径向基神经网络在三维激光切割中的应用

分析了RBF神经网络的特点，并研究了它在三维激光切割中的应用，利用该神经网络及其变形可以预测切缝的宽度，也可以根据输入参数对试样的挂渣与否进行判定。比较了三雏激光切割中切割倾角和切割方向对切割质量的影响，从而实现切割质量的预测和优化工艺参数的选取以指导实际的切割过程。     

基于神经网络建模和遗传算法的线切割工艺参数模糊优化

针对电火花线切割加工工艺难以获得更优异的工艺参数（如同时得到更快的切割速度和更好的表面粗糙度）等问题，先以神经网络建立模型，然后结合遗传算法和模糊理论求解此多目标优化问题，得到优化的工艺参数。结果显示。取得了较好的效果。     

不锈钢板激光切割速度预测方法研究

基于激光与材料相互作用的原理，分析了激光切割不锈钢板过程中的热量传导规律，建立了切割热量模型和热平衡方程，即材料所吸收的总热量等于熔化切缝材料吸收的热量与热传导损失的热量之和，推导出切割速度与其他工艺参数的函数关系。通过此关系，能够对切割速度进行预测。试验结果表明，预测得到的切割速度与试验得到的最佳切割速度十分吻合，切割速度的预测方法可以指导实际加工。     

碳钢板激光切割工艺分析及优化

随着科技的不断发展,激光切割技术已经成为现代工业生产中不可或缺的技术之一,碳钢激光切割件加工技术也得到越来越多的应用。随着对加工工件的质量要求越来越高,对激光加工工艺进行优化以获得更高切割质量与切割精度的钢板具有重要意义。为了优化碳钢的激光加工工艺,文章以3 mm碳钢为对象,采用1 500 W激光器系统,研究了激光功率、切割速度、离焦量和辅助气体压力等切割参数对切割质量的影响。选取切缝宽度、表面粗糙度、挂渣高度作为评价切割质量的标准,综合分析工艺参数对切割质量的影响规律并研究了影响机理,同时通过卷积神经网络对数据进行预测分析,并采用带精英策略的非支配排序的遗传算法对工艺参数进行优化,获得辅助气体为氧气条件下激光切割碳钢的最佳工艺参数为P=1 500 W、v=64 mm/s、F=2.6 bar、h=2.9 mm。通过实验发现优化后的切件切割面光滑,无挂渣现象且切缝宽度较小,达到预期标准。     

机械手逆运动学神经网络算法研究

提出一种基于模糊遗传算法的机械手逆运动学神经网络建模方法。该方法采用3层前向神经网络建立机械手逆运动学模型，应用模糊遗传算法训练神经网络的权系数。此算法可根据种群进化情况，对交叉概率和变 异概率进行在线模糊控制，加快了算法的搜索过程，有效地避免了简单遗传算法中容易出现的初期收敛问题。仿真结果表明，本方法提高了求解精度和收敛速度，不但有效克服了简单遗传算法常出现的初期收敛和BP算法求解精度低、容易陷入局部极小等缺点，而且避免了计算Jacobian矩阵的伪逆，结构简单、容易实现。     

基于模拟退火的BP神经网络激光切割质量控制模型

激光切割金属板材过程中,工艺条件和表面质量之间存在较为复杂的对应关系;提出基于BP神经网络的激光切割质量控制模型,建立工艺条件与切割面粗糙度之间的关系模型,试验测量点取距切割下边缘1.5mm处表面粗糙度Ra;提出利用模拟退火算法提高多层神经网络的拟合精度,改善网络的收敛性能;切割试验样本设计拟采用星点设计法,用于提高神经网络训练样本的信息量和可靠性;经过实际切割不锈钢板材,验证上述方法具有一定的可靠性和应用价值。     

激光切割辅助气体参数对切割质量影响的仿真分析

对激光切割过程中辅助气体在切缝内的速度进行了仿真分析,通过辅助气体在切缝内切割前沿近壁面上的速度云图,分析了辅助气体种类以及压强、离轴量等参数对切割前沿近壁面上气体速度的影响,而切缝内气体速度将直接影响激光切割质量。仿真结果表明:对于2 mm石蜡板,激光功率为5 W、切割速度为0.04 m/s时可采用2 MPa压强,氩气离轴式激光切割的方式能得到较优的切割质量。     

