Improvement and optimisation of hot dip galvanising line using neural networks and genetic algorithms

Abstract

In the present article, an application is present based on the combination of genetic algorithms and neural networks, used to improve the annealing process of a hot dip galvanising line with steel coils. The main objective is to determine the best settings for a furnace in order to reduce the margin of error between the actual strip temperature and expected temperature, not only for each coil that forms the strip, but also in the zones of the strip where transitions are formed by coils with different dimensions or steel types. Basically, the methodology consists in training a multilayer perceptron (MLP), which then determines the settings of the furnace and the speed of the strip according to the type of coil that forms the same strip. Another MLP is used to predict the dynamic behaviour of the strip related to its fluctuations in speed, as well as the temperature of the furnace. In this way, using simulations and genetic algorithms, the optimum settings of the furnace are determined, as well as the speed of the strip in those zones where there are changes in the coils, namely, in dimensions and types of steel.     

Optimum model for predicting temperature settings on hot dip galvanising line

Abstract

Controlling the annealing cycle in a hot dip galvanising line (HDGL) is vital if each coil treated is to be properly galvanised and the steel is to have the right properties. Current HDGL furnace control models usually take into account the dimensions of the coil to be dipped and, in some cases, the type of steel. This paper presents a new model for monitoring furnace temperature settings, which considers not just the coil dimensions but also the chemical composition of the steel. This enables the model to be adjusted more suitably to each type of steel to be dipped, so that the HDGL annealing cycle is optimised and rendered more efficient in dealing with new products. The ultimate aim is to find a model that is equally efficient for new types of steel coil that have not been processed before and whose dimensions and chemical compositions are different from coils processed previously. To find the best model, this paper compares various new and classical algorithms for developing a precise and efficient prediction model capable of determining the three temperature settings for heating on an HDGL located in Avilés (Spain) on the basis of the physical and chemical characteristics of the coils to be processed and the preset process conditions.     

Influence of air knives on liquid zinc flow in galvanising bath

Liquid zinc flow in a galvanising bath was numerically modelled for four cases, and flow field using three-dimensional views was analysed. Flow pattern analysis shows that there are two vortexes (circulations) close to the zinc flow surface and another two below them, which are attributed to strip movement and two guide rolls. Jets from air knives will generate high pressure and zinc backflow. Based on the results of the jets study, the impact of air knives on zinc flow in a galvanising bath was studied on the lengthwise and widthwise section views. It was found that high pressure and zinc backflow due to jets from air knives have little influence on the flow patterns far away from the strip outlet. It was also found that zinc backflow has little influence on the zinc flow surface far away from the strip outlet, while high pressure and zinc backflow together have some influence, that is, the jets of air knives can effectively influence the zinc flow around the strip outlet. This work lays the foundation for the optimisation of the galvanising process to improve the coating quality of steel strips.     

Simulation of special flow affecting dross formation on steel strip in galvanising bath

The zinc flow in a galvanising bath is numerically simulated for three cases, and the flow field is analysed using three-dimensional views. It is found that the flow near the zinc surface directs from the strip to the ingot side when inductors are equipped and whether ingots are melting or not, and the flow direction is opposite to that near the zinc surface for the case without inductor and ingot, which does not exist during the bath operation process. Whirlpools are found to be formed between the snout and the bath side walls, which will increase the chance of dross particles touching the strip surface. Some effective measures are proposed to stop more dross from adhering to strip. This work will lay foundation for the optimisation of the galvanising process to improve the coating quality of steel strip.     

