Constrained Active Fault Tolerant Control Based on Active Fault Diagnosis and Interpolation Optimization

A new active fault tolerant control scheme based on active fault diagnosis is proposed to address the component/actuator faults for systems with state and input constraints. Firstly, the active fault diagnosis is composed of diagnostic observers, constant auxiliary signals, and separation hyperplanes, all of which are designed offline. In online applications, only a single diagnostic observer is activated to achieve fault detection and isolation. Compared with the traditional multi-observer parallel diagnosis methods, such a design is beneficial to improve the diagnostic efficiency. Secondly, the active fault tolerant control is composed of outer fault tolerant control, inner fault tolerant control and a linear-programming-based interpolation control algorithm. The inner fault tolerant control is determined offline and satisfies the prescribed optimal control performance requirement. The outer fault tolerant control is used to enlarge the feasible region, and it needs to be determined online together with the interpolation optimization. In online applications, the updated state estimates trigger the adjustment of the interpolation algorithm, which in turn enables control reconfiguration by implicitly optimizing the dynamic convex combination of outer fault tolerant control and inner fault tolerant control. This control scheme contributes to further reducing the computational effort of traditional constrained predictive fault tolerant control methods. In addition, each pair of inner fault tolerant control and diagnostic observer is designed integratedly to suppress the robust interaction influences between estimation error and control error. The soft constraint method is further integrated to handle some cases that lead to constraint violations. The effectiveness of these designs is finally validated by a case study of a wastewater treatment plant model.     

Image processing-based detection method of learning behavior status of online calssroom students

Due to the outbreak of the new crown epidemic, online teaching is booming, but compared with traditional offline teaching, there are many problems, such as the difficulty of detecting the voice status of students. Therefore, the research on students’ online status detection system is of great significance. In this paper, based on image processing, the detection method of online classroom students’ learning behavior status is studied, and the learning status of students is detected from the perspective of face detection and face recognition fatigue detection. In this study, the students’ learning status is detected by the facial expressions in the video during the students’ learning process. When the students have negative emotions and become tired, the system can detect and record them in time and issue a warning. Therefore, this research can well solve the problems existing in online teaching, and to a certain extent, the teaching quality has been greatly improved.     

Deep learning-based energy-efficient relay precoder design in MIMO-CRNs

This paper investigates the problem of energy efficient relay precoder design in multiple-input multiple-output cognitive relay networks (MIMO-CRNs). This is a non-convex fractional programming problem, which is traditionally solved using computationally expensive optimization methods. In this paper, we propose a deep learning (DL) based approach to compute an approximate solution. Specifically, a deep neural network (DNN) is employed and trained using offline computed optimal solution. The proposed scheme consists of an offline data generation phase, an offline training phase, and an online deployment phase. The numerical results show that the proposed DNN provides comparable performance at significantly lower computational complexity as compared to the conventional optimization-based algorithm that makes the proposed approach suitable for real-time implementation.     

Personalized Image Classification by Semantic Embedding and Active Learning

Currently, deep learning has shown state-of-the-art performance in image classification with pre-defined taxonomy. However, in a more real-world scenario, different users usually have different classification intents given an image collection. To satisfactorily personalize the requirement, we propose an interactive image classification system with an offline representation learning stage and an online classification stage. During the offline stage, we learn a deep model to extract the feature with higher flexibility and scalability for different users’ preferences. Instead of training the model only with the inter-class discrimination, we also encode the similarity between the semantic-embedding vectors of the category labels into the model. This makes the extracted feature adapt to multiple taxonomies with different granularities. During the online session, an annotation task iteratively alternates with a high-throughput verification task. When performing the verification task, the users are only required to indicate the incorrect prediction without giving the exact category label. For each iteration, our system chooses the images to be annotated or verified based on interactive efficiency optimization. To provide a high interactive rate, a unified active learning algorithm is used to search the optimal annotation and verification set by minimizing the expected time cost. After interactive annotation and verification, the new classified images are used to train a customized classifier online, which reflects the user-adaptive intent of categorization. The learned classifier is then used for subsequent annotation and verification tasks. Experimental results under several public image datasets show that our method outperforms existing methods.     

