Abstract:Fault entity recognition is a basis of obtaining the knowledge of aeroengine fault autonomously, which plays an important role in realizing the intelligent fault diagnosis of aeroengine. For building up fleetly accurately aeroengine fault knowledge base, on the basis of the five kinds of aeroengine fault entity type defined, i.e. “unit”, “state failure”, “characterization of signals”, “inspection methods” and “solution”, a kind of BertBiLSTMCRF model of aeroengine fault entity recognition method preliminary is constructed. Based on the largescale data set analysis of an aeroengine, fault entities are extracted and fault knowledge map of abnormal oil pressure is constructed, verifying the effectiveness of the proposed method in identifying heterogeneous fault data of aeroengine

Abstract:Aimed at the problems that damage or rupture of the hydraulic pipe caused by the friction between the hydraulic pipes, or between the hydraulic pipes and the body structure or equipment cabinet can lead to failure of the system, a method is proposed for detecting the frictional damage of the aircraft hydraulic pipe based on acoustic emission technology. A correct judgment of pipes whether there is abrasion damage of the catheter stems from in analyzing the characteristic parameter, feature extraction and damage pattern recognition in the range of time domain through detecting the acoustic emission signal of the abrasion damage of the catheter. In this paper, the definition of pipe grinding is defined, the physical process and mechanical damage mechanism of pipe grinding are introduced in detail, the characteristic parameters of acoustic emission signal are analyzed and processed by using correlation graph analysis, and the damage detection experiment of hydraulic pipe grinding under vibration load is carried out on the aircraft. The results show that the detection method is feasible.

Abstract:A two channel deep convolutional neural network based method is proposed to predict the remaining useful life of aero engines. The method is to utilize maximal information coefficient for reducing data dimensionality noise, and Kalman filter for reducing data noise on the traditional convolution neural network. The realization of data reconstruction is subjected by data slicing and setting the data slice label as the remaining useful life of the last cycle. The segmented and linear remaining useful life decay models are introduced and a method to judge the starting point of life decay is given. The two features before and during life decay as the input of the twochannel network model are used. The results of testing on NASA turbofan Engine Simulation Data Set (C MAPSS) show that the relevant indexes of this method are significantly better than that of other algorithms when the test data range is large. And this method has significant advantages in the remaining useful life prediction of aero engines.

Abstract:Aimed at the problem that the coexistent attractors with different topological properties remain in the nonlinear energy sink (NES) system caused by the multi value character, a nonlinear vibration reduction strategy based on the attractor migration control method is proposed. Firstly, the influence of excitation force amplitude on the global behavior of the NES system in the large parameter range is analyzed. Secondly, an improved parallelized multidegreesoffreedom cell mapping (PMDCM) method is proposed to study the coexistent attractors and their basins under typical parameters. Finally, the attractor transitions with different amplitudes are realized by means of migration control. The simulation results show that the NES system in several typical parameter range presents multistable coexistent attractors phenomenon, the PMDCM method has higher efficiency and precision, and the NES system can be migrated from large amplitude attractors to small amplitude attractors by using open plus nonlinear closed loop (OPNCL), reducing vibration and noise.

Abstract:Aiming at the problems of low efficiency and poor accuracy of traditional steel surface defect detection methods, a steel surface defect detection algorithm based on improved YOLOv5 is proposed in this paper. Firstly, the C3 module and part of convolutional structure in YOLOv5 network are replaced by GhostBottleneck structure to realize the lightweight of network model. Secondly, SE attention mechanism is introduced in Backbone to strengthen the important feature channels. Finally, according to the characteristics of the data set, a detection layer is added to the network to strengthen the feature extraction ability, and a feature fusion structure is added in the Neck part. DW convolution is used to replace part of the standard convolution to reduce the computation. Experimental results show that the improved Yolov5sGSD algorithm reduces the model volume by 10.4%, and the mAP value on the test set is 76.8%. Compared with the original YOLOv5s network, the detection accuracy and speed are obviously higher than some mainstream algorithms. Compared with traditional steel surface defect detection methods, the algorithm proposed in this paper can detect the type and location of steel surface defects more accurately and quickly, and has a smaller model volume, which is convenient for deployment in mobile terminals.

Abstract:Aimed at the problems that the traditional algorithm is low in efficiency and low in accuracy of the direction of arrival (DOA) estimation in coherent targets under conditions of strong interference background, on the basis of constructing the interference blocking matrix to eliminate strong interference in a specific direction, a DOA estimation of coherent targets is made based on the local spatial difference algorithm. First, the blocking matrix is constructed by the angle information of the received signal matrix to eliminate the strong interference information of the covariance matrix, and then the covariance matrix is divided into several submatrices using the local spatial difference algorithm, fully extracting the effective information. Finally, the linear operator method is utilized for completing the DOA of the target estimation. The algorithm effectively improves the data utilization. The simulations show that the proposed method is low in estimation deviation and high in success probability in comparison with the traditional algorithm under conditions of strong interference background.

