Table of Content

  

20 October 2017
Volume 47 Issue 5

Adaptive fault-tolerant containment control for multi-agent systems with unknown global information

YE Dan, ZHANG Tianyu, LI Kui

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  1-6.  doi:10.6040/j.issn.1672-3961.0.2017.168

Abstract ( 1295 )   PDF (1411KB) ( 446 )   Save

Related Articles | Metrics

This paper studied the containment control problem for the multi-agent systems(MASs)with actuator faults, unmatched uncertainty, external disturbance, nonlinear dynamic and unknown global information. The considered actuator failures included loss of effectiveness, stuck and outage. The upper bounds of uncertainty, disturbance and faults were combined and estimated online, whose effects were further compensated by a designed control mechanism. To avoid using the global information, adaptive technique was used to estimate the coupling gain of the controller. The proposed containment controller guaranteed the tracking error converge to zero asymptotically, which implied the followers were driven into the convex hull spanned by the leaders. Finally, a simulation example was given to illustrate the effectiveness of the distributed fault-tolerant containment controller.

A fault filtering method based on an improved extended state filter for nonlinear system

LI Wei, WANG Kehong, CAO Huichao

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  7-14.  doi:10.6040/j.issn.1672-3961.0.2017.173

Abstract ( 1375 )   PDF (3093KB) ( 529 )   Save

References | Related Articles | Metrics

Aiming at the problem of the linearization error and filtering divergence for traditional fault diagnosis method of nonlinear system, a new fault diagnosis method was studied for a class nonlinear system with the measurement noise and actuator fault of time-varying. A new extended state including the nonlinear dynamic and fault of the system was constructed, and a class of the filter containing expansion stateis was constructed for fault diagnosis. The sufficient conditions of the augmented system were presented, meanwhile, the design method of the fault diagnosis filter with the robust H_∞ performance was given. And the filter and diagnosis of the fault could be implemented by the fault isolation for the new extended state, among the known nonlinear dynamic of the original system. Based on the Van der pol oscillator that was typical nonlinear system with measurement noise, the simulation study for the constant value and time-varying failure condition was performed. The results showed that the proposed method could better solve noise filtering and the fault diagnosis problem of nonlinear systems.

Fault diagnosis for manipulators based on Spiking neural networks

WANG Xiuqing, ZENG Hui, XIE Fei, LYU Feng

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  15-21.  doi:10.6040/j.issn.1672-3961.0.2017.165

Abstract ( 1394 )   PDF (3318KB) ( 328 )   Save

Related Articles | Metrics

Because spiking neural networks(SNNs)could convey both temporal and spatial information at the same time, and contain features that were more attractive than those of traditional neural networks(NNs), SNNs were more suitable for analyzing the dynamic time-series signals. A novel fault diagnosis method based on SNNs was proposed to distinguish manipulators' collision and obstruction failure states from the normal working state,as manipulators approaching the grasping position. The architecture of the SNNs for fault diagnosis was discussed, and the results for SNNs fault diagnosis methods with different SNNs' topologic structures and parameters were compared. Experimental results showed that the proposed fault diagnosis method based on SNNs was effective and helpful for manipulators' fault diagnosis, which was also important for manufacture industries' smooth and safe running.

Failure prognosis method based on evidential reasoning for aerospace relay

ZHOU Zhijie, ZHAO Fujun, HU Changhua, WANG Li, FENG Zhichao, LIU Taoyuan

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  22-29.  doi:10.6040/j.issn.1672-3961.0.2017.211

Abstract ( 1399 )   PDF (2811KB) ( 783 )   Save

References | Related Articles | Metrics

To solve the high failure ratio problem of aerospace relay and strong uncertainty in its failure prognosis, a failure prognosis method based on evidential reasoning(ER)by fusing multiple fault characteristics information was proposed. The JRC-7M aerospace relay was chosen as the research object and its characteristic parameters, super-path time and pick-up time, were chosen as the main fault characteristics. In the proposed method, a forecasting model based on the third-order Volterra filter was proposed to online predict the fault characteristics' information, then an adaptive weighting model based on coefficient of variation-based weighting was adopted to calculate the relative weight. To obtain a comprehensive failure prognosis result of the aerospace relay, an safety assessment aggregation scheme based on the ER approach was developed to fuse multiple fault characteristics, and the “history”, “current” and “future” fault state information were synthetically fused. The validity of the proposed method was verified by the testing data collected by the STS2104A electromagnetic relay test system.