Deep-BiGRU based channel estimation scheme for MIMO–FBMC systems

Deep Learning (DL)–based wireless communication systems have the potential to improve the conventional functions and current architecture of communication systems. In this paper, we propose a novel DL-based channel estimation scheme for multiple-input multiple-output filter bank multicarrier with offset quadrature amplitude modulation (MIMO-FBMC/OQAM) systems called deep bidirectional gated-recurrent unit (BiGRU) scheme. This scheme can easily be applied to a single-input single-output (SISO) system. The proposed scheme is divided into two stages: offline and online. The network is first trained in the offline stage. The prediction of channel information and estimation of the channel matrix using the trained network is then performed in the online stage. The simulation results in terms of the normalized mean square error (NMSE) and bit error rate (BER) demonstrate that, under different time-varying channel models, the proposed DL scheme significantly improves the channel estimation performance of FBMC for single and multiple antennas compared to conventional interference approximation method (IAM) channel estimation methods.     

Porcelain Insulator Crack Location and Surface States Pattern Recognition Based on Hyperspectral Technology

A porcelain insulator is an important part to ensure that the insulation requirements of power equipment can be met. Under the influence of their structure, porcelain insulators are prone to mechanical damage and cracks, which will reduce their insulation performance. After a long-term operation, crack expansion will eventually lead to breakdown and safety hazards. Therefore, it is of great significance to detect insulator cracks to ensure the safe and reliable operation of a power grid. However, most traditional methods of insulator crack detection involve offline detection or contact measurement, which is not conducive to the online monitoring of equipment. Hyperspectral imaging technology is a noncontact detection technology containing three-dimensional (3D) spatial spectral information, whereby the data provide more information and the measuring method has a higher safety than electric detection methods. Therefore, a model of positioning and state classification of porcelain insulators based on hyperspectral technology is proposed. In this model, image data were used to extract edges to locate cracks, and spectral information was used to classify the surface states of porcelain insulators with EfficientNet. Lastly, crack extraction was realized, and the recognition accuracy of cracks and normal states was 96.9%. Through an analysis of the results, it is proven that the crack detection method of a porcelain insulator based on hyperspectral technology is an effective non-contact online monitoring approach, which has broad application prospects in the era of the Internet of Things with the rapid development of electric power.     

The characteristic analysis of stochastic resonance and bearing fault diagnosis based on NWSG model driven by trichotomous noise

In this paper, a new system whose potential function (NWSG) based on the joint of New Woods-Saxon potential function(NWSP) and Gaussian potential function(GP), driven by trichonomous is proposed to optimize the perform in bearing fault diagnosis. Firstly, exploring the influence of various system parameters on the shape of NWSG, besides, presenting a method of numerical simulation for trichotomous noise. The results show that the potential function can convert between three state, which is monostable, bistable, and tristable respectively, under different system parameter values. In addition, the mean of signal-noise ratio increase(MSNRI) is served as the measurement index of stochastic resonance (SR) for periodic signal detection, while traditional SR under optimal parameters. Finally, bearing fault diagnosis is carried out. It is found that the performance of the proposed system is better than traditional system which also verified in the bearing fault diagnosis.     

An interpretable intrusion detection method based on few-shot learning in cloud-ground interconnection

An enterprise’s private cloud may be attacked by attackers when communicating with the public cloud. Although traffic detection methods based on deep learning have been widely used, these methods rely on a large amount of sample data and cannot quickly detect new attacks such as Zero-day Attacks. Moreover, deep learning has a black-box nature and cannot interpret the detection results, which has certain security risks. This paper proposes an interpretable abnormal traffic detection method, which can complete the detection task with only a few malicious traffic samples. Specifically, it uses the covariance matrix to characterize each traffic category and then calculates the similarity between the query traffic and each category according to the covariance metric function to realize the traffic detection based on few-shot learning. After that, the traffic images processed by the random masks are input into the model to obtain the predicted probability of the corresponding traffic category. Finally, the predicted probability is linearly summed with each mask to generate the final saliency map to interpret and analyze the model decision. In this paper, experiments are carried out by simulating only 15 and 25 malicious traffic samples. The results show that the proposed method can obtain good accuracy and recall, and the interpretation analysis shows that the model is reliable and interpretable.     

Rolling bearing diagnosing method based on Empirical Mode Decomposition of machine vibration signal

Rolling bearing faults are one of the major reasons for breakdown of industrial machinery and bearing diagnosing is one of the most important topics in machine condition monitoring.The main problem in industrial application of bearing vibration diagnostics is the masking of informative bearing signal by machine noise. The vibration signal of the rolling bearing is often covered or concealed by other structural vibrations sources, such as gears. Although a number of vibration diagnostic techniques have been developed over the last several years, in many cases these methods are quite complicated in use or only effective at later stages of damage development. This paper presents an EMD-based rolling bearing diagnosing method that shows potential for bearing damage detection at a much earlier stage of damage development.By using EMD a raw vibration signal is decomposed into a number of Intrinsic Mode Functions (IMFs). Then, a new method of IMFs aggregation into three Combined Mode Functions (CMFs) is applied and finally the vibration signal is divided into three parts of signal: noise-only part, signal-only part and trend-only part. To further bearing fault-related feature extraction from resultant signals, the spectral analysis of the empirically determined local amplitude is used. To validate the proposed method, raw vibration signals generated by complex mechanical systems employed in the industry (driving units of belt conveyors), including normal and fault bearing vibration data, are used in two case studies. The results show that the proposed rolling bearing diagnosing method can identify bearing faults at early stages of their development.     