Steel annealing furnace robust neural network model

Abstract

In this day and age, galvanised coated steel is an essential product in several key manufacturing sectors because of its anticorrosive properties. The increase in demand has led managers to improve the different phases in their production chains. Among the efforts needed to accomplish this task, process modelling can be identified as the one with the most powerful outputs in spite of its non-trivial development. In many fields, such as industrial modelling, multilayer feedforward neural networks are often proposed as universal function approximators. These supervised neural networks are commonly trained by the traditional, back-propagation learning format, which minimises the mean squared error (mse) of the training data. However, in the presence of corrupted or extremely deviated samples (outliers), this training scheme may produce incorrect models, and it is well known that industrial data sets frequently contain outliers. The process modelled is a steel coil annealing furnace in a galvanising line, which shares characteristics with most of the furnaces used in galvanised lines all over the world. This paper reports the effectiveness of robust learning algorithms compared to the classical mse-based learning algorithm for the modelling of a real industry process. From this model an adequate line velocity (the velocity set point) for a coil, depending on its characteristics and the furnace condition to receive this coil (temperature set points), can be obtained. With this set point generation model the operator could set strategies to manage the line, i.e. set the order of the coil to be treated or preview the line's speed conditions for the transitory situations.     

Stress analysis model of strip winding system with a sleeve for a coil of thin stainless steel

In a strip winding process, the sleeve is a hollow cylinder that is mounted between a strip coil and a man-drel to maintain uniform coil shape when the strip coil is very thin, but its deformation behavior has not been investigated before.Thus, a finite element (FE) model was presented to calculate the stress distribu-tion in a sleeve and strip coil when 1-3 mm-thick stainless steel was wound around the sleeve.The FE model was developed by extending a previous model by adding a sleeve between the mandrel and strip, and by modifying the boundary and interaction conditions.The strip winding process was divided into an initial process and a steady-state process.During the initial process, the minimum and maximum pressure re-quired on the belt wrapper to maintain coil shape by self-friction of the strip was calculated by the FE model when the belt wrapper is ejected at the end of the initial process.After the initial process, an ana-lytical model of the steady-state process was established to calculate the stress distribution and was com-pared with the FE model to validate it.The suggested analytical model took 1 1 s to give the same stress distribution that the FE model took 30 d to produce.     

珠江钢厂CSP热连轧层流冷却热过程模拟研究

分析了带钢层流冷却的传热过程，将相变潜热引入到层流冷却热过程数学模型，用过程模拟的方法研究了珠江钢厂CSP热连轧层流冷却系统。讨论了喷水方式、喷嘴位置、板速和板厚对带钢温度场分布的影响。结果表明建立的模型对于珠钢CSP热连轧层流冷却系统有较好的适应性，能够对实际生产起到较好的指导性作用。     

Coating microstructure development during Zinquench galvanising of sheet steel

Abstract

The influence of an elevated strip entry temperature on the kinetics of the galvanising reaction has been investigated for a titanium stabilised, interstitial free steel. Elevated strip entry temperatures (>550°C) have been shown to accelerate greatly alloying between the substrate and the molten zinc, increasing the growth rate of the δ phase from 1 to ～7 μm s^-1 and preventing the formation of the ζ phase. The substantial Fe–Zn phase layer seen at the coating/substrate interface at high strip entry temperatures (>550°C) is attributed to the dissolution of iron from the substrate into the molten zinc in the first second of immersion, and the limited inhibition provided by the bath aluminium content (<0·15 wt-%effective).     

J55热轧钢卷分层原因分析

用光学金相显微镜等分析手段对J55热轧钢卷在焊管生产时一剖二的过程中发现的分层缺陷进行分析。结果表明其产生的原因是由于钢板内部存在严重的MnS和含有Nb、Ti的CN化合物夹杂以及C、P元素的偏析,偏析来自于连铸坯。     

Study and redesign of high temperature batch annealing furnace for production of grain oriented electrical steel

Abstract

Grain oriented electrical steel coils are batch annealed at 1200°C to develop the magnetic properties of the strip. Temperature gradients are known to exist within the coils, which can affect the three important reactions occurring in the steel during annealing. These gradients result from the anisotropic conduction properties of the coil and the application of furnace heat. Heat is applied from the furnace to the coil in the radial direction. Studies of the conduction properties of an electrical steel coil have shown that the radial coefficient is ～20% of the axial coefficient. Computational fluid dynamics was employed to simulate a furnace redesign, comprising increased axial heating and a larger coil size. The results show that the annealing cycle time may be shortened and annealing conditions simultaneously improved. The difference between process requirements and process performance can be significantly narrowed by use of the redesigned furnace.     

