熔敷金属力学性能人工神经网络模型的研究

在试验的基础上,采用人工神经网络方法建立了基于BP算法的熔敷金属力学性能的预测模型,该模型训练结果与试验值之间有很好的对应关系,说明该模型能准确反映合金元素与熔敷金属力学性能之间复杂的非线性关系。用该模型研究了合金元素对熔敷金属低温韧性的影响,分析得出了与试验一致的结果。     

A hybrid model using genetic algorithm and neural network for process parameters optimization in NC camshaft grinding

Camshaft grinding is more complex comparing with the ordinary cylindrical grinding. Since its quality is mostly influenced by more factors, how to select process parameters quickly and accurately becomes the key to improve its quality and processing efficiency. In this paper, a hybrid artificial neural network (ANN) and genetic algorithm (GA) model is proposed to optimize the process parameters. In this method, a BP neural network model is developed to map the complex nonlinear relationship between process parameters and processing requirements, and a GA is used in order to improve the accuracy and speed based on the ANN model. The results show that the hybrid ANN/GA model is an effective tool for the process parameters optimization in NC camshaft grinding.     

Recent development in finite element analysis of clinched joints

Clinching is a high-speed mechanical fastening technique for point joining of sheet materials. Published work relating to finite element analysis of clinched joints is reviewed in this paper, in terms of process, strength, and vibration characteristics of the clinched joints. It is concluded that the finite element analysis of clinched joints will help future applications of clinching by allowing system parameters to be selected to give as large a process window as possible for successful joint manufacture. This will allow many tests to be simulated that would currently take too long to perform or be prohibitively expensive in practice.     

Prediction of abrasive wear rate of in situ Cu–Al2O3 nanocomposite using artificial neural networks

In this work, artificial neural networks (ANNs) technique was used in the prediction of abrasive wear rate of Cu–Al₂O₃ nanocomposite materials. The abrasive wear rates obtained from series of wear tests were used in the formation of the data sets of the ANN. The inputs to the network are load, sliding speed, and alumina volume fraction. Correlation coefficients between the experimental data and outputs from the ANN confirmed the feasibility of the ANNs for effectively model and predict the abrasive wear rate. The comparison between the ANNs and the multi-variable regression analysis results showed that using ANNs technique is more effective than multi-variable regression analysis for the prediction of abrasive wear rate of Cu–Al₂O₃ nanocomposite materials. Optimization of the training process of the ANN using genetic algorithm (GA) is performed and the results are compared with the ANN trained without GA. Sensitivity analysis is also done to find the relative influence of factors on the wear rate. It is observed that load and alumina volume fraction effectively influence the wear rate.     

基于GA-BP的汽车风振噪声声品质预测模型

目前对于汽车风振噪声的优化研究主要以声压级（Sound pressure level,SPL）作为单一评价指标,既不能全面反映噪声的物理属性,也无法考虑人耳对噪声的主观认知过程。为准确评价风振噪声,引入声品质,运用大涡模拟（Large eddy simulation,LES）对风振噪声进行数值仿真,根据实车道路试验判断仿真的准确性;对仿真结果进行声品质客观评价与主观评价,综合声品质客观评价参数与声品质主观评价试验结果建立BP神经网络预测模型;利用遗传算法（Genetic algorithm,GA）,进一步对BP神经网络的结构参数进行优化,建立GA-BP声品质预测模型。研究结果表明,GA-BP声品质预测模型在训练速度和预测精度上都优于BP神经网络预测模型。预测模型基于声品质主客观评价结果,其预测值可以代替传统的声压级评价指标,为风振噪声提供更为准确合理的评价。     

Parameter optimization of friction stir welding of cryorolled AA2219 alloy using artificial neural network modeling with genetic algorithm

In this paper, parameter optimization of FSW of cryorolled AA2219 alloy was carried out to obtain defect free weld joint with maximum weld strength. To achieve this, artificial neural network (ANN) was used to model the relationship between the input parameters and the mechanical and corrosion properties (output) of the weld joints. The optimal FSW parameters were determined by genetic algorithm (GA). The feasible solution of the GA was tool rotational speed of 1005 rpm, tool travel speed of 20 mm/min and tool tilt angle of 3°. The feasible parameter was used to weld and check the ability of the parameter to produce better weld joint than the L₉ orthogonal array parameters. The weld, subjected to the confirmation test, was investigated by means of metallurgical, mechanical, and corrosion testing. This process reduces the costs associated with trial runs to obtain optimal parameters and also the production cost of the cryorolled (CR) plate which is high.     

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.     