Early diagnosis and life prognosis for slowlyvarying fault based on deep learning

ZHOU Funa, GAO Yulin, WANG Jiayu, WEN Chenglin

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  30-37.  doi:10.6040/j.issn.1672-3961.0.2017.193

Abstract ( 2277 )   PDF (6482KB) ( 1290 )   Save

References | Related Articles | Metrics

In order to overcome the shortcoming of traditional early fault diagnosis methods, a method of combining deep learning with PCA to realize early diagnosis of slowly varying small fault and life prognosis was proposed. Using the deep learning method to extract the sampled data characteristics layer by layer, learning the early fault characteristics and establishing the early fault diagnosis model for slowly varying small fault, the combining deep learning was combined with PCA to integrate the high dimensional fault feature vector extracted by the deep learning into a fault characteristic variable. A data-driven fault precursor could be defined according to the evolution rule of the characteristic variable of the historical fault data, and life prognosis model was established by exponential nonlinear fitting method. The TE benchmark data was used to verify the effectiveness of the proposed algorithm, experimental results showed the validity of the proposed algorithm by comparing with other algorithms.

Effects of multiplicative actuator faults on the fault diagnosis performance in open-loop and closed-loop systems

LIU Yang

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  38-43.  doi:10.6040/j.issn.1672-3961.0.2017.182

Abstract ( 1878 )   PDF (1321KB) ( 841 )   Save

References | Related Articles | Metrics

The effects of multiplicative actuator faults on the fault diagnosis performance in both open-loop and closed-loop systems were investigated. Practical systems were usually subject to control inputs owing to the stability and robustness requirements. When multiplicative actuator faults were encountered, control signals could not be completely decoupled from the residual signals and the faults would unavoidably influence the fault diagnosis performance. The actuator faults were modeled for linear discrete systems and the Luenburger observer was used to generate the residual signals. By studying the one-step state transfer equation and the triangle inequality, the dynamics of the estimation error and residual was discussed and the conditions under which the open-loop and closed-loop residuals were bounded were established which were dependent on the amplitudes of the possible faults. Some numerical examples were provided to show the effectiveness of the proposed analysis for actuator faults with different amplitudes.

A fault detection method based on modified canonical correlation analysis

CHEN Zhiwen, PENG Tao, YANG Chunhua , HE Zhangming, YANG Chao, YANG Xiaoyue

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  44-50.  doi:10.6040/j.issn.1672-3961.0.2017.171

Abstract ( 1556 )   PDF (1953KB) ( 804 )   Save

References | Related Articles | Metrics

In order to improve the effectiveness of the fault detection(FD)method based on standard canonical correlation analysis(CCA), the original residual generation was modified. By analyzing the statistical characteristics of the residual signal and changing the residual generation mode, the improved residual generation method did not depend on the selection of the number of principal components, so that the fault detection performance would be free of such a selection. The proposed method was further applied to the Tennessee Eastman benchmark process, in which four typical faults were simulated. The achieved results showed that the proposed method could successfully detect the faults. Due to the different fault sensitivity of the two test statistics, it could be found that the fault detectability of the two test statistics were different.

Intermittent fault detection method based on sparse representation

YANG Rui

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  51-56.  doi:10.6040/j.issn.1672-3961.0.2017.238

Abstract ( 1173 )   PDF (3780KB) ( 614 )   Save

References | Related Articles | Metrics

Based on the sparsity of intermittent faults in some domains, an intermittent fault detection method based on sparse representation was proposed. The system output data were used to build the overcomplete dictionary and design the fault detection threshold of intermittent fault, which was able to update the over-complete dictionary and fault detection threshold with online measurements. With the simulation verification, the proposed method was suitable for intermittent fault detection in dynamic system and results under different online updating strategies were compared.

Fault-tolerant control of autonomous underwater vehicle based on adaptive region tracking

CHU Zhenzhong, ZHU Daqi

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  57-63.  doi:10.6040/j.issn.1672-3961.0.2017.210

Abstract ( 1594 )   PDF (1108KB) ( 687 )   Save

References | Related Articles | Metrics

An adaptive region tracking fault-tolerant control for the thrusters of autonomous underwater vehicle was proposed. Different from the traditional fault-tolerant control methods of autonomous underwater vehicle, the region tracking control theory was adopted, and the control target was designed as a spatial region. For the uncertainty and thruster fault in the system, the neural network was used to identify them online. Considering the problem of the divergence of neural network caused by the thrust saturation during the thruster fault, a neural network weight adjustment method based on a saturation factor was proposed. The effectiveness of the proposed method was verified by simulation.

