Mathematical modelling of burr height of the drilling process using a statistical-based multi-gene genetic programming approach

Drilling is one of the important machining processes performed extensively in production industry. Literature emphasises that the output process parameters such as burr height, surface roughness, strength, etc. are related to and can be improved by the appropriate settings of the input process parameters. Recently, researchers have applied well-known computational intelligence methods such as regression analysis, artificial neural networks (ANNs), support vector regression (SVR), etc. in the prediction of performance characteristics of the drilling process. Alternatively, an evolutionary approach of multi-gene genetic programming (MGGP) that evolves the model structure and its coefficients automatically can be applied. Despite of being widely applied, MGGP has the limitation for producing models that over-fit on the testing data. One of the reasons attributed for this behaviour is the over-size of the evolved models. Therefore, a statistical-based MGGP (S-MGGP) approach is proposed and applied to the burr height data obtained from the drilling of AISI 316L stainless steel. In this proposed approach, Bayesian information criterion is embedded in its paradigm, which punishes the fitness of larger size models. The performance of S-MGGP and ANN models is found to be better than those of the standardised MGGP and SVR. Further, the parametric and sensitivity analysis conducted validates the robustness of our proposed model and is proved to capture the dynamics of the drilling phenomenon by unveiling dominant input process parameters and the hidden non-linear relationships.     

Measurement of properties of graphene sheets subjected to drilling operation using computer simulation

Drilling being one of the primary machining processes find wide spread applications in manufacturing of functional components. Optimization of drilling process performance requires critical understanding of process parameters which govern the mechanism of drilling process. Machining process at nanoscale level has been studied extensively using numerical modeling approaches owing to complexity in conducting experiments at nanoscale level. In this paper, we propose a new evolutionary approach based on multi-gene genetic programming (MGGP) to numerically model the drilling process of graphene sheet, a two dimensional nanoscale material. The performance of our proposed MGGP model is compared with that of the artificial neural network (ANN) and we observe that our predictions are well in agreement with the data obtained using conventional numerical approach for modeling machining process of nanoscale materials. We anticipate that our proposed MGGP model can find applications in optimizing the machining processes of nanoscale materials.     

Application of Taguchi philosophy for parametric optimization of bead geometry and HAZ width in submerged arc welding using a mixture of fresh flux and fused flux

Taguchi philosophy has been applied for obtaining optimal parametric combinations to achieve desired weld bead geometry and dimensions related to the heat-affected zone (HAZ), such as HAZ width in the present case, in submerged arc welding. The philosophy and methodology proposed by Dr. Genichi Taguchi can be used for continuous improvement in products that are produced by submerged arc welding. This approach highlights the causes of poor quality, which can be eliminated by self-adjustment among the values of the process variables if they tend to change during the process. Depending on functional requirements of the welded joint, an acceptable weldment should confirm maximum penetration, minimum reinforcement, minimum bead width, minimum HAZ width, minimum bead volume, etc. to suit its area of application. Hence, there exists an increasing demand to evaluate an optimal parameter setting that would fetch the desired yield. This could be achieved by optimization of welding variables. Based on Taguchi’s approach, the present study has been aimed at integrating statistical techniques into the engineering process. Taguchi’s L9 (3**3) orthogonal array design has been adopted and experiments have been accordingly conducted with three different levels of conventional process parameters using welding current and flux basicity index to obtain bead-on-plate weld on mild steel plates. Features of bead geometry and HAZ in terms of bead width, reinforcement, depth of penetration and HAZ width have been measured for each experimental run. The slag, generated during welding, has been consumed in further runs by mixing it with fresh unmelted flux. The percentage of slag in the mixture of fused flux (slag) and fresh flux has been defined as slag-mix%. Welding has been performed by using varying slag-mix%, treated as another process variable, in order to obtain the optimum amount of slag-mix that can be used without any alarming adverse effect on features of bead geometry and HAZ. This would lead to ‘waste to wealth’.     

Application of artificial neural network for the prediction of laser cladding process characteristics at Taguchi-based optimized condition

This paper presents an investigation on the optimization of multiple performance characteristics during CO₂ laser cladding process considering clad width and clad depth as performance characteristics. This optimization for multiple quality characteristics has been done using Taguchi’s quality loss function. The process model for laser cladding operation using various techniques like artificial neural network (ANN) has rarely been found in the literature review. In the present work, a number of experiments have been performed to establish the interrelationship between process variables and response variables using the back propagation method of ANN. The essential input process parameters are identified as laser power, scan speed of work table, and powder feed rate. Moreover, the analysis of variance is also employed to determine the contribution of each control parameter on clad bead quality. In order to validate the predicted result, an experiment as confirmatory test is carried out at the optimized cladding condition. It is observed that the confirmatory experimental result is showing a good agreement with the predicted one. However, it has been found that the optimum condition of the cladding parameters for multi-performance characteristics varies with the different combinations of weighting factors.     

