Sensor placement optimization in the artificial lateral line using optimal weight analysis combining feature distance and variance evaluation

Artificial lateral line is a multi-sensor system, mimicking the lateral line of fish to perceive the parameters of flow field. However, it can easily lead to information loss or redundancy with limited number of sensors due to unsuitable sensor placement. An optimal weight analysis algorithm is proposed to solve the problem on sensor placement of robotic fish. Firstly, signal features are extracted from the pressure data, which are collected from candidate sensor locations in different conditions. Then the improved distance evaluation is used to assess each feature, and the feature distance factor is regarded as the weight for distinguishing. Combined with the analysis of variance, the contribution vector of sensor locations is obtained. Three indexes selected by the algorithm are introduced to compare the sensor subsets. The results in both simulation and experiment show the effectiveness of the algorithm. The optimal number of sensors on the robotic fish is also studied.     

Influence of time and length size feature selections for human activity sequences recognition

In this paper, Viterbi algorithm based on a hidden Markov model is applied to recognize activity sequences from observed sensors events. Alternative features selections of time feature values of sensors events and activity length size feature values are tested, respectively, and then the results of activity sequences recognition performances of Viterbi algorithm are evaluated. The results show that the selection of larger time feature values of sensor events and/or smaller activity length size feature values will generate relatively better results on the activity sequences recognition performances.     

Bayesian-inference-based neural networks for tool wear estimation

This paper introduces the application of neural networks based on Bayesian inference, the automatic relevance determination algorithm for selecting relevant features and designing neural estimators for tool wear estimation in face-milling processes. Two types of neural networks are studied: Bayesian support vector machines for regression (BSVR) and Bayesian multilayer perceptrons (BMLP). Sixteen well-known features derived from pre-processing of the milling force signal are considered. The force signal samples are collected from 20 milling experiments under various machining conditions. The feature extraction and selection procedure is then applied to the sampled force signal. The feature selection results from the two neural networks are found to be quite similar. The average force has been proven to be the most relevant feature for tool wear estimation in both cases, among a set of six other features in each case, with each set differing by only one feature. The comparison among the generalization capabilities of the entire, selected, and rejected features shows that the selected features are relatively more relevant to tool wear processes in both cases. The comparison between the estimation results from the two neural networks using the corresponding relevant feature set shows that the BSVR method is more accurate in estimating flank wear than BMLP, but at the cost of a higher computing load.     

基于变分贝叶斯平行因子分解的缺失信号的恢复

现有的工程信号处理方法都是基于完整的数据采集,并没有考虑缺失信号的处理。而在工程实际中,由于人为因素和自然界不可抗拒的因素,有时会造成传感器失效,从而造成信号采集的缺失。为了消除信号缺失对工程信号处理的消极影响,提出了一种基于变分贝叶斯平行因子分解的信号恢复方法。首先利用平行因子分析理论将采集的振动信号构造成三维张量,同时结合贝叶斯方法,引入潜在变量和超参数,建立贝叶斯平行因子概率图模型;其次采用变分贝叶斯算法推导出因子矩阵和超参数的后验分布,从而进一步推断出缺失元素的分布预测;最后通过分析该模型的下界,初始化参数的选择,使该算法更好的解决信号缺失问题。利用均方根误差和相对平方根误差对该算法的性能进行评估,仿真和实验结果表明,随着缺失比例的增大,变分贝叶斯平行因子分解算法相较于传统的低秩张量补全算法,误差更小,能够更加有效的恢复缺失的信号,有效地解决了工程信号处理中因传感器失效而引起的信号缺失的问题。     

结合特征在线选择与协方差矩阵的压缩跟踪算法

本文从3个方面对原始压缩跟踪算法进行改进,以提高其在复杂场景下的鲁棒性和准确性。首先,提出一种结合特征在线选择的压缩跟踪算法,通过计算相邻两帧同维特征所服从的高斯分布曲线的Hellinger距离来度量特征的置信水平,从特征池中选择置信水平较高的特征,并融合特征的置信水平构造贝叶斯分类器。然后,在压缩跟踪框架下引入协方差矩阵以增强算法对目标的表达能力,把Haar-like特征和协方差矩阵相结合构建目标模型,取最大响应值所对应的候选样本作为跟踪结果。最后,优化分类器参数的更新方式,根据目标模板与跟踪结果的相似度来自适应更新分类器参数。改进算法的平均跟踪成功率比原算法提高了25%,平均跟踪精度比原算法提高了22%。相比于原始压缩跟踪算法,本文算法具有更高的跟踪鲁棒性和准确性。     

