竖炉焙烧过程生产质量监控系统

针对竖炉焙烧过程的质量指标磁选管回收率难以实时在线测量问题,基于RBF神经网络与专家系统提出了由磁选管回收率预报模型和生产质量诊断模型构成的竖炉焙烧质量监控系统.经过现场检验,该系统能够准确实时地预报磁选管回收率,并能够对生产质量进行诊断,提出合理的参数调整方法以避免不合格产品的出现.磁选管回收率提高2%,产品合格率提高了50%,有效地保证了竖炉焙烧过程的生产质量.     

Applications of Neural Networks in Spinning Prediction

Th netal network spinning prediction model (BPana RBF Networks) trained by data from the mill canpredict yarn qualities and spinning performance. Theinput parameters of the model are as follows: yarncount, diameter, hauteur, bundle strength, spinningdraft, spinning speed, traveler number and twist.     

基于神经网络的精梳毛纺纱线预测

在分析纺纱原理的基础上，采用人工神经网络建立7个不同的模型，预测精梳毛纺的纱线质量和纺纱性能，分别是纱线不匀、粗细节、断裂强力、强力不匀、断裂伸长和断头率。其中断头率由于其复杂的成因，采用了组合神经网络建模。采用工厂实际生产数据进行验证，前6个指标的预测值与实测值之间的相关系数的平方均超过0．9，断头的预测效果相比而言比较差，但相关系数的平方也超过了0．8，表明人工神经网络技术在精梳毛纺纱线预测方面有很大的应用前景。     

Empirical Model for Evaluating Covered Effect on Polyimide/Metal Wire Cowrapped Yarn by Response Surface Methodology

To accurately evaluate and predict the covered effect of cowrapped yarn,a novel characterization is performed by covered ratio and fineness. Polyimide / metal wire co-wrapped yarn which was designed for applications in aerospace and composites was developed through hollow spindle spinning process. Core yarn speed,hollow spindle rotating speed,and wrapping yarn twist were selected as three main factors that affected spinning process. An empirical model indicating relationship between spinning parameters and covered effect was established based on response surface methodology( RSM). The results show that wrapping yarn twist contributes greatly to smooth wrapping process. Core yarn speed and spindle rotating speed are significant impact factors of covered effect and they interact significantly in covered ratio, but indistinctively in fineness.     

Comprehensive Characterization Model of Integrated Cotton Fiber Quality Index

An integrated cotton fiber quality index (ICFQI) model with cotton fiber qualities which can directly express cotton fiber integrated quality and spinning yarn quality was studied.The fiber length,strength,Micronaire ( fiber fineness and fiber maturity),uniformity of fiber length,and short fiber content are the pivotal indexes expressing ICFQI.All of the results above are the basic knowledge to build up the models of ICFQI.According to spinning consistency index (SCI),spinning strength and spinning yarn integrated quality,ICFQI was the best choice.As the methods of ICFQI had quite a lot of advantages like explicit mechanism,few independent variables.The integrated fiber quality index had a significant positive correlation with yarn strength and spinning consistency,significant negative correlation with yarn evenness and yarn thin places.In additional,the model of the relationship between ICFQI and SCI was established as:SCI =0.235 6 · ICFQI+56.153.It was concluded that ICFQI value was the shared reference index for the testing of fiber inspection agency and the selection and distribution of raw cotton bales by textile mills.     

19.7tex粘胶转杯纺针织纱的纺制

采用转杯纺纱技术开发中细特粘胶针织纱,产品附加值高,经济效益好。通过对粘胶原料、前纺工艺流程及 参数、转环纺元件和工艺参数的合理选择,采用优化工艺纺制的19.7tex粘胶转杯纺针织纱质量优,产量高。     

On-line Cutting Quality Recognition in Milling Using a Radical Basis Function Neural Network

Tool wear, chatter vibration, chip breaking and built-up edge are main phenomena to be monitored in modern manufacturing processes, which are considered as important factors to the quality of products.They are closely related to the cutting parameters, which are to be selected in manufacturing process.However, it is very difficult to measure directly the cutting quality based on on-line monitoring.In this study, the relationship between the cutting parameters and cutting quality is analyzed.A Radical Basis Function (RBF) neural network based on-line quality recognition scheme is also presented, which monitors the level of surface roughness.The experimental results reveal that the RBF neural network has a high prediction success rate.     

