铣削力预测方法和影响因素综述

为减少航空发动机薄壁件铣削加工过程中的加工变形,提高加工质量,需对铣削加工过程中的切削力进行预测。因此,综述了多远回归分析预测模型、微元铣削力预测模型、有限元预测模型和人工神经网络预测模型,并对切削用量、刀具几何参数、工件材料、冷却作用、刀具材料和刀具磨损对铣削力的影响进行了分析。     

粒子群优化人工神经网络在高速铣削力建模中的应用

将粒子群优化人工神经网络理论应用于高速铣削力的建模研究中.采用粒子群算法与反向传播算法相结合的方法,对反向传播神经网络模型进行优化.用粒子群算法训练网络参数,直到误差趋于一稳定值,然后用优化的权值进行反向传播算法运算,以实现高速铣削力的预测.充分发挥了粒子群算法的全局寻优能力和反向传播算法的局部搜索优势.仿真结果表明,与其他几种反向传播算法相比较,粒子群算法与反向传播算法的学习算法训练的神经网络,不仅训练时间明显缩短,而且其预报精度也得到了较大的提高,能够有效地建立铣削力模型,并对铣削力进行准确的预测.     

注塑模具硬态高速铣削力研究

采用高速铣床对4Cr5MoSiV1钢注塑成型模具进行硬态铣削,研究切削加工参数对切削力的影响,通过多因素法正交试验,利用改进的BP神经网络建立了切削力的神经网络模型,将网络预测结果经过现场加工实践检验其准确性,利用MATLAB分析切削参数的影响。结果表明:人工神经网络能准确地预测铣削力的大小,模型具有良好的泛化能力和自适应能力;在高转速、小切深、合适的进给速度以及微量切削液状态下铣削力较小,为优化模具硬态铣削的切削参数并对其实际生产应用提供了较好的依据。     

基于改进粒子群算法的刀具瞬时铣削力预测研究

针对刀具微元铣削力模型预测精度低的问题,建立了基于刀具径向跳动的瞬时铣削力模型,采用改进惯性权重的粒子群算法(IWPSO)对模型进行系数求解。通过改进粒子群算法,避免算法过早收敛而陷入局部最优,提高了算法的速度和精度,降低了模型系数的求解难度,从而减小模型预测铣削力的误差。通过与线性拟合方法求解的铣削力系数对比,在0～0.1s内铣削力预测波形图的波谷与实际铣削力波形图误差较小(5%以内),验证了此模型的精度更高。采用不同铣削参数进行实验,验证了铣削力预测模型预测铣削力的准确性,对实际的铣削加工有着重要意义。     

基于BP神经网络的球头铣刀铣削力建模与仿真

将BP神经网络的理论和算法应用于球头刀具铣削力建模的研究中.采用LM算法建立了铣削力预测的神经网络模型,模型中考虑了影响铣削力的加工参数,选取铣削力试验数据对神经网络模型进行训练,用训练好的神经网络模型对铣削力进行仿真.仿真结果表明,用BP神经网络方法建立的铣削力模型能够对铣削力进行准确的预测.     

高速加工热缩加长刀杆与刀具配合铣削力预测

根据热缩加长刀杆与刀具配合精加工与半精加工的特点,利用反向传播神经网络(BPNN)建立高速加工热缩加长刀杆与刀具配合的铣削力模型。模型除了考虑6个主要影响铣削力的加工条件外,还将时间参量引入输入向量,实现了三向铣削力的瞬态预测。通过大量的加工实验获得网络所需的训练和检验样本,并通过编制Matlab程序实现了网络性能评价和网络参数优化。检验结果表明,铣削力预测结果与实际测量结果之间具有很好的一致性,三向分力的平均预测误差均小于0.18,在预测效率和精度上均优于通常所用的解析模型,并具有很好的扩展性能。     

可转位面铣刀铣削力解析建模与实验研究

以灰铸铁端面铣削为研究对象,建立了可转位面铣刀铣削力解析模型。进行灰铸铁HT250铣削实验,采用快速标定斜角铣削力系数方法进行铣削力系数辨识。结果表明,在选定的工艺参数范围内,建立的铣削力模型具有较高的预测精度。基于该铣削力模型,还研究了铣削参数对铣削力的影响规律。     