Fast genetic algorithm for roundness evaluation by the minimum zone tolerance (MZT) method

According to ISO 1101, “A geometrical tolerance applied to a feature defines the tolerance zone within which that feature shall be contained”.The main goal of the minimum zone tolerance (MZT) method is to achieve the best estimation of the roundness error, but it is computationally intensive. This paper describes the application of a genetic algorithm (GA) to minimize the computation time in the evaluation of CMM roundness errors of a large cloud of sampled points.Computational experiments have shown that by selecting the optimal GA parameters, namely a combination of the five genetic parameters related to population size, crossover, mutation, stop condition, and search space, the computation time can be reduced by up to one order of magnitude, allowing real-time operation.Optimization has been tested using seven CMM samples, obtained from different machining features. The performance of the optimized algorithm has been validated using four benchmark samples from the literature and with certified samples.     

发动机异常检测多目标优化方法

针对柴油发动机异常检测中的特征选择和分类器参数与检测精度之间的耦合关系,提出了一种基于非支配排序粒子群优化的柴油发动机异常检测封装式多目标同步优化方法.利用双树复小波包的分解与重构,对发动机振动信号进行时域、频域和时频域多角度特征提取,构建了较完备的特征参数集,分析了故障诊断中特征选择与分类器参数优化对检测精度的影响,运用非支配排序粒子群优化算法对多个优化目标进行协调和折衷处理,同时追求特征参数子集维数最小化和分类正确率最大化.实验数据分析表明,该方法能够寻找出最优的特征子集和分类器参数,提高柴油发动机异常检测的精度和效率.     

NEW FEATURE SELECTION METHOD IN MACHINE FAULT DIAGNOSIS

Aiming to deficiency of the filter and wrapper feature selection methods, a new method based on composite method of filter and wrapper method is proposed. First the method filters original features to form a feature subset which can meet classification correctness rate, then applies wrapper feature selection method select optimal feature subset. A successful technique for solving optimization problems is given by genetic algorithm (GA). GA is applied to the problem of optimal feature selection. The composite method saves computing time several times of the wrapper method with holding the classification accuracy in data simulation and experiment on bearing fault feature selection. So this method possesses excellent optimization property, can save more selection time, and has the characteristics of high accuracy and high efficiency.     

基于改进粒子群算法的应变传感器优化布置

在结构健康监测和损伤识别研究中,为了应用有限的试验设备资源获取尽可能多的有效测试信息,快速有效地解决应变传感器的优化配置问题,提出了一种基于克隆选择和离散粒子群混合算法优化新型适应度函数的应变传感器优化布置方法,并将该方法应用到拉西瓦拱坝上。结果表明,基于改进克隆选择和离散粒子群混合算法具有更强的全局寻优能力,且提出的应变类适应度函数在保证应变模态正交性和模态应变能方面更有优势。该方法能很好地识别拱坝的应变振型,可在各类结构的模态测试和损伤识别研究中进行推广。     

基于CPSO与LSSVM融合的发酵过程软测量建模

发酵过程是一个复杂的时变、非线性、强耦合过程.发酵过程中的关键参量菌体浓度通常难以用传统物理传感器实时在线检测.为了测量该参数,将CPSO算法与LSSVM相结合构建发酵过程软测量模型.模型采用CPSO算法优化LSSVM软测量模型参数,克服了常规交叉验证法选取参数的耗时和盲目性.仿真结果表明,CPSO-LSSVM软测量模型较LSSVM软测量模型更能在较短的时间内获得较高的收敛精度,其平均误差为2.05％,说明该软测量模型可用于发酵过程不可在线测量的菌体浓度的实时在线软测量,并且预测精度高,预测速度快,预测能力强.该软测量建模方法也为发酵过程其他关键参量的实时在线测量提供了新的途径.     

