基于混合特征LLE融合与SVM的质量异常模式识别

针对生产过程中存在的异常模式识别的问题,提出基于LLE融合与支持向量机的质量异常模式识别方法。首先从动态数据流中提取其原始特征、统计特征、几何特征并将其进行混合,形成动态数据流的混合特征,然后利用LLE算法对混合特征进行降维,将降维后的特征集作为MSVM分类器的输入进行训练,同时采用粒子群算法对MSVM分类器进行参数寻优。最后用训练好的模型对动态数据流进行异常模式的识别。并将所提方法与单一类型特征方法、混合特征方法的识别模型进行比较,仿真结果和应用实例表明,所提方法的识别精度较高,可用于生产过程的质量异常模式识别中。     

一种振动矢量测量及调整的方法

针对消失模生产中振实设备的振动过程不满足工艺要求的情况下会产生的问题,详细阐述了基于消失模振实、紧实过程的测量方法和评价方法,通过使用该种方法可以得出振动矢量和振幅,进而对振实紧实设备和振动紧实工艺进行调整和优化,杜绝干砂振动紧实过程中的各种问题,保证铸件质量。可以普遍应用在干砂振实设备研发、校检、定期维护等各个过程,可以有效提升设备的工艺符合性,减少模样及铸件不合格品的产生。     

Multi-objective optimization of shield construction parameters based on random forests and NSGA-II

Metro shield construction will inevitably cause changes in the stress and strain state of the surrounding soil, resulting in stratum deformation and surface settlement (SS), which will seriously endanger the safety of nearby buildings, roads and underground pipe networks. Therefore, in the design and construction stage, optimizing the shield construction parameters (SCP) is the key to reducing the SS rate and increasing the safe driving speed (DS). However, optimization of existing SCP are challenged by the need to construct a unified multiobjective model for optimization that are efficient, convenient, and widely applicable. This paper innovatively proposes a hybrid intelligence framework that combines random forest (RF) and non-dominant classification genetic algorithm II (NSGA-II), which overcomes the shortcomings of time-consuming and high cost for the establishment and verification of traditional prediction models. First, RF is used to rank the importance of 10 influencing factors, and the nonlinear mapping relationship between the main SCP and the two objectives is constructed as the fitness function of the NSGA-II algorithm. Second, a multiobjective optimization framework for RF-NSGA-II is established, based on which the optimal Pareto front is calculated, and reasonable optimized control ranges for the SCP are obtained. Finally, a case study in the Wuhan Rail Transit Line 6 project is examined. The results show that the SS is reduced by 12.5% and the DS is increased by 2.5% with the proposed framework. Meanwhile, the prediction results are compared with the back-propagation neural network (BPNN), support vector machine (SVM), and gradient boosting decision tree (GBDT). The findings indicate that the RF-NSGA-II framework can not only meet the requirements of SS and DS calculation, but also used as a support tool for real-time optimization and control of SCP.     

一种新的风电场复杂下垫面场地校准模型

针对现有方法模型简单、气象因素考虑不足和缺少异常数据检测等问题,提出一种基于相对密度离群因子（relative RDOF）异常检测和支持向量回归（SVR）组合的场地校准模型。首先,通过RDOF与四分位法对测风塔与风电机组距下垫面相同高度风速间的风速比进行异常数据检测,间接实现风速预处理;同时,根据测风塔与风电机组相同高度风切变序列的相关性,动态引入大气稳定度等级,与多种常规自然因素共同作为模型输入。然后,建立SVR场地校准模型修正风速。最后,通过算例验证,并与国际电工委员会（IEC）标准方法进行对比。结果表明,该方法可有效提升场地校准模型的风速修正精度。     

A combination of near-infrared hyperspectral imaging with two-dimensional correlation analysis for monitoring the content of biogenic amines in mutton

The study investigated the feasibility of using a combination of near-infrared hyperspectral imaging (NIR-HSI) with two-dimensional correlation (2DCOS) analysis for rapid and non-destructive determination of the content of biogenic amines in mutton during refrigerated storage. Total contents of biogenic amines (TBA) were used as the perturbation. By analysing the synchronous and asynchronous two-dimensional correlation spectra, sensitive variables that were closely related to TBA contents were obtained. The results showed that the wavelengths in the spectra range of 1002–1335 nm were the research area for the detection of TBA contents in mutton. The least-squares support vector machines (LSSVM) model based on effective wavelengths selected by competitive adaptive reweighted sampling (CARS) from 2DCOS analysis showed excellent results, with correlation coefficient in prediction (Rp) of 0.91, root mean square error in prediction (RMSEP) of 1.67 mg kg⁻¹ and the ratio of performance deviation (RPD) of 2.76. The research demonstrated that the combination of NIR-HSI and 2DCOS could be used as an effective method for monitoring the content of biogenic amines in mutton.     