Sensor fault tolerant switch strategy for multi-motor synchronous system based on ADRC

MAO Haijie, LI Wei, WANG Kehong, FENG Xiaolin

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  64-70.  doi:10.6040/j.issn.1672-3961.0.2017.167

Abstract ( 1261 )   PDF (1866KB) ( 964 )   Save

References | Related Articles | Metrics

Focusing on multi-motor synchronous control system which was susceptible to variable load and random disturbance, an active disturbances rejection control(ADRC)technique was used to improve the synchronous performance. Based on the principle of sensorless vector control, an extended state observer(ESO)was used in the current loop to obtain the estimated speed with observation of its extended state. The residual signal between actual speed and estimation speed was analyzed to realize fault diagnosis. In order to ensure the safe operation of the fault system, a flexible adaptive switching strategy based on fault diagnosing evaluation of reliability was proposed to ensure smooth transition in case of requiring. Simulation results showed the effectiveness of the proposed method.

Fault tolerant estimation for a class of networked systems with sensor faults

ZHAO Yinghong, HE Xiao, ZHOU Donghua

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  71-78.  doi:10.6040/j.issn.1672-3961.0.2017.254

Abstract ( 1084 )   PDF (2217KB) ( 297 )   Save

Related Articles | Metrics

A new approach to fault tolerant estimation problem was derived for a class of closed-loop networked systems with sensor faults. Closed-loop networked systems implied that there could be network-induced phenomena of random delays and packet dropouts in both the sensor-to-estimator(S-E)channel and the controller-to-actuator(C-A)channel. An indicator function driven individually by two groups of random variables was introduced to characterize the network-induced phenomena in both channels. Then, sensor faults were considered in the estimator design process. A kind of fault tolerant estimator was developed in the unbiased minimum variance sense in order to deal with sensor faults in networked systems with closed loop. Simulations were carried out on a practical networked permanent magnet synchronous motor driving system to test the feasibility and effectiveness of the proposed techniques.

A new distributed formation for multi-agent systems with constant time delays

QIN Liguo, HE Xiao, ZHOU Donghua

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  79-88.  doi:10.6040/j.issn.1672-3961.0.2017.255

Abstract ( 1213 )   PDF (3479KB) ( 744 )   Save

References | Related Articles | Metrics

A new delay independent distributed formation control law was presented for a network of second-order integrators subject to constant time delays. A new augment variable which represented the integration of distributed formation errors was introduced to improve the robustness of the formation control law. Different from current distributed control laws, the presented control law was robust to some constant bias faults. A condition on the existence of the delay independent formation control law was proposed by using Nyquist stability criterion. The simulation demonstrated the effectiveness of the control law.

Fault detection of unmanned aerial vehicle flight control system based on optimal estimation of the L2-norm

CHEN Jie, ZHONG Maiying, ZHANG Ligang

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  89-95.  doi:10.6040/j.issn.1672-3961.0.2017.273

Abstract ( 1515 )   PDF (1453KB) ( 583 )   Save

References | Related Articles | Metrics

In order to realize the rapid online failt detection of unmanned aerial vehicle(UAV)flight control system, a fault detection approach based on optimal estimation of the L2-norm was proposed to the fault detection(FD)of UAV nonlinear flight control system. The nonlinear fault model of UAV flight control system was established, and an optimal estimation of the L2-norm of the unknowninputs was found to be the evaluation function for FD. On the foundation of the approach for linear discrete time-varying systems, the projection in Krein space was applied to calculate the evaluation function recursively, and thus the heavy online computational burden could be solved. The FD for UAV elevator and rate gyros was taken as an example to demonstrate the effectiveness of the proposed method. The results showed that the faults of the UAV flight control system could be detected rapidly through the proposed approach, and the safety of UVA could be guaranteed reliably.

Fault detection for multi-agent systems based on intermediate observer

WU Yanming, WANG Ruiyun, WANG Zhanshan

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  96-102.  doi:10.6040/j.issn.1672-3961.0.2017.169

Abstract ( 1415 )   PDF (1518KB) ( 636 )   Save

References | Related Articles | Metrics

The actuator fault detection was studied with fault detection for a class of linear multi-agent systems under indirect communication network topology. A set of virtual systems were designed, and an actuator fault detection algorithm was proposed based on intermediate observer, which overcame the restriction of observer matching condition. The intermediate variable matrix was used appropriately to simultaneously estimate the states and faults. The states residual signal between adjacent agents was used to detect not only its own faults but also the faults of its nearest neighbors. Based on the Lyapunov stability theory, it was proved that the estimation errors were uniformly ultimately bounded. Simulation results showed the effectiveness of the designed method.

