基于深度时序特征迁移的轴承剩余寿命预测方法

不同工况下轴承退化数据分布不一致导致深度学习等方法对剩余寿命预测效果有限,而已有迁移学习预测方法未能充分挖掘不同工况退化序列的内在趋势性,为此,提出一种基于深度时序特征迁移的轴承剩余寿命预测方法.首先,提出一种深度时序特征融合的健康指标构建模型,利用时间卷积网络挖掘退化趋势的内在时序特征,得到源域多轴承的健康指标;然后,提出一种最小化序列相似度的领域自适应算法,利用源域健康指标作为退化趋势元信息,选取目标域与源域之间的公共敏感特征;最后,采用支持向量机构建预测模型.在IEEE PHM Challenge 2012 轴承全寿命数据集上进行实验,结果表明,所提出方法构建的健康指标可更有效地反映退化趋势,同时明显提升剩余寿命预测的准确度.     

基于深度学习的液压缸寿命预测方法研究

液压缸的工况错综复杂,为了确保液压缸的正常运行,寿命预测系统采集了大量数据以获悉液压缸的寿命状况.针对液压缸监测信号噪声大、单一分类器分类性能不佳的问题,提出了一种基于深度学习的液压缸寿命预测方法.利用DAE算法对噪声数据进行重构,以完成数据的特征提取;利用BP神经网络对数据中各特征子集进行分别训练构成弱分类器,然后采...     

基于PSO-RBF神经网络的刀具寿命预测

有效的刀具寿命预测可以提高加工效率,保证工件加工精度,因此具有重要的研究价值.刀具寿命预测受到刀具材质、切削参数以及加工材料等多因素的影响,导致刀具寿命难以准确预测.针对这一问题提出了一种利用粒子群(particle swarm optimization,PSO)算法优化径向基(radial basis functio...     

基于主成分分析法的惯性器件寿命预测

在基于随机滤波理论的剩余寿命预测模型中,所使用的输入数据为单维,而工程实践中惯性器件有多维监测信息,因此,在现有寿命预测模型中,根据专家经验所选取的单维数据仅使用其中一部分信息。针对以上问题,提出了1种基于主成分分析的寿命预测方法。主成分分析可以将相互关联的多维数据用少数几个不相关的主成分进行代替,这样在减少有用信息损失的同时,使问题得到简化。应用实际的多维历史监测数据和主成分分析的寿命预测模型,进行了某导弹惯性平台的寿命预测实验。实验结果表明,采用基于主成分分析的预测模型能够有效地进行寿命预测,且精度较高。     

基于深度信号处理和堆叠残差GRU的刀具磨损智能预测模型

刀具磨损的智能监测是影响现代机械加工业智能化发展进程的重要因素。在机械加工过程中,大多数机床通过使用传感器采集信号,从而建立刀具磨损与传感器信号之间的关系,在不中断加工过程的情况下中实现刀具的磨损预测,根据是否达到磨损阈值来判断是否自动换刀或报警以实现在线智能监控。能否从传感器信号中提取有效的特征信息,并且建立一个快速响应且精确的预测模型是一个亟待解决的问题。因此,针对上述问题,提出了一种基于深度信号处理和堆叠残差GRU的刀具磨损预测模型。在信号处理方面设计了BiGRU-Self Attention(BGSA)模块,利用双向门控循环单元和内部注意力机制来获取动态时序特征,反映出不同特征的影响程度,以提高建模效率。同时,提出堆叠残差GRU(Stacked-ResGRU)模型实现对刀具实时磨损值的预测,通过残差结构优化了网络结构,加快了模型的收敛速度。通过使用公开的数据集,对铣削过程中刀具的磨损状态预测进行实验研究,实验结果验证了该方法的可行性与有效性。通过对实验结果的对比分析可知,所提模型能够有效地表征刀具磨损程度,并大大减小了预测误差,取得了良好的预测效果。     

K-Similarity降噪的LSTM神经网络水质多因子预测模型

针对水质预测问题,以地表水水质监测因子作为研究对象,提出了一种基于长短期记忆（LSTM）神经网络的水质多因子预测模型,同时利用提出的K-Similarity降噪法对模型的输入数据进行降噪,提高模型预测性能.通过与BP神经网络、RNN和传统的LSTM神经网络预测模型进行对比实验,证明了所提出的方法均方误差最小,预测结果更准确.     