基于多变量时间序列及向量自回归机器学习模型的水驱油藏产量预测方法

提出了一种基于多变量时间序列(MTS)及向量自回归(VAR)机器学习模型的水驱油藏产量预测方法,并进行了实例应用。该方法在井网分析的基础上通过MTS分析对注采井组数据进行优选,并将井组内不同采出井产油量及注入井注水量作为彼此相关的时间序列,通过建立VAR模型从多个时间序列中提取出相互作用规律,挖掘注采井间流量的依赖关系从而进行产量预测。水驱油藏历史生产数据分析结果表明,与数值模拟历史拟合结果相比,机器学习模型产量预测结果具有更高精度,同时不确定性分析提升了预测结果的安全性。通过脉冲响应分析对注入井的采油贡献量进行评价,可为注水开发方案调整提供理论指导。     

基于预训练语言模型的关键词感知问题生成

问题生成任务是指根据给定的文本段落和答案来自动生成对应的问题。针对现有问题生成方法存在的误差累积现象以及问题生成任务固有的“一对多”情况，提出一种带有关键词感知功能的问题生成方法。在预训练语言模型的基础上，实现关键词分类模型与问题生成模型的网络结构设计。输入文本段落中蕴含关键词，为使所生成的问题中包含同样的关键词以保证问题与段落的语义一致性，利用关键词分类模型提取出文本段落中的关键词，将关键词与非关键词的区分特征融入问题生成模型的输入中，该特征作为问题生成过程的全局信息，用以消除问题生成模型仅依赖局部最优解的弊端，减少误差累积与“一对多”情况的发生。在SQuAD数据集上的实验结果表明，该方法能够提升问题生成的质量，其BLEU-4指标值可达24，优于带有复制机制、带有语义监督的问题生成模型，目前已经借助百度百科数据平台实现了大规模工业应用。     

Predicting the mechanical properties of Cu–Al2O3 nanocomposites using machine learning and finite element simulation of indentation experiments

Micromechanics model, finite element (FE) simulation of microindentation and machine learning were deployed to predict the mechanical properties of Cu–Al₂O₃ nanocomposites. The micromechanical model was developed based on the rule of mixture and grain and grain boundary sizes evolution to predict the elastic modulus of the produced nanocomposites. Then, a FE model was developed to simulate the microindentation test. The input for the FE model was the elastic modulus that was computed using the micromechanics model and wide range of yield and tangent stresses values. Finally, the output load-displacement response from the FE model, the elastic modulus, the yield and tangent strengths used for the FE simulations, and the residual indentation depth were used to train the machine learning model (Random vector functional link network) for the prediction of the yield and tangent stresses of the produced nanocomposites. Cu–Al₂O₃ nanocomposites with different Al₂O₃ concentration were manufactured using insitu chemical method to validate the proposed model. After training the model, the microindentation experimental load-displacement curve for Cu–Al₂O₃ nanocomposites was fed to the machine learning model and the mechanical properties were obtained. The obtained mechanical properties were in very good agreement with the experimental ones achieving 0.99 coefficient of determination R2 for the yield strength.     

Design of data feature-driven 1D/2D convolutional neural networks classifier for recycling black plastic wastes through laser spectroscopy

In this study, two types of convolutional neural network (CNN) classifiers are designed to handle the problem of classifying black plastic wastes. In particular, the black plastic wastes have the property of absorbing laser light coming from spectrometer. Therefore, the classification of black plastic wastes remains still a challenging problem compared to classifying other colored plastic wastes using existing spectroscopy (i.e., NIR). When it comes the classification problem of black plastic wastes, effective classification techniques by the laser spectroscopy of Fourier Transform-Infrared Radiation (FT-IR) with Attenuated Total Reflectance (ATR) and Raman to analyze the classification problem of black plastic wastes are introduced. Due to the strong ability of extracting spatial features and remarkable performance in image classification, 1D and 2D CNN through data features are designed as classifiers. The technique of chemical peak points selection is considered to reduce data redundancy. Furthermore, through the selection of data features based on the extracted 1D data with peak points is introduced. Experimental results demonstrate that 2DCNN classifier designed with the help of 2D data feature selection as well as 1DCNN classifier shows the best performance compared with other reported methods for classifying black plastic wastes.     

Learning machines: Rationale and application in ground-level ozone prediction

Multilayer perceptron (MLP) and support vector machine (SVM), two popular learning machines, are increasingly being used as alternatives to classical statistical models for ground-level ozone prediction. However, employing learning machines without sufficient awareness about their limitations can lead to unsatisfactory results in modeling the ozone evolving mechanism, especially during ozone formation episodes. With the spirit of literature review and justification, this paper discusses, with respect to the concerning of ozone prediction, the recently developed algorithms/technologies for treating the most prominent model-performance-degradation limitations. MLP has the “black-box” property, i.e., it hardly provides physical explanation for the trained model, overfitting and local minima problems, and SVM has parameters identification and class imbalance problems. This commentary article aims to stress that the underlying philosophy of using learning machines is by no means as trivial as simply fitting models to the data because it causes difficulties, controversies or unresolved problems. This article also aims to serve as a reference point for further technical readings for experts in relevant fields.