Fault isolability analysis based on improved distance similarity

SONG Yang, ZHONG Maiying

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  103-109.  doi:10.6040/j.issn.1672-3961.0.2017.269

Abstract ( 1264 )   PDF (3892KB) ( 557 )   Save

References | Related Articles | Metrics

A fault isolability analysis approach based on improved distance similarity was proposed to evaluate quantitatively the difficulty level of fault isolation. A parity space-based fault diagnosis residual generator was taken as an example, and the improved fault isolation condition was constructed based on the difference of the residuals in probability distribution. Then the distance similarity and direction similarity of residuals were adopted to evaluate the difficulty level of fault isolation, and the quanlitative evaluation index of fault isolability was put forward. A simulation was carried out to analyze the fault isolability of a fixed-wing unmanned aerial vehicle longitudinal flight control system. The results demonstrated that the method could decide accurately the fault isolation condition, and evaluate quantitatively the difficulty level of fault isolation. The improved fault isolation condition was more intuitional, and the evaluation index could evaluate comprehensively the difference of residuals in both distance and direction compared with existing approaches.

Chemical process monitoring based on two step subspace division

YANG Yawei, SONG Bing, SHI Hongbo

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  110-117.  doi:10.6040/j.issn.1672-3961.0.2017.176

Abstract ( 1285 )   PDF (3277KB) ( 581 )   Save

References | Related Articles | Metrics

In order to solve the problem of high dimension and complex distribution of data collected from modern chemical processes, a method for monitoring chemical process was presented based on two step subspace division(TSSD). In order to reduce the complexity of process analysis, variables with similar characteristic were divided into the same space. Considering the complex distribution of data, the subspace obtained from the first step was divided into Gaussian subspace and non-Gaussian subspace. Principal component analysis(PCA)and independent component analysis(ICA)were used to establish the detection models and construct the statistics. All statistics of subspaces were integrated and used to construct the final statistics based on local outlier factor(LOF). The process results showed that the optimal missed detection rates of TSSD can be obtained for 16 faults, especially 15.375% for fault 10 and 6.75% for fault 16. The superiority monitoring performance of the proposed two steps subspace division method was proved.

Fault diagnostic method for micro-grid based on wavelet singularity entropy and SOM neural network

QIU Lu, YE Yinzhong, JIANG Chundi

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  118-122.  doi:10.6040/j.issn.1672-3961.0.2017.183

Abstract ( 1383 )   PDF (1512KB) ( 571 )   Save

References | Related Articles | Metrics

According to the diversity of micro grids topology, through analyzing the theories of wavelet transform, singular value decomposition and extended shannon-entropy, the wavelet singular entropy could measure the fault signal. A fault diagnosis method for the micro grid system was proposed by integrating the wavelet singular entropy with the self organizing feature map(SOM)neural network. A micro grid fault simulation system was established by PSCAD4.2. The simulation results proved that the proposed diagnosis method was insensitive to the location and the time fault occurs, which had strong adaptability to the variation in structure topology.

A mode-correlation principal component analysis for the fault detection of marine current turbine

ZHANG Milu, WANG Tianzhen, TANG Tianhao, XIN Bin

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  123-129.  doi:10.6040/j.issn.1672-3961.0.2017.166

Abstract ( 1286 )   PDF (3124KB) ( 573 )   Save

References | Related Articles | Metrics

To solve the problem of multi-mode characteristic and frequent mode changes, a detection method for marine current turbine which called mode-correlation principal component analysis was proposed. The influence of modal change on the traditional principal component analysis(PCA)was analyzed in theory. The detection problem caused by multi-mode characteristic was described. A mode normalized algorithm was proposed in the proposed method to dynamic fitting the mode. The statistical difference value of different modes was removed due to relationships between modes. Compared with other methods, the experimental platform was built to verify the effectiveness of the proposed method. Theoretical analysis and experimental results showed that the proposed method could detect the fault quickly and accurately under the condition of variable speed and variable load.

