steels. [13]. The literature does not reveal Bayesian inference investigations that provide detailed parameter analysis examples of probabilistic force predictions in cold rolling of flat metals.…”

Prediction on the Distributions of the Strength and Toughness of Thick Steel Plates Based on Bayesian Neural Network

“…At the same time, not only the target prediction with a point estimation but also an uncertainty estimation arises as an important issue of an artificial intelligence [8] and its various applications. [9][10][11][12][13] Especially, estimating the predictive distribution and uncertainties are important to manage the defect rate of the manufacturing process. [14] The operation data of steelmaking process are exposed to nonquantifiable factors: human errors in the process, noises in the data, the weather, and any other complicated physiochemical factors that cannot be recognized through the data.…”

“…Our work is the first application of MC dropout on a steelmaking process. Although there were several approaches [13,18,19] to predict a physical property and its uncertainty with a Bayesian DOI: 10.1002/srin.202100566…”

“…Our work is the first application of MC dropout on a steelmaking process. Although there were several approaches [ 13,18,19 ] to predict a physical property and its uncertainty with a Bayesian techniques. However, several Bayesian neural network (BNN) approaches were not advanced at the past due to its computational limit.…”