On-line Cutting Quality Recognition in Milling Using a Radical Basis Function Neural Network

Tool wear, chatter vibration, chip breaking and built-up edge are main phenomena to be monitored in modern manufacturing processes, which are considered as important factors to the quality of products.They are closely related to the cutting parameters, which are to be selected in manufacturing process.However, it is very difficult to measure directly the cutting quality based on on-line monitoring.In this study, the relationship between the cutting parameters and cutting quality is analyzed.A Radical Basis Function (RBF) neural network based on-line quality recognition scheme is also presented, which monitors the level of surface roughness.The experimental results reveal that the RBF neural network has a high prediction success rate.     

纯毛纱断裂伸长率预测技术的研究

在毛纺加工流程中,纺纱是关键工序。因为纺纱工序的耗资是整个毛条制造的3-4倍,而且织物的质量在很大程度上取决于纱线质量。因此,预测技术的应用对毛纺厂改善纺纱性能和提高纱线质量具有非常重要的经济价值。运用统计软件SPSS对66个批次的纯毛单纱建立了断裂伸长率的预测模型,并且分析了影响纱线断裂伸长率的主要因素。预测技术作为一种质量控制手段具有以下作用:预测纱线品质和纺纱性能,优化毛条选择,调整生产加工工艺。     

管道纺纱主要工艺参数的优化

管道纺纱是一种新研究出来的自由端纺纱方法,要能实际纺纱还必须合理地选择其工艺参数.本文分析了影响管道纺纱的主要工艺参数.采用二次旋转组合设计方法,通过实验,建立了主要工艺参数与成纱强力的数学模型,确定了相应的约束条件.并采用惩罚函数法进行优化计算,求得最佳工艺条件.再根据优化计算出来的工艺条件,在管道纺纱单头试验机上实际纺制28tex纯棉纱.不仅能稳定地纺纱,而且其纱强力与优化计算结果基本一致.     

基于神经网络多参数融合的钻井过程状态监测与故障诊断

复杂系统状态监测与故障诊断是系统安全运行过程中的重要保障,分析了钻井系统事故状态下特征参数的变化,给出了用神经网络进行故障诊断的流程,在利用样本数据对网络进行训练的基础上建立了稳定的神经网络诊断模型。输入各种状态下的新样本数据,能够得到正确的系统状态识别,通过改进网络算法改进了网络性能。对生产数据的处理结果表明,基于神经网络的多参数融合算法可以很好地识别钻井过程中的不同状态,能够实现状态检测与故障诊断。     

从兔毛织物的掉毛规律研究兔毛产品的加工及开发

用不同磨料对不同配比、不同捻度、不同组织密度、不同机织物组织、不同纺纱系统等制得的兔毛织物，进行模拟掉毛试验，探讨了各种条件下兔毛的掉毛规律．在兔毛产品开发上应注意配毛时兔毛及与其混纺原料的质量和品种，既可开发粗纺呢绒兔毛产品，也可开发精纺及半精纺兔毛产品，也可采用棉纺系统或毛棉混合系统进行加工。     

Yarn Properties Prediction Based on Machine Learning Method

Although many works have been done to constructprediction models on yarn processing quality, the relationbetween spinning variables and yam properties has not beenestablished conclusively so far. Support vector machines(SVMs), based on statistical learning theory, are gainingapplications in the areas of machine learning and patternrecognition because of the high accuracy and goodgeneralization capability. This study briefly introduces theSVM regression algorithms, and presents the SVM basedsystem architecture for predicting yam properties. Model.selection which amounts to search in hyper-parameter spaceis performed for study of suitable parameters with grid-research method. Experimental results have been comparedwith those of artificial neural network(ANN) models. Theinvestigation indicates that in the small data sets and real-life production, SVM models are capable of remaining thestability of predictive accuracy, and more suitable for noisyand dynamic spinning process.     

关于工艺产品加工质量监测优化研究

质量监测可以有效的提高产品质量和生产效率。在复杂产品的生产过程当中,多个质量特性之间相互作用,共同对产品的生产质量产生影响,由于质量特性的数量较多、有些特性的关系是耦合的,因此准确诊断出异常变量是研究的难点。为了高效、准确的诊断出异常变量,提高产品的质量和生产效率,提出了基于改进网格优化的PCA-SVM多元控制图均值偏移诊断模型。在模型训练之前,使用主元分析(PCA)算法对数据进行预处理,降低数据维数和提取数据特征信息,再用改进网格算法对支持向量机(SVM)的参数进行优化,最终得到优化的SVM模型。仿真结果表明,采用的方法与传统方法相比,训练时间更短,且拥有更高的分类准确率。     