A fault detection and estimation scheme for nonlinear stochastic systems based on SCKF

LI Hongyang, HE Xiao

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  130-135.  doi:10.6040/j.issn.1672-3961.0.2017.252

Abstract ( 1512 )   PDF (2618KB) ( 554 )   Save

References | Related Articles | Metrics

The fault detection and estimation problem was investigated for a class of nonlinear stochastic systems based on the square root cubature Kalman filter(SCKF). To estimate the states of complex nonlinear systems, SCKF has the outstanding characterizations of higher accuracy, better stability and lower computational burden. For nonlinear stochastic systems subject to actuator faults, the states were estimated based on the square root cubature Kalman filter. Moreover, according to the estimation results, a residual was designed by using the moving-horizon technique to detect the actuator fault. The faults were estimated based on a state augmentation method. A simulation experiment was given to verify the effectiveness of the proposed scheme.

Fault estimation for discrete-time systems with output dead-zone using two-stage Tobit kalman filter

HUANG Jie, HE Xiao

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  136-142.  doi:10.6040/j.issn.1672-3961.0.2017.212

Abstract ( 1434 )   PDF (2744KB) ( 767 )   Save

References | Related Articles | Metrics

The problem of estimating the fault for discrete-time systems with output dead-zone was addressed via two-stage Kalman filtering approach.Two Bernoulli random vectors were introduced to model the dead-zone effect. A two-stage Tobit Kalman filter(TSTKF)was derived to solve the filtering problem. The covariance matrices of the augmented state Tobit Kalman filter(ASTKF)was decoupled by using a two-stage U-V transformation technique to obtain the TSTKF. A numerical example was provided to illustrate the feasibility and accuracy of the proposed filter in the end which was compared with both standard Kalman filter and Kalman filter with intermittent observations.

Clustering of blast furnace historical data based on PCA similarity factor and spectral clustering

PANG Renming, WANG Bo, YE Hao, ZHANG Haifeng, LI Mingliang

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  143-149.  doi:10.6040/j.issn.1672-3961.0.2017.172

Abstract ( 1432 )   PDF (2051KB) ( 552 )   Save

References | Related Articles | Metrics

The principal component analysis(PCA)similarity factor and spectral clustering algorithms were combined and applied analyze the operational state change in a blast furnace by mining the historical data. The similarity between different data sets generated from moving windows by combining the PCA similarity factor and the distance similarity factor was measured, and the historical data were clustered by constructing the graph from the similarity between different data sets and using spectral clustering algorithm. The effect of operating point drift was reduced and the more accurate clustering result was effectively and steadily achieved by the proposed method. The off-line test proved that, compared with the existing methods which combined the PCA similarity factor and k-means clustering, the proposed method could more effectively recognize the operational state change in a blast furnace.

Parity space-based fault detection for unmanned aerial vehicle flight control systems

ZHAO Xuan, ZHONG Maiying, GUO Dingfei

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  150-156.  doi:10.6040/j.issn.1672-3961.0.2017.270

Abstract ( 1598 )   PDF (1126KB) ( 380 )   Save

References | Related Articles | Metrics

The fault detection(FD)for unmanned aerial vehicle(UAV)flight control system is of great significance to ensure the flight safety of UAV. The parity space approach has the advantage of the decoupling of residual and unknown initial state. However, the increasing of parity order will lead to heavy computational task. Aiming at these problems, a modified parity space approach was put forward for the FD of UAV nonlinear flight control systems. The nonlinear fault model of UAV flight control system was established. On the foundation of parity space approach for linear discrete time-varying systems, the projection in Krein space was applied to calculate the evaluation function recursively, and thus the heavy online computational burden could be solved. The FD for UAV pitot tube and elevator was taken as an example to demonstrate the effectiveness of the proposed method. The results showed that the faults of the UAV flight control system could be detected rapidly through the proposed approach.

Performance assessment of lithium-ion battery based on geometric features and manifold distance

BAO Tala, MA Jian, GAN Zuwang

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  157-165.  doi:10.6040/j.issn.1672-3961.0.2017.268

Abstract ( 1034 )   PDF (3970KB) ( 381 )   Save

References | Related Articles | Metrics

The estimations of state of charge and state of health evolved in li-ion battery health management systems can help managing the reliability and safety of the fielded battery. Considering that many data-driven state of health estimations habituated to model the battery monitoring information in Euclid space with the purpose of assessing battery health status, which often brings about a poor adaptability to operation conditions, manifold learning was used to mine the health information hidden in the battery monitoring data and manifold distance was utilized to measure the battery health condition. At last, a case analysis was conducted to validate the proposed state of health estimation method for the li-ion battery.