遗传算法改进的BP神经网络在协同创新评价中的应用

为了解决客户协同创新中协同工作效率难于评价的问题,提出了一种用遗传算法优化的神经网络对客户协同产品创新进行评价的评价模型：在评价指标方面,设计了一套包括效益、效率和过程的18个指标的评价体系;在评价算法方面,将遗传算法与BP神经网络结合起来,设计了遗传算法改进的BP神经网络算法。该模型充分利用遗传算法的全局搜索能力强与神经网络的局部搜索能力强的特点,克服了遗传算法局部收敛与神经网络收敛速度较慢的问题,是一种非常适用于评价协同工作的模型。最后通过实例训练,证明了该模型的有效性与可行性。     

煤气鼓风机故障模式识别方法研究

为了克服BP神经网络收敛速度慢且容易陷入局部极小问题,提出了将遗传算法与神经网络相结合用于故障诊断的方法。综合遗传算法的全局搜索性和神经网络的并行快速性等特点,通过使用遗传算法优化神经网络的初始权值达到加快其收敛速度的目的。实例结果表明用遗传算法改进后的神经网络在故障模式识别中是可行有效的。     

叠层实体成型中提高激光切割精度的动力学补偿方法

动力学因素,如摩擦力和惯性力,在叠层实体成型加工中常影响激光机床的切割精度。提出的动力学补偿方法结合了闭环位置控制和计算力矩控制两者的优点。既可以避免对名义轨迹的偏差,又可以补偿动力学因素对精度的影响。在每个原动件控制中,用附加的速度前馈来实现动力学补偿。多层前馈型神经网络用来实现机构的逆动力学模型。用周期函数,有限项傅立叶级数,作为激励函数来获取训练样本。复杂的动力学参数辨识过程成为神经网络权值的监督学习过程。实验结果表明,本文提出的方法对提高激光切割的轨迹精度和切口角度精度是有效的。     

S形件翻边成形工艺参数优化设计

建立了S形件成形的参数化有限元模型并进行参数化有限元分析(PFEA).结合人工神经网络(ANN)和遗传算法(GA)进行优化设计,研究了S形件翻边成形中的工艺参数优化.然后利用优化的结果值进行验算,获得了良好的成形件,表明基于PFEA、ANN和GA进行成形工艺参数优化是可行的.     

Parameter optimization of non-vertical laser cutting

In some cases, in order to avoid interference during 3D laser cutting of thin metal a laser head could not be kept vertical to the surface of a work piece. In such situations, the cutting quality depends not only on “typical” cutting parameters but also on the slant angle of the laser head. Traditionally, many tests had to be done in order to obtain the best cutting results. In this paper, an experimental design is employed to reduce the number of tests and an artificial neural network (ANN) is set up to describe quantitatively the relationship between cutting quality and cutting parameters in the non-vertical laser cutting situation. A quality point system is used to evaluate the cutting result of the thin sheet quantitatively. Testing of this novel method shows that the calculated “quality point” using ANN is quite closely in accord with the actual cutting result. The ANN is very successful for optimizing parameters, predicting cutting results and deducing new cutting information.     

OPTIMAL MQL AND CUTTING CONDITIONS DETERMINATION FOR DESIRED SURFACE ROUGHNESS IN TURNING OF BRASS USING GENETIC ALGORITHMS

The evolving concept of minimum quantity of lubrication (MQL) in machining is considered as one of the solutions to reduce the amount of lubricant to address the environmental, economical and ecological issues. This paper investigates the influence of cutting speed, feed rate and different amount of MQL on machining performance during turning of brass using K10 cemented carbide tool. The experiments have been planned as per Taguchi's orthogonal array and the second order surface roughness model in terms of machining parameters was developed using response surface methodology (RSM). The parametric analysis has been carried out to analyze the interaction effects of process parameters on surface roughness. The optimization is then carried out with genetic algorithms (GA) using surface roughness model for the selection of optimal MQL and cutting conditions. The GA program gives the minimum values of surface roughness and the corresponding optimal machining parameters.     