Robotic GMAW online learning: issues and experiments

This paper presents three main contributions: (i) an experimental analysis of variables, using well-defined statistical patterns applied to the main parameters of the welding process. (ii) An on-line/off-line learning and testing method, showing that robots can acquire a useful knowledge base without human intervention to learn and reproduce bead geometries. And finally, (iii) an on-line testing analysis including penetration of the bead, that is used to train an artificial neural network (ANN). For the experiments, an optic camera was used in order to measure bead geometry (width and height). Also real-time computer vision algorithms were implemented to extract training patterns. The proposal was carried out using an industrial KUKA robot and a GMAW type machine inside a manufacturing cell. We present expermental analysis that show different issues and solutions to build an industrial adaptive system for the robotics welding process.     

An improved artificial neural network for laser welding parameter selection and prediction

In the laser welding production, the selection and prediction of welding parameters is essentially important to guarantee weld quality. Artificial neural networks (ANN), which perform a nonlinear mapping between inputs and outputs, are an alternative approach for developing welding parameter forecasting model. In this paper, in order to speed up the convergence and avoid local minimum of the conditional ANN, genetic algorithm simulated annealing (GASA) based on the random global optimization is inducted into the network training. By means of GASA method, weights and threshold of neural networks can be globally optimized with short training time. Meanwhile, the gray correlation model (GCM) is used as a pre-processing tool to simplify the original networks based on obtaining the main influence elements of network inputs. The GCM–GASA–ANN method combines the complementary features of three computational intelligence techniques and owns very good applicability. Through the simulation and analysis of an orthogonal experiment, the proposed method can be proved to have higher accuracy and to perform better than the traditional ANN to forecast the laser welding parameters.     

An integrated system for setting the optimal parameters in IC chip-package wire bonding processes

The wire bonding process is the key process in an IC chip-package. It is an urgent problem for IC chip-package industry to improve the wire bonding process capability. In this study, an integrated system is proposed to identify and control parameters in the wire bonding process in order to achieve high level performance and quality. First, an experimental design with Taguchi method is applied to identify the critical parameters in the wire bonding process. Then, an ANN is used to establish the nonlinear multivariate relationships between wire boning parameters and responses. Finally, a GA is adopted to find the most desired parameter settings by using the output of ANN as the fitness measure. Another popular method, response surface method, for parameter design problems is conducted for comparison purpose. Results of this comparison demonstrate the effectiveness of the proposed approach.     

Solving multi-criteria optimization problem in submerged arc welding consuming a mixture of fresh flux and fused slag

In the present work, application of the Taguchi method in combination with grey relational analysis has been applied for solving multiple criteria (objective) optimization problem in submerged arc welding (SAW). A grey relational grade evaluated with grey relational analysis has been adopted to reveal an optimal parameter combination in order to obtain acceptable features of weld quality characteristics in submerged arc bead-on-plate welding. The idea of slag utilization, in subsequent runs, after mixing it with fresh unmelted flux, has been introduced. The parentage of slag in the mixture of fresh flux and fused flux (slag) has been denoted as slag-mix%. Apart from two conventional process parameters: welding current and flux basicity index, the study aimed at using varying percentages of slag-mix, treated as another process variable, to show the extent of acceptability of using slag-mix in conventional SAW processes, without sacrificing any characteristic features of weld bead geometry and HAZ, within the experimental domain. The quality characteristics associated with bead geometry and HAZ were bead width, reinforcement, depth of penetration and HAZ width. Using grey relational grade as performance index, we have performed parametric optimization yielding the desired features of bead geometry and HAZ. Predicted results have been verified with confirmatory experiments, showing good agreement. This proves the utility of the proposed method for quality improvement in SAW process and provides the maximum (optimum) amount of slag-mix that can be consumed in the SAW process without any negative effect on characteristic features of the quality of the weldment in terms of bead geometry.     