Fault diagnosis and fault-tolerant control methods of X-tail UAV

DENG Junwu, ZHANG Yumin, ZHANG Hongdi, DU Xiaokun

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  166-172.  doi:10.6040/j.issn.1672-3961.0.2017.195

Abstract ( 1426 )   PDF (4281KB) ( 1089 )   Save

References | Related Articles | Metrics

Actuators are the key agency of the UAV. For fault detection and fault diagnosis purpose of problems such as the dead, gain loss and deviation of the actuator, the fault detection filter and Kalman filter were presented in this contribution. The residual vector with actuator fault information was output by using the detection filter, then the threshold detection and residual direction characteristics were used to detect and isolate the fault. After the fault alarm, the Kalman filter was used to estimate the fault parameters, and the nature and extent of the fault were obtained. According to the different forms of fault, the method of control command compensation or reconstruction was finally used for fault-tolerant control purpose. Based on the turning rate model of the X-tails UAV, simulation test showed that the fault diagnosis method was effective and feasible, which could rapidly obtain the fault information, and the fault-tolerant strategy could well restore the system performance.

Multiblock local Fisher discriminant analysis for chemical process fault classification

WANG Lei, DENG Xiaogang, CAO Yuping, TIAN Xuemin

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  179-186.  doi:10.6040/j.issn.1672-3961.0.2017.181

Abstract ( 1244 )   PDF (2284KB) ( 301 )   Save

References | Related Articles | Metrics

Fisher discriminant analysis(FDA)was an effective chemical process fault classification method. However, the local data structure information was not investigated within traditional FDA method. To deal with this problem, a multiblock local Fisher discriminant analysis(MLFDA)method was proposed for more effective chemical process fault recognition. This method analyzed the local data structure characteristics from the variable-dimension and sample-dimension. To mine the local information in the variable-dimension, a variable block division method was designed based on the relevancy between the variables and the principal component subspace of the dataset, with which all the variables could be divided into several local variable blocks. Furthermore, considering the characteristics of local sample structure, the local FDA(LFDA)using local weighting factors was applied to construct classifier for each local variable block. An integrating strategy based on weighting classification performance weighting was presented to combine the results from different classifiers. Simulation results on Tennessee Eastman process showed that the proposed MLFDA method had a lower misclassification rate than traditional FDA and LFDA methods.

Fault diagnosis for industrial processes based on causal topological graph

WANG Mengyuan, ZHANG Xiong, MA Liang, PENG Kaixiang

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  187-194.  doi:10.6040/j.issn.1672-3961.0.2017.239

Abstract ( 1680 )   PDF (2687KB) ( 624 )   Save

References | Related Articles | Metrics

With the combination of the process knowledge and data driven methods, the fault diagnosis method based on causal topological graph could effectively deal with the fault location and fault propagation identification. A correlation index(CI)based on partial correlation coefficient was applied to the causal topological graph to analyze the correlation between variables quantitatively. The proposed CI was monitored via probabilistic principal component analysis method(PPCA)for fault detection. The concept of mean weighted value and causal topological graph were introduced in order to identify the optimal fault propagation path based on reconstruction-based contribution(RBC)after detecting a fault. The effectiveness of the method was verified by the application of hot strip mill process(HSMP). The results showed that the proposed method could effectively identify the fault roots and propagation paths.

Weighted hyper-ellipsoidal support vector data description with negative samples for outlier detection

YAO Yu, FENG Jian, ZHANG Huaguang, HAN Kezhen

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  195-202.  doi:10.6040/j.issn.1672-3961.0.2017.180

Abstract ( 1237 )   PDF (2353KB) ( 335 )   Save

References | Related Articles | Metrics

To solve the influence of the imbalance between positive and negative samples in training sample set, a method named weighted hyper-ellipsoidal support vector data description with negative samples(WNESVDD)was proposed. Mahalanobis distance was introduced such that the information of sample distribution was completely considered. Both normal and negative samples were utilized to modeling. Cost-sensitive learning was introduced to set different weights for different classes. The results showed that the empty areas that decision boundary enclosed were reduced effectively and the decision boundary was refined in the proposed method. The data utilization rate was obviously improved. Several experiments on University of California at Irvine(UCI)data sets and the data set from the semi-conductor manufacturing process were conducted. The experiments results showed that the proposed method had strong ability of anomaly detection, and compared with the similar method, false positives and false negatives were dramatically reduced.

