SUPPORT VECTOR MACHINE FOR STRUCTURAL RELIABILITY ANALYSIS

Support vector machine (SVM) was introduced to analyze the reliability of the implicit performance function, which is difficult to implement by the classical methods such as the first order reliability method (FORM) and the Monte Carlo simulation (MCS). As a classification method where the underlying structural risk minimization inference rule is employed, SVM possesses excellent learning capacity with a small amount of information and good capability of generalization over the complete data. Hence, two approaches, i.e., SVM-based FORM and SVM-based MCS, were presented for the structural reliability analysis of the implicit limit state function. Compared to the conventional response surface method (RSM) and the artificial neural network (ANN), which are widely used to replace the implicit state function for alleviating the computation cost, the more important advantages of SVM are that it can approximate the implicit function with higher precision and better generalization under the small amount of information and avoid the "curse of dimensionality". The SVM-based reliability approaches can approximate the actual performance function over the complete sampling data with the decreased number of the implicit performance function analysis (usually finite element analysis), and the computational precision can satisfy the engineering requirement, which are demonstrated by illustrations.     

基于支持向量机替代模型的可靠性分析

建立了基于支持向量机回归算法和分类算法的替代模型可靠性分析方法,与蒙特卡罗法结合,采用拉丁超立方抽样技术,进行隐式极限状态函数的可靠度计算。讨论了相关参数对支持向量机模型性能的影响,并通过遗传算法进行参数优化,为支持向量机模型的参数选择提供了依据。研究了不同训练样本数量对支持向量机模型预测值精度的影响,进一步证实了支持...     

基于CCH的SVM几何算法及其应用

The support vector machine （SVM） is a novel machine learning tool in data mining. In this paper, the geometric approach based on the compressed convex hull （CCH） with a mathematical framework is introduced to solve SVM classification problems. Compared with the reduced convex hull （RCH）, CCH preserves the shape of geometric solids for data sets; meanwhile, it is easy to give the necessary and sufficient condition for determining its extreme points. As practical applications of CCH, spare and probabilistic speed-up geometric algorithms are developed. Results of numerical experiments show that the proposed algorithms can reduce kernel calculations and display nice performances.     

基于支持向量机的序列可靠性优化方法

在工程设计中,可靠性优化设计通常计算量较大或精度不够。本文提出了一种基于支持向量机（Support Vector Machine）和MPP（Most Probable Point）的可靠性分析方法。用SVM 在MPP处替代原极限状态函数,并利用极限状态函数的梯度信息,使SVM模型穿过M PP并与原函数相切,再基于SVM采用重要抽样法计算失效概率。然后,将SORA（Sequential Optimization and Reliability Assessment ）与基于SVM 的可靠性分析方法相集成,将传统的双循环可靠性优化算法解耦为单循环,并通过基于SVM 的可靠性分析方法修正了SORA中由于线性近似带来的误差,保证了最优设计点处可靠性分析的精度。算例证明,该方法在处理非线性问题时具有精确度高和计算量适度的特点。     

基于一类SVM的综合导航系统信息故障检测方法

为了提高舰艇综合导航系统的可靠性,并考虑到系统准确建模和大量故障数据获取的困难性,提出了一种基于一类支持向量机的信息故障检测方法。该方法主要包括两个过程:第一个过程是根据实测数据,并利用一类支持向量机的分类原理和主元分析法对导航信息进行离线建模;第二个过程是结合主元分析法将该模型应用到实时的信息故障检测中。该方法不依赖于系统模型而且只需要正常的小样本数据对模型进行训练,具有简便易于实现的优点。仿真试验表明,该方法对导航系统的硬故障和软故障都具有较好的检测能力和较短的检测延迟时间,而且该方法对径向基核函数参数的变化具有较低的敏感性,避免了复杂的调参过程。     