复合基质结构优化设计下的应变测量灵敏度提高研究

针对当前微带天线传感器应变测量灵敏度较低的问题,对基质结构进行研究,设计了一种带空气层的复合基质天线传感器.结合遗传算法应用HFSS软件对传感器结构参数进行优化设计,仿真研究最优解下的复合基质天线传感器"谐振频率-应变"关系.对所设计的复合基质天线传感器进行拉伸对比试验,其应变测量灵敏度是传统微带贴片天线传感器的3.1...     

A roller bearing fault diagnosis method based on the improved ITD and RRVPMCD

Targeting that the measured vibration signal of roller bearing contains the characteristics of non-stationary and nonlinear, and the extraction features may contain smaller correlation and redundancy characteristics in the roller bearing fault diagnosis, the vibration signal processing method based upon improved ITD (intrinsic time-scale decomposition) and feature selection method based on Wrapper mode are put forward. In addition, in the design of the classifier, targeting the limitation of existing pattern recognition method, a new pattern recognition method-variable predictive model based class discriminate (VPMCD) is introduced into roller bearing fault identification. However, the parameters are fitted by using least squares in VPMCD method, while least squares regression is sensitive to “abnormal value”. Therefore, a robust regression-variable predictive mode-based class discriminate (RRVPMCD) method is proposed in this paper, robust regression is adopted to estimate parameters and the effect of “abnormal value” in the estimation of parameters would be reduced by giving each feature a weight. Firstly, improved ITD method and feature selection method based on Wrapper mode are combined to extract the fault features of roller bearing vibration signals, and feature vector matrixes are established, then a predictive model is built through the method of RRVPMCD, finally, the established predictive model is used for pattern recognition. Experimental results show that the model based on the improved ITD, the Wrapper feature selection and RRVPMCD method can effectively identify work status and fault type of roller bearing.     

Fault diagnosis of rotary kiln using SVM and binary ACO

This paper proposes a novel hybrid algorithm for fault diagnosis of rotary kiln based on a binary ant colony (BACO) and support vector machine (SVM). The algorithm can find a subset selection which is attained through the elimination of the features that produce noise or are strictly correlated with other already selected features. The BACO algorithm can improve classification accuracy with an appropriate feature subset and optimal parameters of SVM. The proposed algorithm is easily implemented and because of use of a simple filter in that, its computational complexity is very low. The performance of the proposed algorithm is evaluated through two real Rotary Cement kiln datasets. The results show that our algorithm outperforms existing algorithms.     

Effect of geometrical parameters on high frequency performance of Rogowski coil for partial discharge measurements

Partial discharges (PD) in a power system component is a sever threat, indicating a high likelihood of the imminent and complete failure of the insulation. The reliability of partial discharge diagnostics depends upon the design and accuracy of the measuring sensors. In this paper the Rogowski coil is employed as a PD measuring sensor. The selection of suitable geometrical parameters has been identified as an important aspect regarding the proper operation and installation of the coil around the under test power component. These parameters significantly affect the measuring performance of the coil in terms of its sensitivity and bandwidth. The measurement performance of different geometrical designs for the high frequency Rogowski coils is evaluated. The variation in parameters such as core and coil diameter, diameter of the copper wire used for winding and number of turns, has been experimentally investigated. In addition the return winding, as a non-conventional method of creating a return loop for the Rogowski coil, is compared with a return wire (loop) to analyze sensor performance. The comparative study of variation in mechanical design features provides a brief guideline to select the optimal design of the coil.     

A data driven approach for condition monitoring of wind turbine blade using vibration signals through best-first tree algorithm and functional trees algorithm: A comparative study

Wind energy is one of the important renewable energy resources available in nature. It is one of the major resources for production of energy because of its dependability due to the development of the technology and relatively low cost. Wind energy is converted into electrical energy using rotating blades. Due to environmental conditions and large structure, the blades are subjected to various vibration forces that may cause damage to the blades. This leads to a liability in energy production and turbine shutdown. The downtime can be reduced when the blades are diagnosed continuously using structural health condition monitoring. These are considered as a pattern recognition problem which consists of three phases namely, feature extraction, feature selection, and feature classification. In this study, statistical features were extracted from vibration signals, feature selection was carried out using a J48 decision tree algorithm and feature classification was performed using best-first tree algorithm and functional trees algorithm. The better algorithm is suggested for fault diagnosis of wind turbine blade.     