Training Process of Unsupervised Learning Architecture for Gravity Spy Dataset

Transient noise appearing in the data from gravitational-wave detectors frequently causes problems, such as instability of the detectors and overlapping or mimicking gravitational-wave signals. Because transient noise is considered to be associated with the environment and instrument, its classification would help to understand its origin and improve the detector's performance. In a previous study, an architecture for classifying transient noise using a time–frequency 2D image (spectrogram) is proposed, which uses unsupervised deep learning combined with variational autoencoder and invariant information clustering. The proposed unsupervised-learning architecture is applied to the Gravity Spy dataset, which consists of Advanced Laser Interferometer Gravitational-Wave Observatory (Advanced LIGO) transient noises with their associated metadata to discuss the potential for online or offline data analysis. In this study, focused on the Gravity Spy dataset, the training process of unsupervised-learning architecture of the previous study is examined and reported.     

基于粒子滤波状态估计的滚动轴承故障识别方法

提出了一种基于粒子滤波状态估计的滚动轴承故障识别方法,该方法主要包括故障模型建立和故障识别两个步骤。在故障模型建立部分,首先依据滚动轴承不同故障状态下的振动信号,建立对应的自回归模型,作为故障模型;在故障识别部分,将正常状态下对应的模型,转化为状态空间模型,设计粒子滤波器,然后对不同的故障状态进行估计,提取其残差的相关特征,并结合模型参数特征应用BP神经网络识别算法进行故障识别。最后以美国凯斯西储大学的滚动轴承振动数据为例,验证了该方法的有效性。     

基于LGSPP-Bayes的故障检测与辨识方法

针对主元分析(Principal component analysis, PCA)和局部保持投影(Locality preserving projections, LPP)方法在降维过程中分别只能保留数据集的整体信息和局部信息,提出一种基于局部整体结构保持投影的贝叶斯故障检测与辨识方法(Local and global structure preserving projections and bayes, LGSPP-Bayes)。首先,将正常工况操作下的原始数据通过局部整体结构保持投影方法投影到低维特征空间,得到高维到低维的数据转换矩阵;然后通过设计贝叶斯分类器来进行故障检测;最后当检测到故障后通过计算贝叶斯分类函数的大小来识别故障种类。将LGSPP-Bayes方法应用于TE过程,仿真结果表明对故障的检测优于其他方法,并且可以很好地将故障种类识别出来。     

过程控制系统中多条件约束的多传感器故障检测与诊断

针对工业过程控制系统中的故障具有类型多样、时空独立和非线性等特点,使得检测与诊断效率降低,系统性能下降等问题,提出了一种基于自定义多条件约束的多传感器故障检测与诊断机制。该机制,首先考虑了系统的稳态和时空特征建立了非线性过程控制系统多故障模型,并给出了满足条件判定法则;然后对于系统中的单故障,并发故障和通信故障等类型给出了多条件约束法则及独立特性判断;最后提出了通过自定义多条件约束的多传感器故障检测与诊断机制。实验结果表明,在平均检测概率、稳态特征保持能力和系统功耗等方面明显优于无条件约束的机制,可以显著改善过程控制系统性能。     

Identification of multiple transient faults based on the adaptive spectral kurtosis method

This paper proposes a multi-fault detection method based on the adaptive spectral kurtosis (ASK) analysis of the vibration signal from single sensor. A theoretical model of multiple bearing faults is established in this paper. Compared with the kurtogram and protrugram techniques, the proposed method can more effectively extract signatures of multiple bearing faults even in the presence of strong background noise. The performance of the proposed method in fault detection of the rolling element bearings is validated using simulation data and experimental signals from a bearing with multiple faults and two faulty bearings.     