基于主轴和进给轴电流的铣削力间接预测方法

提出了一种基于主轴和进给轴电流最优变权法的瞬时铣削力预测方法。首先,分析了主轴电流与x向瞬时铣削力的映射关系,基于互相关方法考虑了电流信号的延迟效应;其次,基于Devavit Hartenberg法对五轴机床进行运动学建模,将进给轴驱动力矩从机床坐标系映射到刀具坐标系,基于力雅可比矩阵得到进给轴驱动力矩和瞬时铣削力的映射关系;最后,基于最优变权法,综合考虑了主轴和进给轴电流对瞬时铣削力的影响,进行了瞬时铣削力预测实验。实验结果表明,基于主轴和进给轴电流最优变权法的瞬时铣削力预测误差在10%以内,能够有效预测加工过程的瞬时铣削力。     

基于神经网络的铣削复杂薄壁件受力变形分析和建模研究

铣削过程的复杂性使加工变形问题很难得到精确的解析解。为研究铣削过程中复杂薄壁件受力变形模型,将人工神经网络引入到摆线轮加工变形模型研究过程中,以有限元仿真结果为依据,通过改进的BP神经网络算法,建立了高速铣削轴承钢摆线轮铣削力与变形之间的非线性映射模型。结果显示所建立的网络模型具有较高的精度和良好的泛化能力,为进一步实现变形控制提供科学依据。     

基于加工特征的铣削力预测研究

航空发动机广泛采用钛合金薄壁结构,薄壁件在铣削加工过程中受铣削力的影响易于产生加工变形,影响加工质量。为减少加工变形,提高加工质量,需对铣削加工过程中的铣削力进行预测。为此,以Johnson-Cook本构方程为基础,考虑材料热力学动态性能和断裂准则对铣削力的影响,建立了基于加工特征的钛合金Ti-6Al-4V铣削力预测模型。首先,利用UG/Open工具模块对UG软件进行二次开发,创建了零件加工特征知识库。然后,利用Deform-3D仿真软件对材料本构模型、切屑分离和切屑断裂准则等进行描述,建立钛合金Ti-6Al-4V铣削加工有限元模型,对铣削力进行预测。铣削力实验证明了预测模型的可行性。最后,利用建立的有限元模型研究了工件曲率半径对铣削力的影响。结果表明,圆弧内轮廓铣削过程中的铣削力较大,圆弧外轮廓铣削过程中的铣削力较小。     

采用深度学习的铣刀磨损状态预测模型

为提高刀具磨损监测的预测精度与泛化性能,研究了基于深度学习的铣刀磨损状态预测,提出了基于堆叠稀疏自动编码网络与卷积神经网络的两种预测模型。堆叠稀疏自动编码网络对特征向量进行降维并将其纳入分类器来实现预测,可避免特征选择对先验知识的依赖;卷积神经网络将铣削振动数据转化为小波尺度图并输入模型完成分类,精简了传统建模流程。最后将提出的两种模型与传统神经网络模型进行比较,验证了所提模型的效率与精度。     

基于铣削力识别的铣削监测与铣刀故障诊断

为提高铣削过程监测与刀具故障诊断精度，通过测量铣床的频响函数和在铣削加工中的铣床振动加速度响应信号，用载荷识别的方法计算铣削力，分别得到了用4刀齿和2刀齿加工时横向铣削力的识别结果，所得到的铣削力曲线与加工工况吻合良好。以所识别铣削力为特征参量，用ART2神经网络进行了铣削过程监测与铣刀故障诊断，其结果比直接用振动响应信号进行监测与诊断更可靠，从而得到较好的监测诊断结论。     

Neural networks approach to the determination of the machining parameters

A neural networks based approach to determine the appropriate machining parameters such as speed, depth of cut and feed is proposed in this study. In this approach neural networks were used for building automatic process planning systems. Training of neural networks was performed with back propagation method by using data sets sampled in a standard handbook. These networks consist of simple processing, elements or nodes capable of processing information in response to external inputs. This approach saves computing time and storage space. In addition, it provides easy extendability as new data become available. Currently, the system provides three neural networks: for turning, for milling and for drilling operations. The performance of the trained neural network for drilling is evaluated to examine how well it predicts the machining parameters. Test results show that the neural network for the turning operation is able to predict the machining parameter values within an acceptable error rate.     