Separation behaviour of zinc dross from hot dip galvanising melts of different aluminium concentrations by alternating magnetic field

Abstract

Dross in zinc pots is the most important factor that impairs the surface quality of hot dip galvanising automobile steel sheet. Separation behaviour of zinc dross from galvanising melts of different Al concentrations (0·12, 4·5 and 55·0 wt-% by an alternating magnetic field has been investigated in order to reduce the dross defects on steel sheets. The properties of different zinc drosses were investigated by means of EDAX and the quantitative metallographic method and the conductivities were measured by physics property measurements system. The experimental results show that the zinc dross in different hot dip galvanising melts has different morphologies, compositions and structures, and the average particle size of the dross increases with the Al concentration. In the case of extra iron (above the solubility limit) in the zinc melt, zinc dross quantity also increases with the Al concentration. All of the three types of zinc dross particles >5 μm can be successfully separated when the magnetic frequency is 17·5 kHz, effective magnetic flux intensity is 0·05 T, imposed time is 15 s and the cross-section of the ceramic square pipe is 5 × 5 mm.     

Stress–Strain State of Coiled Steel

In a new mathematical model of the stress–strain state of steel strip in the course of cooling, the nonplanarity, surface roughness, and transverse thickness variation (convexity of the cross section) are taken into account. The stress–strain state of a coil of thin steel sheet has a significant influence on factors such as the temperature distribution in the coil; the scale formation on cooling in the course of hot rolling; the adhesion of adjacent turns in the annealing of cold-rolled strip; and the shape of the coil itself. The mathematical model is based on representation of the coil as individual nested hollow cylinders of finite length. The cylinders are divided into sections over the width. The sum of solutions of the Lame equation for individual sections is shown to converge to the solution for the cylinder as a whole. The model permits calculation of the coil’s stress–strain state, taking account of gap formation between adjacent turns as a result of the transverse variation in strip thickness. The modeling results show how the radial and tangential stress formed in strip winding is distributed within the coil. The model permits calculation of the stress–strain state of the coil in the winding of even strip; in the winding of convex even strip with no tension; in the loose winding of convex even strip with tension less than that in tight winding; in tight winding of even convex strip with the correct tension; and in the winding of convex uneven strip without tension. The decrease in distance between contacting rough surfaces is calculated on the basis of a probabilistic approach. An algorithm is presented for calculation of the coil’s stress–strain state. The result obtained for the stress distribution in the coil is typical for the winding of steel strip. The model is verified for the winding of hot-rolled strip, in terms of the size of the region with tight contact of adjacent turns. The tightness of contact is assessed on the basis of the temper color on the edges of the hot-rolled strip. The discrepancy between the calculated and measured size of the region with tight contact is 3%.     

Mechanical properties of alloy Ti–6Al–4V and of stainless steel 316L processed by selective laser melting: influence of out-of-equilibrium microstructures

Abstract

Ti–6Al–4V and stainless steel 316L have been processed by selective laser melting under similar conditions, and their microstructures and mechanical behaviours have been compared in details. Under the investigated conditions, Ti–6Al–4V exhibits a more complex behaviour than stainless steel 316L with respect to the occurrence of microstructural and mechanical anisotropy. Moreover, Ti–6Al–4V appears more sensitive to the build-up of internal stresses when compared with stainless steel 316L, whereas stainless steel 316L appears more prone to the formation of ‘lack of melting’ defects. This correlates nicely with the difference in thermal conductivity between the two materials. Thermal conductivity was shown to increase strongly with increasing temperature and the thermophysical properties appeared to be influenced by variations in the initial metallurgical state.     