Modeling and optimization of features of bead geometry including percentage dilution in submerged arc welding using mixture of fresh flux and fused slag

Quality has now become an important issue in today’s manufacturing world. Whenever a product is capable of conforming to desirable characteristics that suit its area of application, it is termed as high quality. Therefore, every manufacturing process has to be designed in such a way that the outcome would result in a high quality product. The selection of the manufacturing conditions to yield the highest desirability can be determined through process optimization. Therefore, there exists an increasing need to search for the optimal conditions that would fetch the desired yield. In the present work, we aim to evaluate an optimal parameter combination to obtain acceptable quality characteristics of bead geometry in submerged arc bead-on-plate weldment on mild steel plates. The SAW process has been designed to consume fused flux/slag, in the mixture of fresh flux. Thus, the work tries to utilize the concept of ‘waste to wealth’. Apart from process optimization, the work has been initiated to develop mathematical models to show different bead geometry parameters, as a function of process variables. Hence, optimization has been performed to determine the maximum amount of slag--flux mixture that can be used without sacrificing any negative effect on bead geometry, compared to the conventional SAW process, which consumes fresh flux only. Experiments have been conducted using welding current, slag-mix percentage and flux basicity index as process parameters, varied at four different levels. Using four³ full factorial designs, without replication, we have carried out welding on mild steel plates to obtain bead-on-plate welds. After measuring bead width, depth of penetration and reinforcement; based on simple assumptions on the shape of bead geometry, we calculated other relevant bead geometry parameters: percentage dilution, weld penetration shape factor, weld reinforcement form factor, area of penetration, area of reinforcement and total bead cross sectional area. All these data have been utilized to develop mathematical models between predictors and responses. Response surface methodology (RSM), followed by the multiple linear regression method, has been applied to develop these models. The effects of selected process parameters on different responses have been represented graphically. Finally grey relational analysis coupled with the Taguchi method (with Taguchi’s orthogonal array) has been applied for parametric optimization of this welding technique. Confirmatory experiments have been conducted to verify optimal results.     

基于混合算法的磁悬浮隔振器模型辨识与验证

针对磁悬浮隔振器动态电磁力模型存在非线性及磁滞且很难建立其精确模型的问题,提出了基于BP算法、改进遗传（MGA）算法的混合算法的BP神经网络的模型辨识方法,建立了磁悬浮隔振器动态电磁力气隙电流关系的模型。结果表明,基于混合训练算法辨识得到的模型具有更高的精度,能够满足磁悬浮隔振器动态电磁力模型辨识需求。最后,搭建了磁悬浮隔振实验平台,建立控制模型,并验证了辨识模型的有效性。     

Development of RSM- and ANN-based models to predict and analyze the effects of process parameters of laser-hardened commercially pure titanium on heat input and tensile strength

Laser transformation hardening (LTH) is an innovative and advanced laser surface modification technique as compared to conventional transformation hardening processes and has been employed in aerospace, marine, chemical applications, heat exchangers, cryogenic vessels, components for chemical processing and desalination equipment, condenser tubing, airframe skin, and nonstructural components which introduces the advantageous residual stresses into the surface, improving the mechanical properties like wear, resistance to corrosion, tensile strength, and fatigue strength. In the present study, LTH of commercially pure titanium, nearer to ASTM grade 3 of chemical composition was investigated using continuous wave 2 kW, Nd: YAG laser. The effect of laser process variables such as laser power, scanning speed, and focused position was investigated using response surface methodology (RSM) and artificial neural network (ANN) keeping argon gas flow rate of 10 lpm as fixed input parameter. This paper describes the comparison of the heat input (HI) and ultimate tensile strength (σ) (simply called as tensile strength) predictive models based on ANN and RSM. The paper also presents the effect of laser process variables on the HI and ultimate σ. The research work also emphasizes on the effect of HI on σ. The experiments were conducted based on a three-factor, three-level Box–Behnken surface statistical design. Quadratic polynomial equations were developed for proper process parametric study for its optimal performance characteristics. The experimental results under optimum conditions were compared with the simulated values obtained from the RSM and ANN model. Adequacy of the developed models was tested by analysis of variance technique. A multilayer feed-forward neural network with a Levenberg–Marquardt back-propagation algorithm was adopted to develop the relationships between the laser hardening process parameters, HI, and ultimate σ. The performance of the developed ANN models were compared with the second-order RSM mathematical models of HI and σ. There was good agreement between the experimental and simulated values of RSM and ANN. The comparison clearly indicates that the ANN models provide more accurate prediction compared to the RSM models. It has been found that there is a trend of increased tensile strength with the decrease of hardening heat input and a trend of increased tensile strength with the increase of hardening cooling rate. As heat input decreases, there will be a faster cooling rate. Considering the effect of HI on ultimate σ, it was found that the lower the heat input, the faster cooling rate. The details of experimentation, model development, testing, validation of models, effect of laser process variables on heat input and ultimate σ, effect of HI on σ, and performance comparison of RSM and ANN models are presented in the paper. The results of Box–Behnken design of RSM and ANN models also indicate that the proposed models predict the responses adequately within the limits of input parameters being used. It is suggested that regression equations can be used to find optimum conditions for HI and σ of laser-hardened commercially pure titanium material.     