基于灰色预测和KID离群点检测的AUV故障诊断

提高故障诊断能力对于确保水下机器人AUV系统的稳定运行具有重要意义。针对水下机器人推进器系统,提出一种基于离群点检测的AUV故障检测方法。首先,将传感器采集的数据进行灰色预测处理;然后,提出了一种结合K-mean和DBSCAN的改进迭代聚类(Iterative K-mean DBSCAN,IKD)算法进行离群点检测;最后,与K-mean和DBSCAN算法相比,仿真实验结果表明基于灰色预测和KID离群点检测算法的故障检测准确率高,能够有效地实现水下机器人AUV的无监督故障诊断。     

基于机器视觉的高精密轴承在线检测

高精密轴承是一种圆柱型零件,针对圆柱型零件高曲率表面缺陷及外形尺寸不能同时进行在线检测的问题,设计并实现了基于机器视觉的在线检测系统。检测时,为了解决金属件表面反光的问题,设计了专用的光源系统和照明方式,通过光学系统和机械旋转平台的配合,圆柱型零件在旋转的过程中被光学系统成像,从而可以采集到完整的圆柱面图像;经过快速的图像处理技术,可以检测到微米级的轴承表面缺陷及外形尺寸。检测结果表明系统具有高效率、精度高、易于使用等特点,可有效解决高精密轴承表面缺陷及尺寸在线检测的问题。     

基于改进的Mask R-CNN的乳腺肿瘤目标检测研究

乳腺癌是全球女性死亡率最高的恶性肿瘤之一,早期发现有助于提升患者的存活率。本文利用深度学习中的目标检测网络对乳腺X线图像中的肿瘤病变区域进行定位和分类;然后选取Mask R-CNN网络作为目标检测模型,对Mask R-CNN的基准网络D-ShuffleNet进行改进,提出了一种新的网络——Mask R-CNN-II网络,并在Mask R-CNN-II网络中应用迁移学习算法。通过实验验证了Mask R-CNN-II网络比Mask R-CNN网络的检测精度更高,而且验证了所提基准网络、所使用的融合图像的思想以及迁移学习算法是有效的。Mask R-CNN-II有利于提高乳腺肿瘤的定位与分类,可为放射科医生提供辅助诊断意见,具有一定的临床应用价值。     

基于ODDD水下机器人故障诊断方法

近年来数据挖掘技术的快速发展使得利用水下机器人作业过程中积累的大量数据进行故障诊断成为可能。基于数据挖掘的故障诊断技术能够从数据中获取潜在的诊断知识。针对水下机器人推进器系统数据特征,提出一种基于聚类和距离的离群点检测方法(Outlier Detection based on DBSCAN and Distance,ODDD)。首先,对数据进行粗聚类,然后采用剪枝规则进行离群点检测,来实现故障诊断。仿真实验结果表明算法能够实现水下机器人快速有效的故障检测。     

基于互信息熵-近红外光谱的过程模式故障检测

近红外光谱分析在工业过程故障检测方面具有独特的优势,是一种准确且高效的方法。结合互信息熵和传统的主成分分析,对近红外光谱特征信息进行提取,通过构建过程的模式来刻画工业过程的运行状态。利用近红外光谱数据,从有机分子含氢基团振动信息中获取工业系统的过程模式,从微观分子层面探索提高工业过程故障检测准确率的有效方法,结合贝叶斯统计学习技术,提出了基于近红外光谱数据的工业过程故障检测技术。针对近红外光谱信息量丰富,谱带较宽,特征性不强的特点,首先对工业过程不同运行状态下的近红外光谱吸光度数据进行一阶导数预处理,采用主成分分析法（principal component analysis,PCA）压缩光谱数据量,扩大不同运行状态下光谱特征信息的差异性,提取光谱的内部特征信息。然后采用互信息熵（mutual information entropy,MIE）作为光谱特征信息相关性度量函数,基于最小冗余最大相关算法进一步减少光谱特征信息间的冗余并最大化光谱特征信息与类别的相关性,弥补了PCA无监督特征波长选择的不足,提出一种基于PCA-MIE的过程模式构建方法,获得的过程模式子集更紧凑更具类别表现力。再利用贝叶斯统计学习算法,根据后验概率对构建的模式子集进行决策,判别生产过程的正常状态和故障状态。由于过程模式子集结合了PCA浓聚方差的优势和互信息熵相关性测度的特征信息选择方法,蕴含了更多的近红外光谱的本质信息与内在规律,从而更能刻画工业过程的运行状态。接着,设置测试准确率TA作为评估标准,用以评价故障检测方法的性能效果。最后利用某化工厂提供的原油脱盐脱水过程近红外光谱数据对所提方法进行验证,并与传统近红外光谱特征信息提取方法PCA和MIE方法性能进行对比分析,结果表明基于PCA-MIE的过程模式故障检测方法几乎在所有维数子集上性能都优于其他两种方法,在特征维数为18维时获得最高的准确率94. 6%,证明了方法的优越性。