A challenge to many real-world data streams is imbalance with concept drift, which is one of the most critical tasks in anomaly detection. Learning nonstationary data streams for anomaly detection ...has been well studied in recent years. However, most of the researches assume that the class of data streams is relatively balanced. Only a few approaches tackle the joint issue of imbalance and concept drift. To overcome this joint issue, we propose an ensemble learning method with generative adversarial network-based sampling and consistency check (EGSCC) in this paper. First, we design a comprehensive anomaly detection framework that includes an oversampling module by generative adversarial network, an ensemble classifier, and a consistency check module. Next, we introduce double encoders into GAN to better capture the distribution characteristics of imbalanced data for oversampling. Then, we apply the stacking ensemble learning to deal with concept drift. Four base classifiers of SVM, KNN, DT and RF are used in the first layer, and LR is used as meta classifier in second layer. Last but not least, we take consistency check of the incremental instance and check set to determine whether it is anormal by statistical learning, instead of threshold-based method. And the validation set is dynamic updated according to the consistency check result. Finally, three artificial data sets obtained from Massive Online Analysis platform and two real data sets are used to verify the performance of the proposed method from four aspects: detection performance, parameter sensitivity, algorithm cost and anti-noise ability. Experimental results show that the proposed method has significant advantages in anomaly detection of imbalanced data streams with concept drift.
Chest computed tomography (CT) becomes an effective tool to assist the diagnosis of coronavirus disease-19 (COVID-19). Due to the outbreak of COVID-19 worldwide, using the computed-aided diagnosis ...technique for COVID-19 classification based on CT images could largely alleviate the burden of clinicians. In this paper, we propose an Adaptive Feature Selection guided Deep Forest (AFS-DF) for COVID-19 classification based on chest CT images. Specifically, we first extract location-specific features from CT images. Then, in order to capture the high-level representation of these features with the relatively small-scale data, we leverage a deep forest model to learn high-level representation of the features. Moreover, we propose a feature selection method based on the trained deep forest model to reduce the redundancy of features, where the feature selection could be adaptively incorporated with the COVID-19 classification model. We evaluated our proposed AFS-DF on COVID-19 dataset with 1495 patients of COVID-19 and 1027 patients of community acquired pneumonia (CAP). The accuracy (ACC), sensitivity (SEN), specificity (SPE), AUC, precision and F1-score achieved by our method are 91.79%, 93.05%, 89.95%, 96.35%, 93.10% and 93.07%, respectively. Experimental results on the COVID-19 dataset suggest that the proposed AFS-DF achieves superior performance in COVID-19 vs. CAP classification, compared with 4 widely used machine learning methods.
High-speed ultrasonic vibration cutting (HUVC) applies an ultrasonic vibration to a cutting tool transverse to the cutting direction. The advantage of this method is that the high-speed cutting ...exceeds the critical cutting speed restrictions of ultrasonic vibration cutting. However, irregular dimple arrays are machined on the work-piece surface, increasing the surface roughness. In this paper, a phase shift control method is proposed, which aims to overlap dimples in a regular pattern to decrease the surface roughness for HUVC. First, this paper investigates the theory of phase shift control to guarantee the surface quality of HUVC. Then, a closed-loop direct digital synthesis (DDS) system is designed to keep the phase shift of the machined dimples stable and controllable. A contrast experiment is conducted to compare the performance of the closed-loop DDS and the open-loop DDS methods in HUVC process. The results show that the closed-loop DDS method can render the dimple phase stable and controllable during the HUVC process. The phase shift value influences the surface roughness. Setting the phase shift to 0.94π with an appropriate feed-rate, the surface roughness is reduced to 0.303, compared to a surface roughness of 0.785 when using the open-loop DDS method.