传感器动态补偿的神经网络逆系统方法

给出一种基于神经网络逆系统的传感器动态补偿策略。无需传感器具体模型和参数，即可实现传感器系统的近似单位线性化，达到动态补偿的目的。仿真实验和动态标定试验结果表明，应用这种新型的易于工程实现的动态补偿方法可显著地提高传感器的动态特性，有效改善传感器的动态品质。     

Nd:YAG激光焊接殷钢薄板材料的工艺研究

利用Nd:YAG脉冲激光作为焊接热源,对殷钢材料Invar36分别进行了平板单道焊接试验和对焊试验,分析了工艺参数（激光功率、焊接速度、脉冲宽度和离焦量）变化对焊缝的表面形貌、熔宽以及熔透性的影响。对0.85mm厚度的殷钢薄板对焊接头的硬度和成分的变化以及拉伸强度进行了检测。结果表明：激光功率和脉宽是影响焊缝熔深、熔宽和热影响区大小的主要因素;扫描速度对焊缝表面的鱼鳞状条纹间距影响尤为明显;离焦量主要影响焊缝的宽度和熔透性;合理匹配工艺参数能够实现0.85mm厚度薄板对焊,并且获得形貌良好的焊缝。焊缝的组织成分没有发生明显变化,拉伸强度和基体的相当,显微硬度略低于基体硬度。     

Build orientation determination of multi-feature mechanical parts in selective laser melting via multi-objective decision making

Selective laser melting (SLM) is a unique additive manufacturing (AM) category that can be used to manufacture mechanical parts. It has been widely used in aerospace and automotive using metal or alloy powder. The build orientation is crucial in AM because it affects the as-built part, including its part accuracy, surface roughness, support structure, and build time and cost. A mechanical part is usually composed of multiple surface features. The surface features carry the production and design knowledge, which can be utilized in SLM fabrication. This study proposes a method to determine the build orientation of multi-feature mechanical parts (MFMPs) in SLM. First, the surface features of an MFMP are recognized and grouped for formulating the particular optimization objectives. Second, the estimation models of involved optimization objectives are established, and a set of alternative build orientations (ABOs) is further obtained by many-objective optimization. Lastly, a multi-objective decision making method integrated by the technique for order of preference by similarity to the ideal solution and cosine similarity measure is presented to select an optimal build orientation from those ABOs. The weights of the feature groups and considered objectives are achieved by a fuzzy analytical hierarchy process. Two case studies are reported to validate the proposed method with numerical results, and the effectiveness comparison is presented. Physical manufacturing is conducted to prove the performance of the proposed method. The measured average sampling surface roughness of the most crucial feature of the bracket in the original orientation and the orientations obtained by the weighted sum model and the proposed method are 15.82, 10.84, and 10.62 μm, respectively. The numerical and physical validation results demonstrate that the proposed method is desirable to determine the build orientations of MFMPs with competitive results in SLM.     

Applying ANN to predict the forming load and mechanical property of magnesium alloy under hot extrusion

This study establishes the database concerning magnesium alloy hot extrusion, and uses it to conduct various investigations. Firstly, artificial neural networks (ANN) analysis is used to determine the die shapes of various extrusion ratios. Secondly, the process parameters for the hot extrusion of magnesium alloy are determined, and thirdly, the tensile strength and maximum extrusion load of the finished product are predicted. The database includes 11 parameters, associated with 108 sets of experiment, determined by material type (AZ31 and AZ61), extrusion ratio (14.41, 35.9 and 55.85), product shape (tubular and sheet), semicone angle of the die (90° and 30°), extrusion speed, temperature to which the billet is heated, temperature to which the container is heated, lubricant, hold-time at a specified temperature, extrusion load and tensile strength. ANN is applied to learn from this database, and backward propagation analysis is conducted to find the mechanical properties of the products under various extrusion ratios. This study adopts the orthogonal array of the Taguchi method to hot extrusion experiments that involve dies with different extrusion ratios, and sets the tensile strength and extrusion load of the finished product as the quality characteristics, to acquire the optimal parameter condition. Then, based on the results obtained from the additive model, confirmatory experiments are performed. An Analysis of Variance (ANOVA) analysis is then performed to investigate and analyze the influence of factors on the hot extrusion process. The weight of important factors in the database is increased, and subsequently, the forming load and mechanical properties of magnesium alloy under extrusion are accurately predicted.     