Fault tolerant cooperative control for a class of complex networks

HUANG Chengkai, YANG Hao, JIANG Bin, CHENG Shuyao

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  203-209.  doi:10.6040/j.issn.1672-3961.0.2017.194

Abstract ( 1234 )   PDF (1235KB) ( 377 )   Save

References | Related Articles | Metrics

A fault tolerant cooperative control strategy for a class of complex network systems with the time varying interconnection topology was proposed. The general neighboring rule-based linear cooperative protocol was developed and a sufficient synchronization condition was derived. It was showed that the target point was still reached with the proposed fault tolerant cooperative control method when fault occurred. An example of a complex consisting of four one-link manipulators was presented to demonstrate the effectiveness of the proposed method.

Fault diagnosis and fault tolerant control based on sliding mode observer for speed sensor in asynchronous motor

XIE Xiaolong, JIANG Bin, LIU Jianwei, JIANG Yinhang

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  210-214.  doi:10.6040/j.issn.1672-3961.0.2017.175

Abstract ( 1363 )   PDF (2112KB) ( 435 )   Save

References | Related Articles | Metrics

Aiming at the fault appearing in speed sensor of asynchronous motor in vector control system, a sliding-mode observer based fault diagnosis and fault tolerant control method was proposed. The state space model of asynchronous motor was established by coordinate transformation method, the sliding mode observer was then constructed the stability conditions and speed adaptive law were derived by Lyapunov stability theory. The switching control strategy was designed to guarantee the safety of the induction motor, which changed system working mode into speed sensorless vector control when the residuals were larger than the given thresholds. The developed method could detect faults within 0.2 s and recover the performance of the system, whose rapidity and effectiveness was validated via a simulation experiment.

Application of dynamic kernel principal component analysis in unmanned aerial vehicle fault diagnosis

LI Minghu, LI Gang, ZHONG Maiying

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  215-222.  doi:10.6040/j.issn.1672-3961.0.2017.274

Abstract ( 1301 )   PDF (2390KB) ( 417 )   Save

References | Related Articles | Metrics

The flight control system(FCS)was the core subsystem of unmanned aerial vehicle(UAV), performing FD for it could greatly improve the safety and reliability of UAV. When the mathematical model of UAV was unknown or uncertain, data-driven methods were more suitable than model-based FD methods. Considering that FCS of UAV was a typical nonlinear dynamic system, a nonlinear principal component analysis(PCA)method was used instead. A dynamic kernel principal component model under normal state was established for UAV, then fault detection was performed by T² and SPE statistics; When a fault was detected, a method called reconstruction-based contribution was used for fault isolation. The simulation results showed that the proposed method could achieve better fault diagnosis effect for typical faults of actuators and sensors than linear DPCA model. Besides, DKPCA could achieve high sensitivity for small faults of UAV.

Fault diagnosis of asynchronous motor based on wavelet packet entropy and wavelet neural network

WU Jianping, JIANG Bin, LIU Jianwei

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  223-228.  doi:10.6040/j.issn.1672-3961.0.2017.179

Abstract ( 1353 )   PDF (1904KB) ( 418 )   Save

References | Related Articles | Metrics

A method based on wavelet packet entropy and wavelet neural network was presented for asynchronous motor to realize fault diagnosis. The signal with faulty information was pretreated by wavelet packet, the wavelet packet energy spectrum entropy and coefficient entropy was extracted. The feature vector of information entropy was constructed. When the feature vector was input into the wavelet neural network, we trained it to detect and output the fault mode, so as to realize the fault diagnosis. This method had good adaptive resolution and fault tolerance, and it could avoid local minima and slow convergence effectively. The experiment results showed that this method could be used for fault diagnosis of induction motors, which was better than BP neural network model with the same parameters.

A fault diagnosis and fault-tolerant control strategy for multilevel inverter

LIU Zhuo, WANG Tianzhen, TANG Tianhao, FENG Yefan, YAO Junqi, GAO Diju

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  229-237.  doi:10.6040/j.issn.1672-3961.0.2017.170

Abstract ( 1500 )   PDF (3383KB) ( 432 )   Save

References | Related Articles | Metrics

To solve the open circuit faults of the switching device in multilevel inverters, a fault diagnosis and fault-tolerant control strategy was proposed for a cascaded H-bridge multilevel inverter. The fast fourier transform(FFT)and the principal component analysis(PCA)were utilized to reprocess and feature extraction respectively without changing the circuit topology of the cascaded H-bridge multilevel inverter. The faults of cascaded H-bridge multilevel inverters could be fast diagnosed with constructing Bayesian network(BN). According to the fault diagnosis results, output voltages of the multilevel inverter could still meet the demands to keep the system operating smoothly. The way was based on a voltage reconstruction fault tolerant control method by changing the modulation waves and carrier waves, and the performance of the strategy was verified by simulation and experiment.