刀具磨损自动识别及检测系统

对加工状态下刀具磨损值及时有效的检测,可以在保证加工精度的前提下提高加工效率,同时为刀具剩余寿命预测提供有力的数据支撑.针对刀具磨损值测量过程中人工参与的不足、易受主观因素影响、检测精度低等问题,建立了刀具磨损图像自动检测系统,基于最大类间方差及遗传算法迭代寻找最佳阈值,进行磨损区域分割;利用磨损区域呈现"线状"的特点,进行磨损区域特征识别及滤波;基于多级Hough变换以及磨损区域二次识别的方法,将刀具磨损区域在图像中清晰地提取出来.在此基础上,基于Canny算子边缘检测方法建立刀具磨损曲线,计算出磨损区域的磨损值.在刀具磨损检测系统计算的磨损量与超景深显微镜的人工测量结果进行比对,经实验表明,该磨损检测系统检测误差在±6％以内,并且提高了刀具磨损值的检测效率.     

基于混合智能计算的铣刀状态监测

提出了一种基于混合智能融合技术进行铣刀磨损量监测和预测方法。利用多传感器对切削力和振动信号进行监测,通过频率变换提取切削力特征量,采用小波包分解技术提取振动信号特征量。通过信号特征值的组合,分别探讨了几种混合智能数据融合技术-小波神经网络,遗传神经网络,遗传小波神经网络对刀具磨损量的预测效果。实验分析表明,提出的几种基于多传感器的混合智能数据融合技术均能够有效地完成刀具磨损量监测和预测,同时对它们各自的特点进行了比较分析。     

基于计算智能算法的铣刀状态监测

本文提出了基于智能融合技术进行铣刀磨损量监测和预测方法。利用多传感器对切削力和振动信号进行监测,通过频率变换提取切削力特征量,采用小波包分解技术提取振动信号特征量。通过信号特征值的组合,分别探讨了几种计算智能数据融合技术-小波神经网络、遗传神经网络、遗传小波神经网络对刀具磨损量的预测效果。实验分析表明,本文提出的几种计算智能数据融合技术均能够有效地完成刀具磨损量预测。     

基于PPRMS和循环神经网络的刀具磨损趋势预测

针对数控机床的刀具加工过程中噪声干扰大、磨损特征微弱导致其寿命退化趋势难以预测等问题，提出了一种基于PPRMS和循环神经网络的刀具磨损趋势预测方法。采用峰峰值（Peak to Peak）计算刀具寿命状态数据的幅值波动性大小，利用高斯4σ准则剔除异常值，抑制随机噪声的干扰；采用均方根值（Root Mean Square, RMS）计算刀具健康因子，结合Bi-LSTM和Bi-GRU网络搭建刀具磨损趋势退化预测网络模型，提取出刀具磨损退化微弱特征。在全寿命公开数据集PHM-2010上验证了所提方法的有效性，预测误差MSE、 RMSE和MAE均优于单一的Bi-LSTM模型。     

Online tool wear prediction system in the turning process using an adaptive neuro-fuzzy inference system

Tool wear is a detrimental factor that affects the quality and tolerance of machined parts. Having an accurate prediction of tool wear is important for machining industries to maintain the machined surface quality and can consequently reduce inspection costs and increase productivity. Online and real-time tool wear prediction is possible due to developments in sensor technology. Recently, various sensors and methods have been proposed for the development of tool wear monitoring systems. In this study, an online tool wear monitoring system was proposed using a strain gauge-type sensor due to its simplicity and low cost. A model, based on the adaptive network-based fuzzy inference system (ANFIS), and a new statistical signal analysis method, the I-kaz method, were used to predict tool wear during a turning process. In order to develop the ANFIS model, the cutting speed, depth of cut, feed rate and I-kaz coefficient from the signals of each turning process were taken as inputs, and the flank wear value for the cutting edge was an output of the model. It was found that the prediction usually accurate if the correlation of coefficients and the average errors were in the range of 0.989–0.995 and 2.30–5.08% respectively for the developed model. The proposed model is efficient and low-cost which can be used in the machining industry for online prediction of the cutting tool wear progression, but the accuracy of the model depends upon the training and testing data.     