Classification-driven model selection approach of genetic programming in modelling of turning process

Turning is a widely used machining process, but the process complexity and uncertainty lead to empirical modelling techniques being preferred over physics-based models for predicting the process performance. The literature reveals that empirical methods such as artificial neural networks (ANN), support vector regression (SVR), regression analysis and fuzzy logic have been extensively applied in the modelling of turning process. The present work introduces genetic programming (GP) for the modelling of turning, but it is observed that the optimal models selected from the GP population based on training and validation errors do not perform well on testing data (unseen samples). Selecting the best GP model from the population of models is therefore a vital step. In view of this, the classification-driven model selection approach of GP (C-GP) is proposed in this paper. In this methodology, potential classification techniques such as Bayes multinomial, partitioning and regression trees, classification and regression trees and decision trees are integrated with GP to predict the class (best or bad) of the GP models. The model that is classified as the “best” by the most number of classification techniques is selected, and its performance is compared to those from ANN and SVR. It is found that the C-GP model has accuracy on par with ANN and gives satisfactory performance on testing data.     

Slag recycling in submerged arc welding and its influence on weld quality leading to parametric optimization

Slag generated during conventional submerged arc welding (SAW) has been recycled by mixing varying percentages of crushed slag with fresh flux to use in subsequent runs. The influence of using flux-slag mixture on various aspects of SAW weld parameters of bead geometry have been investigated in a quantitative basis. Slag has been reprocessed and reused in submerged arc welding to produce bead-on-plate weld on mild steel plates. Apart from conventional process parameters: voltage (OCV), wire feed rate, nozzle to plate distance (stick-out) and traverse speed, welding has been carried out using various percentages of flux-slag mixture; the % of fused flux in the mixture has been treated as a process parameter. Various bead geometry parameters viz. bead width; reinforcement, depth of penetration and depth of HAZ have been measured for each of weld prepared in the study. Using experimental data, a grey-based Taguchi approach has been applied for parametric optimization of this non-conventional SAW process. The aim was to reveal the optimal amount of slag-mix%, which could be applied in SAW process without imposing any adverse effect on features of bead geometry and HAZ. Optimal result has been checked through confirmatory test.     

Estimation of discharge with free overfall in rectangular channel using artificial intelligence models

In open channels, free overfall can be used as a flow-measuring device in various flow regimes for different shapes of channels. The aim of this paper is to accurately estimate the discharge in rectangular channels by applying four soft computing models [i.e., artificial neural network (ANN), gene expression programming (GEP), multivariate adaptive regression spline (MARS) and M5 tree model]. The variables including brink depth, roughness coefficient, channel width and channel bed slope taken from earlier published works were used as inputs for the models. Data were divided according to three different splitting scenarios, 50%–50%, 60%–40% and 75%–25%, for training-testing phases to obtain more robust evaluation of the models. The computed discharges were then compared with experimental results. In order to evaluate the accuracy of discharge estimations by ANN, GEP, MARS and M5 tree models, six statistical measures including correlation coefficient (R), mean absolute error (MAE), root mean square error (RMSE), Nash-Sutcliffe (NS), Willmott index (WI) and Legates and McCabe index (LMI) have been applied. For different training-testing scenarios, the performance of the ANN model is better than the other methods. Regarding 50%–50% training-testing scenario, ANN has the best accuracy with R, RMSE, MAE, NS, WI and LMI of 0.994, 0.004 m³/s/m, 0.002 m³/s/m, 0.986, 0.999 and 0.909 by considering all inputs.     

冲压件成形计算机模拟工艺参数优化方法研究

分析了常规有限元金属板料成形模拟的不足，提出了参数化有限元分析的概念，在对人工神经网络、遗传算法进行深入分析研究的基础上，采用参数化有限元分析方法进行分析，得到了训练样本。提出了采用人工神经网络技术建立冲压件成形多参数映射关系模型，采用遗传算法进行多参数组合优化，实现冲压件成形计算机模拟工艺参数优化的方法。实际应用结果表明，优化结果与试验结果基本吻合，该优化方案实用可行。