基于SVM回归的圆柱壳体抗脉冲强度预估模型

在小子样结构响应试验数据样本的基础上,利用支持向量机回归的方法模拟了圆柱壳体动态极限应变峰值同壳体几何尺寸和外加脉冲载荷大小的非线性函数关系,同时通过改进的模拟退火单纯形混合算法优化了支持向量机的性能参数,并将支持向量机回归分析的预测性能同BP人工神经网络方法做了比较,验证了具有优化性能参数组合的支持向量机在小样本条件下更好的预测和推广能力. 最后,从支持向量机回归模型导出了大尺寸圆柱壳体抗脉冲载荷的强度极限同自身几何尺寸的多元函数关系,从而为该类型壳体设备抗脉冲载荷的强度分析提供了一个可借鉴的预估模型. 研究结果表明了支持向量机在机械结构的强度预估和可靠性分析等力学领域具有广泛的应用前景.      

一种基于多尺度和改进支持向量机的光纤陀螺温度漂移建模与补偿方法

为了提升光纤陀螺温度漂移模型建模的准确性及补偿的效果,提出了一种基于改进支持向量机的多尺度建模和回归方法。首先分析了造成光纤陀螺温度漂移的关键因素,给出了建模的属性参数和温度试验。然后根据经验模态分解得到的本征模态函数排列熵的变化趋势,得出了回归精度和熵之间的变化关系,进而提出了基于信号分解的多尺度回归方法。为了提高上述多尺度回归算法的适应性,在传统支持向量机的基础上,提出了基于组合核函数的支持向量机回归算法,以适应不同特性的回归数据集。为了进一步提高回归精度,基于降低回归数据复杂度的分段回归思想,在上述多尺度回归的基础上提出了双-多尺度回归,并验证了方法的有效性。最后,将提出的算法以实际的光纤陀螺温度漂移数据进行验证,结果表明,相比于传统的支持向量机和反向传播神经网络具有更好的回归精度,温度漂移模型也更加精确,以均方误差指标为例,回归精度提升了两个数量级。     

基于网格搜索的支持向量机砂土液化预测模型

在使用支持向量机对砂土液化进行预测时,不可避免地会遇到惩罚因子和核函数参数如何选取的问题,目前还没有确定这两个参数的选择模式,解决这一问题比较常用的办法有经验公式法、遗传算法和网格搜索法.对此本文基于网格搜索方法,使用LIBSVM软件包,寻找砂土液化训练样本的结构风险最小值所对应的支持向量机最优参数;使用最优参数,建立了支持向量机砂土液化预测模型.研究结果表明:预测样本的正确率可达87.5%,而且预测结果稳定;同时通过比较BP神经网络的砂土液化预测情况,可知支持向量机有更好的泛化能力,收敛速度也更快.     

Hybrid GR-SVM for prediction of surface roughness in abrasive water jet machining

This paper presents a hybridization model of support vector machine (SVM) and grey relational analysis (GRA) in predicting surface roughness value of abrasive water jet (AWJ) machining process. The influential factors of five process parameters in AWJ, namely traverse speed, water jet pressure, standoff distance, abrasive grit size and abrasive flow rate, need to be analyzed using GRA approach. Then, the irrelevance factors of process parameters are eliminated. There is a need of determining the influential factors of process parameters to the surface roughness as to develop a robust prediction model. GRA acts as feature selection method in preprocessing process of hybrid grey relational-support vector machine (GR-SVM) prediction model. Efficiency of the proposed model is demonstrated. GR-SVM presents more accurate result than conventional SVM as it removes the redundant features and irrelevant element from the experimental datasets.     

结构可靠性分析的支持向量机响应面法

针对隐式极限状态可靠性分析问题,提出了一种支持向量机响应面法,该方法采用了与经典响应面法类似的迭代思想,由支持向量回归机替代经典响应面法中的固定多项式函数来构建响应面,并结合一次二阶矩法形成迭代过程.在此基础上,还对训练样本提出了一种改进的选取方法,从而进一步提高方法的效率.文中将所提方法与多种经典可靠性分析方法的计算结果对比分析,改进的支持向量机响应面法精度较高,调用结构分析程序的次数最少.     