基于SVR-GA算法的广义加工空间机床切削稳定性预测与优化研究

针对机床零件加工位置和进给方向不确定造成刀尖频响函数变化,导致切削稳定性叶瓣图与无颤振工艺参数预测具有不确定性问题,提出一种耦合支持向量回归机(SVR)与遗传算法(GA)的切削稳定性预测与优化方法。该方法采用锤击法模态实验和空间坐标变换,获取样本空间不同加工位置与进给方向的刀尖频响函数;进而结合传统切削稳定性预测方法构建以各向运动部件位移、进给角度、主轴转速、切削宽度、每齿进给量为输入的极限切削深度SVR预测模型;采用该SVR模型作为切削稳定性约束建立材料切除率优化模型,通过遗传算法求解各运动轴位移、进给角度与切削参数的最优配置。以某型加工中心展开实例研究,实验结果表明获取的优化配置能实现稳定切削,验证了该方法的有效性。     

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.     

激光加热辅助车削淬硬钢的白层形成临界切削速度预测与实验研究

淬硬钢已加工表面极易产生白层,对零件的服役性能造成很大影响,采用激光加热辅助车削替代常规车削来加工淬硬钢可解决白层形成的问题.利用VDFLUX用户子程序建立了热力耦合移动激光加热辅助车削过程有限元模型并进行仿真.基于有限元模型仿真结果和白层形成临界切削速度预测模型,分析了不同激光预热参数对白层形成临界切削速度的影响.结...     

基于遗传算法的采煤机螺旋滚筒多目标优化设计

为实现采煤机螺旋滚筒截煤和装煤时综合性能最优,基于虚拟样机技术和离散元理论,得到了滚筒各截割性能指标与装煤率随不同结构及运动参数的变化规律,依据机械优化设计理论建立了各性能指标的评价函数。选取螺旋升角、截线距、转速和牵引速度为设计变量,建立了以不同性能指标为分目标的多目标优化模型,利用遗传算法求解得到了最优的结构参数和运动参数。结果表明,利用遗传算法优化后滚筒的最大切削面积增大247mm2,截割比能耗减小0.014kW·h/m3,截割功率减小10.8kW,截割阻力减小7085kN,装煤率提高1.7%,有效地提升了滚筒的综合性能。研究结果为滚筒结构参数和运动参数的选取提供了数据支撑,具有一定的工程应用价值。     

Artificial neural network based on genetic learning for machining of polyetheretherketone composite materials

In this paper an artificial neural network (ANN) aiming for the efficient modelling of a set of machining conditions for orthogonal cutting of polyetheretherketone (PEEK) composite materials is presented. The supervised learning of the ANN is based on a genetic algorithm (GA) supported by an elitist strategy. Input, hidden and output layers model the topology of the ANN. The weights of the synapses and the biases for hidden and output nodes are used as design variables in the ANN learning process. Considering a set of experimental data, the mean relative error between experimental and numerical results is used to monitor the learning process obtaining the completeness of the machining process modelling. Also a regularization term associated to biases in hidden and output neurons are included in the GA fitness function for learning. Using a different set of experimental results, the optimal ANN obtained after learning is tested. The optimal number of nodes on the hidden layer is searched and the positive influence of the regularization term is demonstrated. This approach of ANN learning based on GA presents low mean relative errors in learning and testing phases.     

Prediction and analysis of surface roughness characteristics of a non-ferrous material using ANN in CNC turning

Surface roughness, an indicator of surface quality is one of the most-specified customer requirements in a machining process. For efficient use of machine tools, optimum cutting parameters (speed, feed, and depth of cut) are required. So it is necessary to find a suitable optimization method which can find optimum values of cutting parameters for minimizing surface roughness. The turning process parameter optimization is highly constrained and non-linear. In this work, machining process has been carried out on brass C26000 material in dry cutting condition in a CNC turning machine and surface roughness has been measured using surface roughness tester. To predict the surface roughness, an artificial neural network (ANN) model has been designed through feed-forward back-propagation network using Matlab (2009a) software for the data obtained. Comparison of the experimental data and ANN results show that there is no significant difference and ANN has been used confidently. The results obtained conclude that ANN is reliable and accurate for predicting the values. The actual R _a value has been obtained as 1.1999???m and the corresponding predicted surface roughness value is 1.1859???m, which implies greater accuracy.