一种利用LSTM和稳态时间序列的光伏阵列故障诊断方法

受最大功率点跟踪算法和时变环境条件的影响,光伏阵列的电气工作参数包含了复杂的暂态过程以及工频干扰噪声,严重影响了故障特征质量以及诊断算法性能。针对该问题,本文首先提出了一种基于最大功率点（MPP）的稳态时间序列预处理方法,以自动过滤数据中的暂态过程和干扰噪声,获取连续的稳态时间序列电气特征数据,作为故障诊断模型的输入参数;然后,提出了一种基于长短期记忆网络（LSTM）的深度网络模型,以实现对光伏阵列常见故障的检测及分类;最后,在一个小型光伏并网发电系统及其Simulink仿真模型上,进行故障模拟及仿真以验证所提出的故障诊断方法。实验结果表明所提出的故障诊断方法具有良好的精度和泛化性能,并且优于常规的反向传播神经网络（BPNN）和循环神经网络（RNN）。     

一种利用LSTM和稳态时间序列的光伏阵列故障诊断方法

针对光伏阵列电气工作参数包含复杂的暂态过程及工频干扰噪声严重影响故障诊断模型性能的问题,提出一种基于最大功率点的稳态时间序列预处理方法.首先,以自动过滤数据中的暂态过程和干扰噪声,获取连续的稳态时间序列电气特征数据,作为故障诊断模型的输入参数;然后,提出一种基于长短期记忆网络的深度网络模型,以实现对光伏阵列常见故障的检测及分类;最后,在一个小型光伏并网发电系统及其Simulink仿真模型上,进行故障模拟及仿真,以验证所提出的故障诊断方法.实验结果表明,所提出的故障诊断方法具有良好的精度和泛化性能,并且优于常规的反向传播神经网络和循环神经网络.     

化学化工实验室无线安全监测系统故障诊断方法研究

针对化学化工实验室的安全问题设计了一种无线安全监测系统,通过无线通信模块传输现场检测数据和经压缩后的图片,实现了实验室安全远程无线监测与监视的功能. 由于监测系统中的无线通信模块能耗较高,长期运行容易发生故障,其故障检测方式也较为复杂. 为了在线检测无线通信模块的故障和识别其类型,以确保无线安全监测系统的可靠性,研究了无线通信模块的电流特性,建立了基于模糊神经网络的故障诊断模型,可实现对无线通信模块在不同状态下的故障进行诊断. 实验结果表明,与BP神经网相比,采用模糊神经网络的无线通信模块故障诊断方法训练耗时短、收敛快、训练误差和验证误差小、诊断正确率高,能够在线检测无线通信模块的多种类型故障,明显提高了无线安全监测系统的可靠性,具有较好的实际应用价值.      

提高成纱条干均匀度的工艺研究和技术措施

成纱条干均匀度与原料中的短纤维含量、牵伸工艺配置、机械设备状态、操作和清洁等因素有关.为了提高成纱条干均匀度,本文基于系统工程纺纱学理论,从原料选配至细纱的每一道工序优选了工艺参数和纺纱零部件,并通过加强生产管理,进行精细化控制,使纱线达到质量要求.     

圆盘式旋流纺纱的工艺分析

介绍了圆盘式旋流纺纱的装置和流程，探讨了影响圆盘式旋流纺纱成纱性能的主要工艺参数，成功地纺制了包芯纱．为了解和掌握圆盘式旋流纺纱设备的纺纱性能、工艺特点，对所纺纱线的强度和耐磨性进行了测试．根据测试结果，分析了圆盘式旋流纺纱的主要工艺参数与成纱性能之间的关系，发现第一喷嘴压力、第二喷嘴压力、梳辊转速、喂棉量、棉条号数、喂入张力、芯纱退绕张力是圆盘式旋流纺纱的主要工艺参数．     

基于BP神经网络的纺纱质量预报模型

简述了纺纱质量预报的理论与实际意义，应用了BP神经网络的标准算法。通过建立BP神经网络模型对细纱条干不匀率(CV）断裂强力(BS)与纺纱断头率(ED)进行预报，并与多元线性回归方法进行比较，验证了BP神经网络预报的准确性与高效性。最后以纺纱断头率(ED)预报为例，指出了今后神经网络在纺纱质量预报应用中亟待解决的问题。