空间曲线的特征识别与高质量非均匀采样

为避免传统均匀采样方法因忽视曲线重要特征而生成不理想的采样结果,获得给定数量且由特征点和辅助点组成的采样点序列,提出基于特征识别的高质量空间曲线非均匀采样方法.首先使用抛物线插值法得到曲线上所有曲率极大值点和挠率极大值点的近似位置,经筛选后产生特征点,以更好地抓住空间曲线的轮廓特征.然后定义基于弧长、曲率和挠率加权组合的特征函数,并以此自适应地选取曲线上的辅助点.与3种主流采样方法比较的实验结果表明,该方法能够获得更高质量的采样结果且具有更好的实用性,从而进一步改善空间曲线的B样条拟合效果.     

基于ERNIE-BiGRU模型的摘要语步自动识别研究

学术文献的摘要是对文献主要内容的浓缩,摘要不同部分的语步具有不同的信息,语步的自动识别和抽取对于学术摘要的后续研究有着重要的应用价值,而目前语步识别的研究相对较少,并且相关算法的效果还需要提高。针对上述问题,该文提出了一种基于ERNIE-BiGRU模型的语步识别算法。该算法首先结合中文句法分析理论提出基于句法依存关系的多语步结构拆分法,对学术文献摘要多语步结构进行自动拆分,获得多个单语步结构;然后构建用于训练的单语步结构语料库,并利用知识增强语义表示预训练模型,训练出句子级词向量;最后将训练出的单语步结构词向量信息输入双向门限循环单元(BiGRU)进行摘要语步自动化识别,取得了良好的效果。实验结果表明,该算法具有较好的鲁棒性和较高的识别精度,在结构化和非结构化摘要上的识别准确率分别达到了96.57%和93.75%。     

Effects of multiple freeze–thaw cycles on meat quality,nutrients, water distribution and microstructure in bovine rumen smooth muscle

This study investigated the effect of 5 freeze–thaw cycles (freezing at −18°C for 12 h and then thawing at 4°C for approximately 12 h) on the meat quality, proximate composition, water distribution and microstructure of bovine rumen smooth muscle (BSM). As the number of freeze–thaw cycles increased, BSM pH, shear force, water content and protein content decreased by 3.06%, 35.50%, 14.49% and 21.11%, respectively, whereas BSM thawing loss, cooking loss, pressing loss, total aerobic count (TAC), ash content and fat content increased by 108.12%, 47.75%, 78.33%, 90.99%, 105% and 35.20%, respectively. The freeze–thaw cycles resulted in greater protein and lipid oxidation, as evidenced by a 36.46% reduction in the sulfhydryl content and a 209.06% and 338.46% increase in the carbonyl and malondialdehyde contents, respectively. Ice crystal formation disrupted the structural integrity of the muscle tissue. Low-field nuclear magnetic resonance results showed that the freeze–thaw cycles prolonged the relaxation times (T_2b, T₂₁ and T₂₂), indicating that immobile water shifted to free water, and consequently, free water mobility increased. After 3 freeze–thaw cycles, the decline in shear force slowed, the increase in thawing loss became accelerated, and the TAC approached the domain value (6 log colony-forming units/g). Therefore, the number of freeze–thaw cycles of smooth muscle during transport, storage and distribution should be controlled to 3 or fewer. The current results provide a theoretical basis and data support for the further utilisation and culinary processing of smooth muscle.     

基于SVD和SSA-VMD降噪的轴承故障特征提取

针对强噪声背景下轴承故障特征提取困难的问题，提出一种基于奇异值分解和参数优化变分模态分解联合降噪的轴承故障特征提取方法(SSVMD)：首先，对原始信号进行奇异值分解(Singular Value Decomposition，SVD)处理，运用奇异值差分谱法选取有效奇异值并将原始信号重构得到初步降噪信号；其次，为防止故障信息丢失，将残余信号进行麻雀算法(Sparrow Search Algorithm，SSA)优化的变分模态分解(Variational Mode Decomposition，VMD）算法处理，得到最佳的模态个数K和惩罚参数α，选取峭度值最大、包络熵最小的IMF分量与初步降噪信号叠加得到最终降噪信号，并对信号进行包络分析；最后，通过仿真和试验数据分析得出，该方法能在信噪比很低的情况下降低噪声含量并提取轴承故障特征，为设备的状态监测和故障诊断提供理论依据。     

CDBN‑IKELM的轴承变工况故障诊断方法

针对现有方法在轴承变工况方面存在的诊断精度低、人工提取特征不充分等问题，提出了基于卷积深度置信网络（convolutional deep belief network，简称CDBN）与改进核极限学习机 （improved Kernel?based extreme learning machine，简称IKELM）的滚动轴承故障智能识别方法。首先，由卷积深度置信网络对原始信号内的故障特征进行深层自适应提取；其次，利用等距特征映射对提取的多维特征进行降维，去除冗余特征信息；然后，采用改进的核极限学习机对特征进行分类，使用粒子群（particle swarm optimization， 简称PSO）对模型重要参数进行优化，实现滚动轴承变工况下的故障识别；最后，将所提方法应用于不同工况下多种轴承故障的诊断。实验结果表明，该方法能够智能有效地识别变工况的轴承故障，诊断结果优于已有的智能故障诊断方法。     

Improving accuracy of automatic optical inspection with machine learning

Electronic devices require the printed circuit board(PCB)to support the whole structure,but the assembly of PCBs suffers from welding problem of the electronic components such as surface mounted devices(SMDs)resistors.The automated optical inspection(AOI)machine,widely used in industrial production,can take the image of PCBs and examine the welding issue.However,the AOI machine could commit false negative errors and dedicated technicians have to be employed to pick out those misjudged PCBs.This paper proposes a machine learning based method to improve the accuracy of AOI.In particular,we propose an adjacent pixel RGB value based method to pre-process the image from the AOI machine and build a customized deep learning model to classify the image.We present a practical scheme including two machine learning procedures to mitigate AOI errors.We conduct experiments with the real dataset from a production line for three months,the experimental results show that our method can reduce the rate of misjudgment from 0.3%–0.5%to 0.02%–0.03%,which is meaningful for thousands of PCBs each containing thousands of electronic components in practice.     

基于VMD和改进TCN的短期光伏发电功率预测

光伏发电功率存在波动性,且光伏出力易受各种气象特征影响,传统TCN网络容易过度强化空间特性而弱化个体特性。针对上述问题,文中提出一种基于VMD和改进TCN的短期光伏发电功率预测模型。通过VMD将原始光伏发电功率时间序列分解为若干不同频率的模态分量,将各个模态分量以及相对应的气象数据输入至改进TCN网络进行建模学习。利用中心频率法确定VMD的最优分解模态分解个数。在传统TCN预测模型的基础上,使用DropBlock正则化取代Dropout正则化以达到抑制卷积层中信息协同的效果,并引入注意力机制自主挖掘并突出关键气象输入特征的影响,量化各气象因素对光伏发电的影响,从而提高预测精度。以江苏省某光伏电站真实数据为例进行仿真实验,结果表明所提预测方法的RMSE为0.62 MW,MAPE为2.03%。