Selection of optimum cutting condition of cobalt-based superalloy with GONNS

Machining of new superalloys is challenging. Automated software environments for determining the optimal cutting conditions after reviewing a set of experimental results are very beneficial to obtain the desired surface quality and to use the machine tools effectively. The genetically optimized neural network system (GONNS) is proposed for the selection of optimal cutting conditions from the experimental data with minimal operator involvement. Genetic algorithm (GA) obtains the optimal operational condition by using the neural networks. A feed-forward backpropagation-type neural network was trained to represent the relationship between surface roughness, cutting force, and machining parameters of face-milling operation. Training data were collected at the symmetric and asymmetric milling operations by using different cutting speeds (V _c), feed rates (f), and depth of cuts (a _p) without using coolant. The surface roughness (Ra_asymt, Ra_symt) and cutting force (Fx_asymt, Fy_asymt, Fz_asymt, Fx_symt, Fy_symt, Fz_symt) were measured for each cutting condition. The surface roughness estimation accuracy of the neural network was better for the asymmetric milling operation with 0.4% and 5% for training and testing data, respectively. For the symmetric milling operations, slightly higher estimation errors were observed around 0.5% and 7% for the training and testing. One parameter was optimized by using the GONNS while all the other parameters, including the cutting forces and the surface roughness, were kept in the desired range.     

基于神经网络的数控铣削变形预测

数控铣削加工变形问题一直是自动化制造领域的瓶颈问题。铣削过程的复杂性及引起变形的多因素性使加工变形问题很难得到精确的解析解。本文在相关课题研究的基础上 ,将自动控制领域的前沿科学———人工神经网络引入该问题的研究进程之中 ,采用三层反向传播BP网络模拟铣削参数与变形间的非线性关系 ,为加工变形预测及进一步实现变形控制提供科学依据。     

数控加工铣削试验数据的清洗与处理

样本数据作为神经网络模型训练的导师,其质量直接影响了神经网络的学习能力以及神经网络模型的预测能力。以DMC60H为试验平台,以壳体类铝合金零件加工为研究对象,提取数控铣削加工试验数据;通过对数控铣削参数试验数据的分析与研究,提出了试验数据与样本数据的处理原则,分析了验证数据的构成以及所占数量。实现了样本数据的优化,并同时剔除了样本数据中的错误信息。     

MULTIPLE REGRESSION AND COMMITTEE NEURAL NETWORK FORCE PREDICTION MODELS IN MILLING FRP

This work utilizes the mechanistic modeling approach for predicting cutting forces and simulating the milling process of fiber-reinforced polymers (FRP) with a straight cutting edge. Specific energy functions were developed by multiple regression analysis (MR) and committee neural network approximation (CN) of milling force data and a cutting model was developed based on these energies and the cutting geometry. It is shown that both MR and CN models are capable of predicting the cutting forces in milling of unidirectional and multidirectional composites. Model predictions were compared with experimental data and were found to be in good agreement over the entire range of fiber orientations from 0 to 180°. Furthermore, CN model predictions were found to greatly outperform MR model predictions.     

MULTIPLE REGRESSION AND COMMITTEE NEURAL NETWORK FORCE PREDICTION MODELS IN MILLING FRP

This work utilizes the mechanistic modeling approach for predicting cutting forces and simulating the milling process of fiber-reinforced polymers (FRP) with a straight cutting edge. Specific energy functions were developed by multiple regression analysis (MR) and committee neural network approximation (CN) of milling force data and a cutting model was developed based on these energies and the cutting geometry. It is shown that both MR and CN models are capable of predicting the cutting forces in milling of unidirectional and multidirectional composites. Model predictions were compared with experimental data and were found to be in good agreement over the entire range of fiber orientations from 0 to 180°. Furthermore, CN model predictions were found to greatly outperform MR model predictions.     

An intelligent sensor for monitoring milling cutter breakage

A new intelligent sensor system using neural networks to separate tool breakage clearly from the cutter run-out or cutting transients in milling is proposed. The features of the spindle displacement signal are fed into the input layer of the neural network. With the back propagation training algorithm, the output of the neural network can be used to identify the milling cutter with or without tool breakage. Experiments show that this new approach can monitor tool breakage in milling operations successfully.