基于BP神经网络的TC4钛合金超塑性变形后组织及性能预测研究

分别用Visual Fortran语言和MATLAB软件建立了TC4钛合金超塑性变形时变形参数与其力学性能和晶粒尺寸之间的BP神经N络模型，通过用较少的力学性能和晶粒尺寸的试验数据进行训练，进而对其性能进行预测。结果表明，BP神经网络用于材料超塑性变形后的力学性能及晶粒尺寸预测是可行的，其预测误差小于7%。     

Influence of process parameters in mechanical clinching with extensible dies

The influence of clinching tool design in joining metal sheets by the clinching process with extensible dies is investigated. The material flow during the clinching process was examined experimentally and numerically. The geometrical and mechanical characteristics of joints produced under different processing conditions, that is, forming loads, were used to calibrate and validate a 3D finite element model of the clinching process. Then, the model was utilized to evaluate the influence of clinching tool design parameters, namely the punch diameter, the punch corner radius, the fixed die depth, the fixed die diameter, and the die corner radius. The effects of design parameters on the cross section of a clinched joint, the required forming load and the joint strength were analysed and the appropriate processing window was determined. According to the achieved results, the main benefits and drawbacks of each configuration are discussed.     

人工神经网络在除湿机故障诊断中的应用

本文介绍了人工神经网络中使用最广泛的BP神经网络,通过除湿机的特征参数与故障原因之间的关系,建立了神经网络模型,并用Matlab神经网络工具箱函数对网络进行了训练,结果达到要求,能较好地识别故障.     

Precise volume fraction prediction in oil–water–gas multiphase flows by means of gamma-ray attenuation and artificial neural networks using one detector

Artificial neural network (ANN) is an appropriate method used to handle the modeling, prediction and classification problems. In this study, based on nuclear technique in annular multiphase regime using only one detector and a dual energy gamma-ray source, a proposed ANN architecture is used to predict the oil, water and air percentage, precisely. A multi-layer perceptron (MLP) neural network is used to develop the ANN model in MATLAB 7.0.4 software. In this work, number of detectors and ANN input features were reduced to one and two, respectively. The input parameters of ANN are first and second full energy peaks of the detector output signal, and the outputs are oil and water percentage. The obtained results show that the proposed ANN model has achieved good agreement with the simulation data with a negligible error between the estimated and simulated values. Defined MAE% error was obtained less than 1%.     

基于混合遗传算法的磁控形状记忆合金驱动器磁滞模型优化

为了消除或减小磁滞非线性特性对磁控形状记忆合金驱动器定位精度的影响,应用BP神经网络建立了磁控形状记忆合金驱动器磁滞模型。针对BP网络算法存在的不足,以及网络结构、初始连接权值和阈值的选择对BP网络训练的影响很大等问题,提出一种混合遗传算法对神经网络磁滞模型的权值和阈值进行优化。将优化后的参数赋值给BP神经网络重新训练,结果表明,优化后的磁滞模型训练误差绝对值由25nm减小到5nm,有较好的收敛性。     

Prediction of contraction scour using ANN and GA

This paper emphasizes on the application of soft computing tools such as artificial neural network (ANN) and genetic algorithm (GA) in the prediction of scour depth within channel contractions. The experimental data of earlier investigators are used in developing the models and ANN and GA Toolboxes of MATLAB software are utilized for the purpose. The multilayered perceptron (MLP) neural networks with feed-forward back-propagation training algorithms were designed to predict the scour depth. The mean squared error and correlation coefficient are used to check the performance of networks. It is found that the ANN architecture 4-16-1 having trained with Levenberg-Marquardt ‘trainlm’ function had best performance having mean squared error of 0.001 and correlation coefficient of 0.998. In addition, the suitability of ‘trainlm’ method over other training methods is also discussed. The scour depths predicted by ANN model were compared with those computed by the two analytical models (with and without sidewall correction for contracted zone) and an empirical model proposed by Dey and Raikar [1]. In addition, heuristic search technique called genetic algorithm is used to develop the predictor for maximum scour depth within channel contraction. The population size for GA was 500 members with total generations of 1000, crossover fraction of 0.8 and Gaussian operator for mutation. It is promising to observe that the GA model predicts the maximum scour depth equally well as that of empirical model of Dey and Raikar [1]. Hence, both ANN and GA models can be satisfactorily used to predict the scour depth within channel contractions.     