Nonuniform structure and properties in hot-rolled low-carbon steel coils

The austenite phase transitions in the industrial production of hot-rolled steel coils are studied magnetometrically. Nonuniformity of the structure and mechanical properties of low-carbon steel strip over its length and width may be attributed to incomplete phase transitions in the strip on the output roller conveyer and nonuniformity of the subsequent coil cooling in regular production conditions. Inheritance of the structural inhomogeneity formed after hot rolling may be responsible for the thickness variation observed after cold rolling of the strip. Recommendations are made for reducing or eliminating the structural inhomogeneity so as to improve product quality.     

Application of Microstructure Engineering in Steel Coil Cooling Process

Aconventionalhotstripmillconsistsmainlyof areheatingfurnace,descalingunits,roughing stands,finishingstands,runouttableanddowncoil ers.Metallurgicalprocessesofsteelstripassociated withhotrollingareausteniterecoveryandrecrystal lizationinrollingstands,pha…     

转鼓式飞剪在1780热轧带钢生产线的应用

飞剪是热带钢连轧设备中不可缺少的剪切设备,它对带钢的咬入、减少带坯对轧辊的冲击、防止轧制过程中出现卡钢事故、使钢卷尾部平齐、易于卷取和打包都起到决定性作用,对产品质量的好坏和成材率有直接联系。从工艺角度介绍了承钢1780热轧带钢生产线中转鼓式飞剪的结构及特点、工艺要求、控制及连锁要求、在线监测控制。针对生产中存在的问题,提出了解决措施。     

Online prediction and monitoring of mechanical properties of industrial galvanised steel coils using neural networks

In galvanising line of cold rolling mill, mechanical properties, i.e. yield strength (YS) and ultimate tensile strength (UTS), are achieved by controlling the key process parameters within specified limits. In this paper, a feed-forward back-propagation artificial neural network (ANN) is proposed to predict the mechanical properties of a coil from its chemical composition, thickness, width and key galvanising process parameters. Principal component analysis is used to avoid redundancy and collinearity effects in input variables for the ANN. The model predicted the YS and UTS with an accuracy of ±10?megapascal (MPa) for 90% of the data. The model was implemented in the continuous galvanising line of Tata Steel, India. An online quality monitoring system was developed to monitor the predicted mechanical properties and process parameters of a galvanised coil. This system helps quality team in decision making.     

Online mechanical property prediction system for hot rolled IF steel

Abstract

Traditionally, mechanical property estimation is carried out by destructive testing, which is costly and time consuming. Sometimes, the time schedule in the mill is so tight that coils are dispatched, while the samples are still under investigation; thus, knowledge of the strip quality immediately after rolling without mechanical testing can save a lot of time and money. As the rolling process is complex and final mechanical properties of steel depend on many parameters, it is almost impossible to develop an accurate first principle based mathematical model, so an artificial neural network based model to predict the mechanical properties of hot rolled steel strip has been developed. This paper describes the neural network based online system that helps in predicting mechanical properties of interstitial free (IF) steel strip and also elaborates how this models can help in capturing various metallurgical phenomena during rolling.     

Influence of process parameters on scale reduction with H2

Abstract

A new pickling method that reduces the metal oxide on steel strip using H₂ has been studied through an extensive campaign of experimental tests. The process, referred to as 'acid free pickling', is environmentally friendly and does not use acid. It is most suitable for application to hot rolled low carbon steel strip. Experimental studies are in progress to validate the application of hydrogen on different metals.     

Prediction of shape defects during cooling of hot rolled low carbon steel strip

Abstract

Shape defects are found in hot rolled steel strip when unwrapping tightly wound coils. This problem is particularly acute in thin strips that were considered to be defect free while processing. A model developed to predict the occurrence and magnitude of such defects in hot rolled low carbon steel strip is described in the present paper. The model assumes that the strip is free of shape defects as it exits the last stand of a continuous mill, but, as a result of processing conditions, thermal and microstructural gradients are present across the width of the strip. It is considered that the variation of ferrite and austenite mixture is caused by the chemical composition of the steel and the actual temperature of the strip. On cooling to room temperature, the distribution of both temperature and microstructure will cause variation in the local contraction that the steel is subjected to, and will promote shape defects.