基于SVM的结构可靠度分析响应面方法

响应面法是解决隐式极限状态方程结构可靠度分析问题比较理想的方法,其关键问题是响应面函数的重构。根据响应面方法经验点集的小样本特点,利用支持向量机(SVM)对小样本数据良好的学习和泛化能力,用SVM重构结构响应面方程,建立了基于SVM的隐式极限状态方程结构可靠度分析的响应面方法。在此基础上,文中提出了改进SVM响应面方法,改进的方法充分利用每次有限元计算成果,大幅减少了有限元计算次数。算例表明本文方法具有很好的计算精度和计算效率。     

Predicting the critical heat flux in concentric-tube open thermosiphon: a method based on support vector machine optimized by chaotic particle swarm optimization algorithm

This study presents a method based on support vector machine (SVM) optimized by chaotic particle swarm optimization algorithm (CPSO) for the prediction of the critical heat flux (CHF) in concentric-tube open thermosiphon. In this process, the parameters C, ε and δ² of SVM have been determined by the CPSO. As for a comparision, the traditional back propagation neural network (BPNN), radial basis function neural network (RBFNN), general regression neural network (GRNN) are also used to predict the CHF for the same experimental results under a variety of operating conditions. The MER and RMSE of SVM–CPSO model are about 45% of the BPNN model, about 60% of the RBFNN model, and about 80% of GRNN model. The simulation results demonstrate that the SVM–CPSO method can get better accuracy.     

基于更新支持向量的大跨度斜拉桥体系可靠度分析

针对斜拉桥静力体系可靠度分析中隐式功能函数重构和繁杂失效路径的特点,提出了一种基于更新支持向量的体系可靠度分析方法,将传统的用于构件可靠度分析的支持向量机(SVM)改进并应用于斜拉桥体系可靠度分析。该方法主要有4个步骤:首先通过构件的敏感分析识别斜拉桥的主要失效路径;其次采用最小二乘支持向量机(LS-SVM)对斜拉桥的隐式功能函数进行重构,并通过Monte-Carlo抽样得出构件的可靠指标;然后根据已经更新的有限元模型对支持向量进行更新,得出相关构件失效后的剩余构件的条件可靠指标;最后由结构体系的失效树和串并联关系得出斜拉桥的体系可靠度。主跨为420m的混凝土斜拉桥算例分析表明了上述算法的有效性和实用性,同时也获得了该斜拉桥的主要失效路径并识别了影响其体系可靠度的主要构件。     

基于支持向量机回归的结构系统可靠性及灵敏度分析方法

提出了一种基于支持向量机回归近似极限状态方程的系统可靠性分析方法,所提方法首先由支持向量机拟合系统各失效模式的极限状态方程,将复杂或隐式极限状态方程近似等价为显式极限状态方程,然后根据系统各个失效模式的逻辑结构,由高精度的显式极限状态方程方法计算系统的失效概率和参数灵敏度.与线性展开和响应面法近似极限状态方程相比,文中方法由于采用了基于结构风险最小化原理的支持向量机回归,因而在拟合非线性极限状态方程上表现优越,计算精度高.与直接蒙特卡洛模拟相比,由于该方法采用较少的样本即可近似出概率等价的显式极限状态方程,因而计算效率大幅提高.工程实例表明:所提方法可以处理串联、并联和混合系统的可靠性与可靠性灵敏度分析,具有工程运用价值.     