基于遗传算法的人工神经网络在航空发动机磨损预测中的应用

提出了一种将遗传算法和BP算法相结合的学习算法来训练BP神经网络，实现网络结构的优化，并用优化后的BP人工神经网络建立了航空发动机的磨损预测模型。将该模型预测结果与BP算法和多元线性回归法的预测结果进行了比较。检验结果表明：基于遗传算法的BP神经网络优于BP算法及多元线性回归法，具有良好的预测效果。     

基于气体传感器阵列的混合气体定量分析

优选CO和H2气体敏感的半导体气体传感器组成阵列，建立实时数据采集系统，结合BP神经网络模式识别技术，实现了混合气体组分的定量分析。讨论了不同响应时间下的阵列输出值、不同的数据预处理算法及不同的神经网络结构等主要影响因素对网络输出结果的影响。结果表明，采用RRD预处理算法对3min响应时间下的阵列输出值进行处理，再输入到有12个隐层神经元数的3层BP神经网络进行训练，预测的效果最好。该处理模式能较准确地完成CO和H2混合气体组分的定量分析。     

圆柱面过盈连接的力学特性及设计方法

传统的圆柱面过盈连接设计将结合直径、结合宽度等因素视为定量,并且忽略边缘效应所引起的应力集中。为解决此问题,更好地研究圆柱面过盈连接的力学特性,寻求更合理的设计方法,结合有限元法和BP神经网络各自的优势,以ABAQUS为工具分析圆柱面过盈连接接触面的应力特性及结合直径、结合宽度、包容件外径及过盈量等因素对它的影响,将通过分析得到的大量接触边缘最大应力作为神经网络的训练样本,建立接触边缘最大应力的BP神经网络模型。将接触面应力的各种影响因素视为可调变量,结合接触边缘最大应力的BP神经网络模型,提出一种以接触边缘最大应力为优化目标的圆柱面过盈连接设计的BP神经网络动态调整算法。通过实例分析,表明该算法比传统的设计方法更为合理。     

拉深孔成形的工艺参数优化与实验研究

为了探讨拉深孔成形技术对提高板料成形性能的有效性，以圆筒形件为研究对象，研究了拉深孔成形条件下的成形性能。采用数值模拟、人工神经网络和遗传算法进行板材成形工艺参数优化，得到了最优化的压边力和拉深孔相对密度等拉深工艺参数。根据优化后的结果设计并完成了相关的工艺实验，取得了与有限元模拟相一致的结果，证明了拉深孔成形技术是一种提高板材成形性能的行之有效的方法。     

Joining of aluminium and copper materials with friction welding

Radial basis network (RBN), a special type of artificial neural networks (ANN), is introduced to the field of machining process modeling and simulation. This feed-forward three-layer fully interconnected neural network is successfully used to establish the relationship between the machining conditions (inputs) and process parameters (outputs) for the case of ball end milling. A set of four key input parameters is selected to represent the cutting conditions, while four important characteristics of the instantaneous cutting force are used as the output set. Experiments are conducted to train as well as to validate and assess the performance of the proposed network. In addition, a case study, consisting of a typical machining scenario found in industry, is performed to test and verify the model. A very good agreement is observed between the forces predicted by the new model and their experimental counterparts, thus validating the new approach.     

Experimental and Prediction of Abrasive Wear Behavior of Sintered Cu-SiC Composites Containing Graphite by Using Artificial Neural Networks

Experimental investigations were undertaken to determine the abrasive wear behavior of various percentages of Cu-SiC-Gr hybrid composites. Wear tests were carried out using a pin-on-disc type machine using various input parameters like load, sliding distance, and sliding velocity with various SiC abrasive papers of grit size 80, 220, and 400, having an average particle size of 192, 102, and 45 μm. Neural networks are employed to study the tribological behavior of sintered Cu-SiC-Gr hybrid composites. Optical microscope, scanning electron microscope (SEM), X-ray diffraction (XRD), and energy-dispersive spectral observations are used to evaluate the characteristics. The proposed neural network model used the measured parameters, namely, the weight percentage of graphite, abrasive size, sliding speed, load, and sliding distance, to predict the wear loss of the composite. In order to improve the accuracy and obtain better results, an artificial neural network (ANN) with a genetic algorithm (GA) function was used. Optimization of the training process of the ANN using a GA is performed and the results are compared with the ANN trained without a GA. The predicted values from the proposed networks coincide with the experimental values.     

GA-BP网络在钢轨磨损量预测中的应用

基于GA算法及BP神经网络优点,将GA算法优化的BP网络应用于钢轨磨损量预测。通过赫兹模拟试验方法,在JD-1轮轨模拟试验机上获得不同曲线半径、轴重、运行速度工况下对应的钢轨磨损量,即神经网络所需样本,建立GA算法优化三层BP神经网络,对钢轨磨损量进行预测。结果表明:GA算法优化的BP神经网络对钢轨磨损量具有良好的预测性能,较好地反映了曲线半径、轴重、运行速度对钢轨磨损量的影响规律。