Deep-holes are typical parts of aircraft structures, which is difficult to be machined. Boring assisted with ultrasonic vibration-assisted cutting has been proved to greatly enhance machining ...performance, especially for Ti6Al4V aviation alloy. This paper focuses on the machining extra-large aspect ratio (exceeding 20) of Ti6Al4V aviation deep-hole with the axial ultrasonic vibration-assisted boring (AUVB) method. First, the kinetics of the AUVB process is analyzed and a retrospective of its separation cutting feature is provided. Subsequently, a multi-stepped cantilever beam model of boring bar is established to analyze its static rigidity and dynamic stability. The aperture error is deduced, and then size coefficient is put forward to represent the static rigidity of the boring bar, which is inversely proportional to the diameter. In addition, two different vibration cases, namely modal-coupling vibration and regenerative vibration are considered for dynamic stability analysis. Next, the morphology of bored surface is analyzed, and the geometric height of peaks formed by AUVB and CB are calculated. Phase shift φ= π is suggested for obtaining a better surface in AUVB. Finally, the feasibility of AUVB on the machining of extra-large aspect ratio Ti6Al4V titanium alloy aviation deep-hole is verified through systematic experiments. Results demonstrate that AUVB has obvious advantages in reducing boring force, improving boring accuracy, suppressing vibration and promoting surface quality. Furthermore, the aperture error decreases to 50% and vibration amplitudes decrease to only 20–25%. The overall surface roughness of the deep-hole part stays below Ra=0.8μm with rotational speeds of 60r/min and 80r/min, and the surface residual stress state is transferred from the tensile state to a compressive one. As a result, not only AUVB can provide better boring accuracy and surface finish, but it also can enhance the surface fatigue properties.
The issue of a low minority class identification rate caused by data imbalance in anomaly detection tasks is addressed by the proposal of a GAN-SR-based intrusion detection model for industrial ...control systems. First, to correct the imbalance of minority classes in the dataset, a generative adversarial network (GAN) processes the dataset to reconstruct new minority class training samples accordingly. Second, high-dimensional feature extraction is completed using stacked asymmetric depth self-encoder to address the issues of low reconstruction error and lengthy training times. After that, a random forest (RF) decision tree is built, and intrusion detection is carried out using the features that SNDAE retrieved. According to experimental validation on the UNSW-NB15, SWaT and Gas Pipeline datasets, the GAN-SR model outperforms SNDAE-SVM and SNDAE-KNN in terms of detection performance and stability.
Due to the tool-workpiece separation effect, high-speed ultrasonic vibration cutting has the proven ability to reduce tool flank wear and prolong tool life by up to approximately three times when ...turning Ti-6Al-4V alloys in the recent research. Temperature reduction caused by the cooling medium penetrating into the gaps of the tool-workpiece contact zone is considered to be the primary reason for this remarkable performance. In this paper, a tool-workpiece thermocouple system was developed for the temperature measurement of high-speed ultrasonic vibration cutting Ti-6Al-4V alloys. First, the calibration of the tool-workpiece thermocouple was conducted to identify the piecewise linear relation between thermoelectric power and cutting temperature. Next, the insulation of the thermoelectric circuit from both the vibration-stimulated circuit and the experiment platform was realized. Finally, comparative cutting experiments on high-speed ultrasonic vibration cutting and conventional cutting with and without cutting fluid were carried out. The experimental results demonstrate that in the dry cutting experiments, no obvious temperature difference between these two methods occurred. However, when cutting fluid was applied, compared to conventional cutting, a maximum 30% temperature reduction could be realized by high-speed ultrasonic vibration cutting with a duty cycle value of 0.55 at cutting speeds of 250 to 300 m/min. Therefore, cutting fluid is thought to be an essential part of effective cooling in the process of high-speed ultrasonic vibration cutting and its cooling behavior during such a short separation interval needs further researches.