Parameter identification of systems with multiple disproportional local nonlinearities

Identification of nonlinear systems, especially with multiple local nonlinearities exhibiting disproportional ratios of the degree of nonlinearity and present at a single or multiple spatial locations, is a highly challenging inverse problem. Identification of such complex nonlinear systems cannot be handled easily by the existing conventional restoring force or describing function methods. Further, noise-corrupted measured time history responses make the parameter identification process much more difficult. Keeping this in view, we propose a new meta support vector machine (meta-SVM) model to precisely identify the type, spatial location(s) and also the nonlinear parameters present in disproportionate levels using the noisy measurements. Apart from the conventional SVM model, we also explore the effectiveness of the non-batch processing models like incremental learning for lesser computational cost and increased efficiency. Both incremental and conventional support vector regression models are explored to precisely identify the nonlinear parameters. A numerically simulated multi-degree of freedom spring-mass system with limited multiple local nonlinearities at a few selected spatial locations is considered to illustrate the proposed meta-SVM model for nonlinear parametric identification. However, the extension of the proposed meta-SVM model is rather straightforward to include all types of nonlinearities and cases with the simultaneous existence of multiple numbers of same or different nonlinearities (i.e. combined nonlinearities) at single or multiple locations. It is also clearly established from the numerical simulation studies that the proposed incremental meta-SVM model paves way for online real-time identification of nonlinear parameters which is not yet been addressed in the existing literature.

     

A note of hybrid GR-SVM for prediction of surface roughness in abrasive water jet machining

The paper attempts at reviewing a previous research entitled “Hybrid GR-SVM for prediction of surface roughness in abrasive water jet (AWJ) machining” and some problems were found in its reducing parameters process and prediction model. The authors presented a hybrid of grey relation analysis (GRA) and support vector machine (SVM) to estimate the roughness surface in a certain dataset of AWJ machining. Their proposed method may not work in real case of AWJ machining as it is claimed. This paper gives a counter model in order to illustrate these remarks.     

基于支持向量机的导管泄爆容器压力峰值预测

为了预测导管泄爆容器压力峰值,根据文献提取出影响导管泄爆容器压力峰值的因素,将这些因素作为输入变量,采用支持向量机算法对压力峰值与各因素的内在关系进行了研究,建立导管泄爆容器压力峰值预测模型,对模型的有效性及预测能力进行了验证。将预测模型与现有经验公式进行比较,表明支持向量机模型具有较好的预测能力,且预测能力优于经验公式。     

Composite feedforward-feedback controller for generator excitation system

This paper proposes a composite feedforward-feedback controller for a generator excitation system. By integrating both feedforward and feedback compensation, the composite controller composes a feedforward approximate controller and an error-feedback proportional-integral-derivative (PID) compensator. The feedforward controller is derived based on an approximation model method, which not only avoids complex computation, but also avoids online learning or adjustment. Only a general support vector machine (SVM) modeling technique is involved in a feedforward controller implement. Meanwhile, the feedback controller is constructed using a PID controller. Then the attenuation of both disturbances and estimated errors is guaranteed. Several simulations illustrate the effectiveness of the composite excitation controller.     

基于支持向量机的结构损伤分步识别研究

支持向量机是一种基于统计学习理论的机器学习算法,能够较好地解决小样本的学习问题。本文介绍了支持向量机分类和回归算法,提出了基于支持向量机的结构损伤分步识别方法:以模态频率作为损伤特征,首先根据支持向量机分类算法的概率估计确定可能的损伤位置,重新构造训练样本,然后利用支持向量机回归算法计算损伤位置;最后估计损伤程度。以梁的损伤识别为例进行了验证,结果表明该方法可以提高损伤识别的精度。     

基于高斯过程分类的结构贝叶斯可靠性分析

贝叶斯可靠性方法是处理不完备信息条件下结构可靠性问题的有效途径之一。在实际应用中,由于可靠性分析的计算量较大,常须采用各种近似替代模型以提高计算效率。传统的替代模型方法是对结构的功能函数予以近似建模。这种方法不易定量考虑模型误差对可靠性分析的影响,且难以应用于诸如功能函数不连续和失效域不连通等情况。为此,本文提出一种基于高斯过程分类的替代模型,直接辨识结构的极限状态曲面,并将其应用于结构贝叶斯可靠性分析之中。分析了替代模型不确定性对可靠性预测结果的影响,给出了失效概率分布参数的方差算式,进而提出了改善模型精度的补充采样准则。通过算例验证了方法的适用性和有被性．