Abstract:Delay forwarding deceptive jamming is an important jamming method in the light of the inverse synthetic aperture radar (ISAR) imaging. In order to counter the false target caused by jamming, an ISAR deceptive jamming recognition and elimination method is proposed based on orthogonal pulse diversity. Firstly, a jamming signal model of ISAR is established, and the constraint conditions of orthogonal pulse design for this jamming are studied and given. A group of semi orthogonal pulses are designed to eliminate and identify the delay forwarding deception jamming. This method can detect and eliminate the false target caused by delay forwarding jamming by orthogonal design of inverse synthetic aperture radar waveform, and achieve the corresponding antijamming of inverse synthetic aperture radar. Finally, through the simulation experiment, the sidelobe waveform with 0.3 times the peak value is generated under the condition of 20 dB interference signal ratio, and the ISAR image is successfully obtained. At the same time, the existing methods being compared with high signal to noise ratio of more than 5 dB, the peak signal to noise ratio of ISAR image is improved by about 5 dB, and the method is valid

Abstract:An adaptive iterative physical optics algorithm is proposed by studying the complex electromagnetic scattering characteristics of two dimensional multi roughness layered rough surface and target above, and a composite model of twodimensional layered rough surface and the target above is established by adopting the Monte Carlo method in combination with the generated Gaussian rough surface with Gaussian spectrum function based on sub regional modeling method. The direct induced electromagnetic current of the layered rough surface and the target is obtained by using the physical optics method and the equivalence principle. Based on the surface integral equations, the iterative mechanism of the coupled electromagnetic current between the layered rough surfaces and between the rough surface and the target is deduced. The traditional iterative physical optics method is improved by introducing the change rate of induced electromagnetic current energy, making the algorithm convergent automatically. The results are compared with those of the multilevel fast multipole algorithm and iterative physical optics method, and the accuracy and efficiency of the algorithm are verified. On this basis, the bistatic RCS calculation results and scattering characteristics of layered rough surfaces with different targets and different roughness are studied. The effect of the distance between layered rough surfaces and on the bistatic RCS calculation results and scattering characteristics is discussed. This study provides data support and theoretical basis for the detection, classification and recognition of ultra low altitude targets above a layered environment.

Abstract:Aimed at the problems that traditional emitter signal identification methods often need to carry out artificial feature extraction and signals are difficult to be identified accurately under condition of low SNR environments, a method of emitter signal recognition based on improved UNet3+ network is proposed. By trimming the UNet3+ network hierarchy, the feature fusion ability is retained while the complexity of the network is reduced. The attention mechanism is introduced to optimize the model performance, and a new network model is constructed. The simulation results of eight common radar signals show that the recognition accuracy of the improved model reaches 96.63%. Compared with some classical network models, the total training time is shorter, and the ability to identify the radiation source signal is more effectively under condition of low SNR environments. And the proposed model can also be adapted to the complex electromagnetic environments.

Abstract:In satellite communication systems, frequencies and channels are very rare resources. How to use reliable and efficient methods to develop resources has become a severe problem that needs to be solved urgently. This paper proposes a dynamic satellite channels allocation algorithm DRLDCA. This algorithm is to model satellite and environment interaction on a Markov decision making process, improving satellite decision making ability through environmental feedback, realizing efficient response to user business requests, improving the service quality of satellite communication, and reducing the probability of communication blocking. The simulation analysis shows that the proposed algorithm can effectively improve the communication throughput and reduce the communication blocking rate.

Abstract:Due to low temperature and harsh environment, there is ice coating problem in distribution network lines, which has great harm to the reliability of power supply system. Hydrophobic film or coating can be used to effectively prevent the ice coating problem. Preparation of hydrophobic coatings for adaptive circuits with high efficiency and low cost is a practical subject. A hydrophobic surface was prepared by plasma modification of polydimethylsiloxane (PDMS) with mixed gas of He(Ar)/CH4/C4F8. The surface can cover the outer layer of the circuit to provide hydrophobic properties. The effects of discharge voltage level, distance between discharge electrodes, discharge time and discharge frequency on the preparation of hydrophobic surface modified by plasma were studied. COMSOL software was used to analyze the electric field intensity distribution of PDMS microchannel and explore the distribution of plasma. Based on the simulation results, physical experiments were designed to verify and analyze the mechanism of each factor[HJ] by measuring the contact angle, roughness and surface topography of PDMS modified surface and taking gas discharge images. Finally, suitable conditions were selected to successfully prepare the PDMS hydrophobic surface in line with expectations, and its good stability was tested.