In-process tool condition forecasting based on a deep learning method

It is widely acknowledged that machining precision and surface integrity are greatly affected by cutting tool conditions. In order to enable early cutting tool replacement and proactive actions, tool wear conditions should be estimated in advance and updated in real-time. In this work, an approach to in-process tool condition forecasting is proposed based on a deep learning method. A long short-term memory network is designed to forecast multiple flank wear values based on historical data. A residual convolutional neural network is built to enable in-process tool condition monitoring, using raw signals acquired during the machining process. The integration of them enables in-process tool condition forecasting. Median-based correction and mean-based correction are adopted to improve the accuracy. IEEE PHM 2010 challenge data has been used to illustrate and validate this approach. Experimental study and quantitative comparisons showed that future flank wear values could be precisely forecasted during the machining process. The proposed approach contributes to prompt and reliable cutting tool condition forecasting, which will support the decision-making about cutting tool replacement in data-driven smart manufacturing.     

Research on multi-signal milling tool wear prediction method based on GAF-ResNext

A key aspect impacting the quality and efficiency of machining is the degree of tool wear. If the tool failure is not discovered in time, the quality of workpiece processing decreases, and even the machine tool itself may be harmed. To increase machining quality, efficiency and facilitate the intelligent advancement of the manufacturing industry, tool wear prediction is crucial. This research offers a multi-signal tool wear prediction method based on the Gramian angular field (GAF) and depth aggregation residual transform neural network (ResNext), enabling fast and accurate tool wear prediction. Specifically, the required one-dimensional signal is obtained through preprocessing including intercepting, splicing and wavelet threshold denoising of the force and vibration signals, and GAF is used to encode the obtained one-dimensional signal to generate a (224 × 224) data matrix. ResNext automatically extracts the features of the data matrix, establish the relationship between features and tool wear, and creates a tool wear prediction model based on GAF-ResNext. The ability of this method to predict tool wear has been trained and tested by milling experimental data. The experimental findings demonstrate the real-time, accuracy, dependability and universality of this method. This method has a better effect when compared to other research methods. The study's findings can boost machining productivity and offer technical support for intelligent tool wear early warning and intelligent manufacturing.     

基于双通道卷积神经网络的航班延误预测模型

针对航班延误预测数据量大、特征提取困难而传统算法处理能力有限的问题,提出一种基于双通道卷积神经网络（DCNN）的航班延误预测模型。首先,该模型将航班数据和气象数据进行融合,应用DCNN进行自动特征提取,采用批归一化（BN）和Padding策略优化,提升到港延误等级的分类预测性能;然后,在卷积神经网络（CNN）基础上加入直通通道,以保证特征矩阵的无损传输,增强深度网络的畅通性;同时引入卷积衰减因子对卷积通道的特征矩阵进行稀疏性限制,控制不同网络深度的特征叠加比例,维持模型的稳定性。实验结果表明,所提模型与传统模型相比,具有更强的数据处理能力。通过数据融合,航班延误预测准确率可提高1个百分点;加深网络深度后,该模型能保证梯度的稳定,从而训练更深的网络,使准确率提升至92.1%。该基于DCNN算法的模型特征提取充分,预测性能优于对比模型,可更好地服务于民航决策。     

Design of a decision support system for machine tool selection based on machine characteristics and performance tests

Economic globalization, together with heightened market competition and increasingly short product life cycles are motivating companies to use advanced manufacturing technologies. Use of high speed machining is increasingly widespread; however, as the technology is relatively new, it lacks a deep-rooted knowledge base which would facilitate implementation. One of the most frequent problems facing companies wishing to adopt this technology is selecting the most appropriate machine tool for the product in question and own enterprise characteristics. This paper presents a decision support system for high speed milling machine tool selection based on machine characteristics and performance tests. Profile machining tests are designed and conducted in participating machining centers. The decision support system is based on product dimension accuracy, process parameters such as feed rate and interpolation scheme used by CNC and machine characteristics such as machine accuracy and cost. Experimental data for process error and cycle operation time are obtained from profile machining tests with different geometrical feature zones that are often used in manufacturing of discrete parts or die/moulds. All those input parameters have direct impact on productivity and manufacturing cost. Artificial neural network models are utilized for decision support system with reasonable prediction capability.     