Fault diagnosis and fault tolerant control for non-Gaussian uncertain singular stochastic distribution control systems

SUN Yuancheng, YAO Lina

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  238-245.  doi:10.6040/j.issn.1672-3961.0.2017.184

Abstract ( 1361 )   PDF (2144KB) ( 314 )   Save

References | Related Articles | Metrics

A tracking controller was constructed to make the output PDF track the desired one before fault occurring. A robust fault diagnosis and fault tolerant tracking control algorithm of non-Gaussian uncertain singular stochastic distribution control systems with probability density function approximation error was proposed. A fault diagnosis observer was designed to estimate the fault information with the disturbance of uncertainty and probability density function approximation error. When the fault occurred, the tracking controller was reconfigured to ensure that the post-fault output probability density function still tracked the desired distribution. The Lyapunov stability theory was used to analyze the stability of the observation error dynamic system, the closed-loop system and the tracking error dynamic system, the gain matrices were obtained by solving the corresponding linear matrix inequalities. An illustrated example was given to demonstrate the effectiveness of the proposed algorithm for time-varying fault.

Quality prediction method based on hybrid MPLS for multiphases process

YE Xiaofeng, WANG Peiliang, YANG Zeyu

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  246-253.  doi:10.6040/j.issn.1672-3961.0.2017.178

Abstract ( 1161 )   PDF (1471KB) ( 281 )   Save

References | Related Articles | Metrics

When the traditional multi-way partial least squares(MPLS)method was carried out for quality prediction, it performed the problems of low prediction accuracy and lack of local capacity. Considering to multiphase characteristics in industrial batch process, a mixture MPLS model method was proposed for the multiphases quality prediction. The Gauss mixture model(GMM)was employed to model the high dimensional spatial distribution of the measurement and the quality variables,which was used to identity sub-phase data blocks in each batch. Due to the problem of inequality length in identical sub-phase, the dynamic time warping(DTW)algorithm corresponding to the maximum path length of time was carried out to synchronize each data block in same sub-phase, whats more, the sub-phase MPLS model was set up according to the variable expansion method. The Fisher discriminate analysis(FDA)was introduced to minimize the relationship among sub-phase blocks, and then the kernel density method was used to monitor phases switch online by estimating the probability density distribution of less relationship sub-phase blocks. The multiphases quality prediction was established by mixing several sub-phase MPLS models according to Bays principle. Furthermore, the result of quality prediction for penicillin fermentation process showed the effectiveness of proposed method.

On least squares fault estimation with incorrect measurement noise statistics

ZHAO Ye, HE Xiao, ZHOU Donghua

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  254-262.  doi:10.6040/j.issn.1672-3961.0.2017.253

Abstract ( 934 )   PDF (1544KB) ( 290 )   Save

Related Articles | Metrics

The least squares fault estimation and performance analysis problem was investigated for a class of systems with event-triggered measurement transmission and incorrect measurement noise statistics. Usually, measurement noise statistics was not completely known. In this case, an event-triggered scheme was proposed to transmit the measurement output to remote estimator when a pre-set condition was violated and an event was triggered. A filter was designed in order to minimize an upper bound of filtering error covariance with event-triggered measurement transmissions and incorrect measurement noise statistics. The desired filter parameters were calculated recursively for online computation in the framework using the least squares method. Performance of the proposed filter with incorrect measurement noise statistics was further studied and analyzed. A numerical simulation with actual background was exploited to illustrate the effectiveness of the proposed algorithm and demonstrate the performance difference.

Fault estimation design for linear multi-agent systems with switching topologies

CUI Yang, ZHANG Ke, JIANG Bin

JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE). 2017, 47(5):  263-270.  doi:10.6040/j.issn.1672-3961.0.2017.177

Abstract ( 1334 )   PDF (1034KB) ( 367 )   Save

References | Related Articles | Metrics

This paper concerns with the actuator fault estimation of linear multi-agent systems with switching topologies by means of mode-dependent average dwell time. Using relative output estimation error, the observer and the global estimation error dynamics were constructed. Switching topologies and the existence of disturbance were taken into consideration. Based on the stability theory of the switched systems and fault diagnosis theory, the existence conditions of the observer and the design criterion of the parameters were given via the mode-dependent average dwell time method. Using the appropriate fault estimation algorithm the sufficient condition for the global fault estimation error system to achieve exponential stability and H_∞ performance was given by linear matrix inequalities. An example was given to verify the effectiveness of the proposed method.