Integral DOE and MANOVA techniques for classification feature selection: using solder joint defects as an example

This study presents a novel approach for feature selection using an integrated DOE and MANOVA technique to classify solder joint defects for print circuit boards (PCBs). The main selection procedure includes three stages. The first stage adopts a single feature variable selection algorithm to eliminate poorly discriminated feature variables. The second stage, Plackett-Burman (PB) resolution III design, is then constructed to select the remaining feature variables. The MANOVA technique is then used to calculate the Pillai statistic as the response to the PB design of experiment, and statistical analysis is then executed to obtain the optimal multiple feature variables for multiple groups. The discriminate function classifier is used to evaluate the classification results. The experimental analysis results show that the proposed analysis procedure can acquire an optimum subset of features for classification.     

Intelligent fault diagnosis of rolling bearing based on kernel neighborhood rough sets and statistical features

Intelligent fault diagnosis benefits from efficient feature selection. Neighborhood rough sets are effective in feature selection. However, determining the neighborhood value accurately remains a challenge. The wrapper feature selection algorithm is designed by combining the kernel method and neighborhood rough sets to self-adaptively select sensitive features. The combination effectively solves the shortcomings in selecting the neighborhood value in the previous application process. The statistical features of time and frequency domains are used to describe the characteristic of the rolling bearing to make the intelligent fault diagnosis approach work. Three classification algorithms, namely, classification and regression tree (CART), commercial version 4.5 (C4.5), and radial basis function support vector machines (RBFSVM), are used to test UCI datasets and 10 fault datasets of rolling bearing. The results indicate that the diagnostic approach presented could effectively select the sensitive fault features and simultaneously identify the type and degree of the fault.     

Fault classification on vibration data with wavelet based feature selection scheme

Fault classification based upon vibration measurements is an essential building block of a conditional based health usage monitoring system. Multiple sensors are incorporated to assure the redundancy and to achieve the desired reliability and accuracy. The shortcoming of using multiple sensors is the need to deal with a high dimensional feature set, a computationally expensive task in classification. It is vital to reduce the feature dimension via an effective feature extraction and feature selection algorithm. A simple wavelet based feature selection scheme is proposed herein, uniquely built by local discriminant bases and genetic optimization. This scheme overcomes the disadvantages faced by the existing feature selection methods by producing a generic feature set, reducing the dimensionality of features, and requiring no prior information of the problem domain. The proposed feature selection scheme is based upon the strategy of "divide and conquer" that significantly reduce the computation time without compromising the classification performance. The simulation results show the proposed feature selection scheme provides at least 65% reduction of the total number of features at no cost of the classification accuracy.     

Optimal feature selection using genetic algorithm for mechanical fault detection of induction motor

Time-domain vibration signals are measured in all horizontal, axial, and vertical directions for induction motor mechanical fault diagnostics. Many features are extracted from these signals. The problem is how to find the good features among the feature set in order to receive reliable classifications. Based on specific distance criteria, a genetic algorithm (GA) is introduced to reduce the number of features by selecting optimized ones for fault classification purpose. A decision tree and multi-class support vector machine are used to illustrate the potentiality and efficiency of this selection method. Comparisons show that the diagnostic systems after selecting specific features perform better than the original system.     

Data correlation analysis for optimal sensor placement using a bond energy algorithm

Optimal sensor placement is one of the crucial and fundamental factors for constructing a cost-effective structural health monitoring system and is related to the effective evaluation of the state of the structure. Structural responses are correlated to some extent, as the structural behavior is continuous. Based on the above two considerations, the question arises of how to obtain the maximum amount of information for understanding the structure using measurements from limited sensors and not be limited to direct monitoring at the placements where the limited sensors are located. Data correlation analysis for optimal sensor placement is proposed using a bond energy algorithm, in which the objectives, such as structural response evaluation covering the maximum structural responses using measurements from sensors located at the optimal placements, are taken into account. The data correlation analysis is conducted for the structural responses, and the correlation matrix is established. Furthermore, the optimal sensor placements and the correlation of the responses at element locations can be determined using the bond energy algorithm. A Schwedler single-layer spherical lattice dome-like structure, which is a common large space steel structure, is used to simulate the structural responses and verify the effectiveness of the proposed method by discussion of different scenarios of parameter selection.