基于模态区间-马尔科夫链的动刚度预测

动刚度反映着数控加工过程中的稳定性,对加工精度有直接影响.对动刚度进行预测,预知动刚度的变化趋势,可为处理数控机床动刚度劣化问题提供指导作用.传统的马尔科夫链模型预测方法难以处理不确定性问题,将降低预测结果的可靠性.将模态区间数学理论应用于传统马尔科夫链中,提出一种基于模态区间-马尔科夫链模型的预测方法,以提高预测结果的可靠性.为了验证提出的预测方法的有效性,设计了一个数控机床相对激振实验.通过相对激振方法获取历史动刚度数据,采用模态区间-马尔科夫链模型对动刚度劣化趋势进行预测.结果表明:提出的预测方法有一个更好的预测精度.     

A sensor fusion and support vector machine based approach for recognition of complex machining conditions

During the machining process of thin-walled parts, machine tool wear and work-piece deformation always co-exist, which make the recognition of machining conditions very difficult. Existing machining condition monitoring approaches usually consider only one single condition, i.e., either tool wear or work-piece deformation. In order to close this gap, a machining condition recognition approach based on multi-sensor fusion and support vector machine (SVM) is proposed. A dynamometer sensor and an acceleration sensor are used to collect cutting force signals and vibration signals respectively. Wavelet decomposition is utilized as a signal processing method for the extraction of signal characteristics including means and variances of a certain degree of the decomposed signals. SVM is used as a condition recognition method by using the means and variances of signals as well as cutting parameters as the input vector. Information fusion theory at the feature level is adopted to assist the machining condition recognition. Experiments are designed to demonstrate and validate the feasibility of the proposed approach. A condition recognition accuracy of about 90 % has been achieved during the experiments.     

伴随时空特性的雷电预测BP-ANN模型研究

为提高雷电预测模型的准确率和学习性能，提出一种基于增量学习和时空特性的雷电预测BP-ANN二项分类器。通过增量方式和依据数据的时空特征进行历史数据的学习，建立多种BP-ANN模型，分别对新的数据进行预测分类，然后采用多数投票方式确定新数据的类别。分别构建基于增量学习的BP-ANN模型、基于时空特性的BP-ANN模型以及结合基于增量学习和时空特性的BP-ANN模型这3种雷电预测模型，并在真实雷电数据集上进行预测准确度和学习性能的测试，结果表明了增量学习、时空特性以及二者结合的优劣。     

In-process regressions and adaptive multicriteria neural networks for monitoring and supervising machining operations

The authors develop a monitoring and supervising system for machining operations using in-process regressions (for monitoring) and adaptive feedforward artificial neural networks (for supervising). The system is designed for: (1) in-process tool life measurement and prediction; (2) supervision of machining operations in terms of the best machining setup; and (3) catastrophic tool failure monitoring. The monitoring system predicts tool life by using different sensors for gathering information based on a regression model that allows for the variations between tools and different machine setups. The regression model makes its prediction by using the history of other tools and combining it with the information obtained about the tool under consideration. The supervision system identifies the best parameters for the machine setup problem within the framework of multiple criteria decision making. The decision maker (operator) considers several criteria, such as cutting quality, production rate and tool life. To make the optimal decision with several criteria, an adaptive feedforward artificial neural network is used to assess the decision maker's preferences. The authors' neural network approach learns from the decision maker's complex behavior and hence, in automatic mode, can make decisions for the decision maker. The approach is not computationally demanding, and experiments demonstrate that its predictions are accurate.     

基于深度卷积神经网络的车型识别

传统的基于卷积神经网络的车型识别算法存在识别相似车型的准确率不高,以及在网络训练时只能使用图像的灰度图从而丢失了图像的颜色信息等缺陷。对此,提出一种基于深度卷积神经网络(Deep Convolution Neural Network,DCNN)的提取图像特征的方法,运用深度卷积神经网络对背景较复杂的车型进行网络训练,以达到识别车型的目的。文中采用先进的深度学习框架Caffe,基于AlexNet结构提出了深度卷积神经网络的模型,分别对车型的图像进行训练,并与传统CNN算法进行比较。实验结果显示,DCNN网络模型的准确率达到了96.9%,比其他算法的准确率更高。