Axial ultrasonic vibration-assisted cutting (AUVC) has been proved to have better machining performance compared with conventional cutting methods; however, the effect of numerous and complex ...influencing factors on machining performance has not been clearly revealed, and a recommended combination of cutting conditions has not been proposed yet, especially for difficult-to-machine material such as Ti6Al4V alloy. This paper focuses on the experimental and theoretical investigation into Ti6Al4V machining performance with AUVC method. First, a retrospective of the separation characteristics of AUVC is provided, and the variable parameter cutting characteristics are demonstrated. The influencing factors on machining performance are classified into four categories: machining parameters, vibration parameters, tool choice, and cooling conditions. The relationship between these factors in terms of their effect on machining performance is established theoretically. Then, it describes experiments to determine the influence of these factors on cutting force, tool life, and surface roughness. For absolute influence, the orders for cutting force, tool life, and surface roughness are respectively cutting depth > amplitude > feed rate > rotation speed, rotation speed > feed rate > amplitude > cutting depth, and feed rate > amplitude > cutting depth > rotation speed. However, for relative influence, the order is unified as amplitude > feed rate > rotation speed > cutting depth. Finally, it suggests a smaller feed rate, larger amplitude, moderate rotation speed, and smaller cutting depth in addition to a WC tool coated with TiAlN and used under HPC cooling condition for optimal performance of AUVC. This recommendation is based on the theoretical analysis and experimental results of cutting force, surface roughness, and tool life.
In the present study, an analytical transient cutting force model for high-speed ultrasonic vibration cutting has been developed. Micro-machining characteristics are figured out when both the tool ...nose round and round cutting edge are taken into consideration in the cutting mechanism. Four cutting zones (i.e., elastic recovery, plowing, shearing, and tool-chip friction) are proposed and analyzed to calculate the transient cutting force on each divided cross section in the normal direction of the cutting tool along the cutting edge. A non-free cutting coefficient is added to evaluate the influence of the tool shape on the cutting force compared to the orthogonal cutting model. Afterwards, transient cutting force and average cutting force are simultaneously measured. The results show that the shape and value of the transient cutting force correspond well with what was calculated using the proposed model, and the availability of the proposed model is verified by the cutting force shapes and values.
Due to the continuous progress and development of new media technology, the living habits and ways of Chinese people have undergone profound changes due to the emergence and development of the ...Internet. Many college students are bombarded with information from the Internet, and they also gradually feel great social pressure. In addition, the Internet has become an important channel and platform for many college students to make suggestions, communicate and study. However, due to the characteristics of hidden, open and free travel of the Internet, college students can not only get access to positive and positive information, but also suffer from many unhealthy information infringements. Under the dual influence of social pressure and public opinion, college students will have a negative impact on the physical and mental development of college students. In serious cases, online mass incidents will occur, which will be detrimental to the construction of a harmonious society. Based on this, it is necessary to correctly handle the social pressure and public opinion related content of college students in the new media environment. In this paper, the connotation, characteristics and environment of new media are analyzed, combined with regression calculation, the pressure brought by new media to college students is further analyzed from four angles, and the four characteristics of college students’ public opinion are discussed in detail. Then, it puts forward the countermeasures of social pressure and public opinion from five dimensions.
Axial ultrasonic vibration–assisted cutting has been proposed and proved preponderance on cutting performance; however, the transient separation cutting characteristic has not been clearly revealed. ...This paper focuses on the transient separation cutting characteristic of axial ultrasonic vibration–assisted cutting and its influence on the cutting performance. First, the separation cutting criteria, including feed coefficient and phase shift, are established and influence factors, such as feed rate, amplitude, and phase shift, are systematically analyzed. Subsequently, the duty ratio is deduced and calculated, and some features and a control mechanism are introduced. In addition, a transient separation cutting model with four stages is demonstrated. The reason for the average cutting force reduction by the transient separation cutting characteristic is also stated. Finally, the verification of the separation cutting and transient separation cutting model and the comparison of the feed thrust force and parameter influences on the cutting performance are experimentally performed. The transient signals of different duty ratios 0.55 to 1 are obtained, and 10 to 40% reductions of feed thrust are measured both by a PCB sensor and a Kistler dynamometer. The influence degree of the feed coefficient, spindle rotation speed, and cutting depth on the performance is analyzed. A parameter combination of smaller feed coefficient, moderate spindle rotation speed, and smaller cutting depth is suggested to obtain a better cutting performance by fully considering the cutting force, surface roughness, and tool life.
PDF