Abstract:Aimed at the problems that the performance of laser modulator with high power chips in optical modules becomes deteriorated, and bit error rate caused by the clock reference deviation between transceivers is high, a low power and high jitter tolerance clock and data recovery circuit (CDR) is proposed in this paper. By adopting a technology of voltage controlled oscillator (VCO) type and full speed CDR system architecture, and using a technology of an inductance peaking in clock buffer, the power consumption of the CDR chip is reduced. By introducing a zero point compensation resistor in the CDR integration path, the jitter tolerance of the CDR is improved. The CDR is designed with CMOS 65 nm process and supplied with 1.1 V. The back end simulation results show that when the CDR circuit works at 28 Gbps, the power consumption is 2.18 pJ/bit. When the frequency difference of the transceiver is 5 000 ppm, the jitter peak to peak value of the recovered clock is 5.6 ps, and the jitter tolerance meets the needs of design index and the CIE25/28G protocol specification.

Abstract:In view of low bandwidth and pin efficiency in Dieto Die (D2D) of high performance CPU, GPU, AI and other high end chips, a 125 Gb/s transmitter with low power consumption with high pin efficiency for ultra short reach (USR) is proposed. To improve pin efficiency, this circuit adopts correlated non return to zero (CNRZ) encoding technology. To reduce the transmitter’s power consumption, this paper adopts a precoding soucre series terminated（SST） driver technology. To solve the problem of high power consumption of traditional two stage 2∶1 MUX, this paper adopts a CMOS 4∶1 MUX technology. This transmitter is designed with 28 nm CMOS technology and supplied with 0.9 V voltage. The simulation results show that when the transmitter based on CNRZ technology works at 125 Gb/s, the minimum eye width of the output signal can reach 0.41 UI (1 UI=40 ps), and the system power consumption is 1.1 pJ/bit, the pin efficiency increases from 5bit/10wire to 5bit/6wire.

Abstract:In consideration of the influence of infection timedelay on the transmission of zero-day virus, the time-delay transmission model of zero-day virus and its stability under the background of platform dynamic defense are studied, the following measures should be taken in the research. Firstly, the influence of time-delay on the propagation of zero-day virus is analyzed, and the Dynamic Platform Defense model is introduced, and a SIZRO model of zero-day virus propagation is established. Secondly, according to the Lyapunov stability criterion, the local stability of equilibrium point in system is proved, and the basic reproduction number and its effect on the zero-day virus transmission are analyzed. Finally, the number of infected nodes in SIZDR model being compared with that in SIZRO model, the feasibility of using platform dynamic defense to solve zero-day virus transmission problem is demonstrated. The theoretical analysis and simulation results show that the proposed model can objectively reflect the time-delay propagation and immune law of zero-day virus, and the Dynamic Platform Defense can effectively improve the defense effect of the system against zero-day virus.

Abstract:Aimed at the problems that the selection of kernel function and parameter adjustment in SVM algorithm are not scientific, and the detection is poor in accuracy of classification, a Grey Wolf Optimization Algorithm based on Particle Swarm Optimization (PSOGWO) algorithm is proposed to improve the Intrusion Detection System (IDS) based on SVM. This method is to utilize PSOGWO algorithm for optimizing the parameters of SVM to improve the overall performance of intrusion detection based on SVM. The optimal detection model of SVM classifier is determined by the fusion of PSOGWO algorithm and SVM. The comparison experiments are made based on NSL-KDD dataset, and the results show that the intrusion detection method based on PSOGWO-SVM achieves the optimization of the parameters of SVM, improving significantly the detection rate, the convergence speed and the model balance. And this algorithm is feasible.

Abstract:In view of the insufficient durability of ordinary mortar in practical engineering application,the durability modification effect of VAE emulsion powder on cement mortar was studied.For VAE latex powder modified mortar with different polymer content (8%, 9%, 10%, 11% and 12%),through shrinkage,cracking sensitivity,impermeability,chloride corrosion resistance and carbonation experiments,the shrinkage,total weight of cracking,water seepage height,water permeability pressure,carbonation depth and chloride ion penetration height were tested.Combined with SEM experiments,the micro modification mechanism of VAE emulsion powder on mortar is revealed.The results show that the addition of VAE emulsion powder significantly improves the impermeability of mortar.When the polymer content is 12%,the total cracking weight of VAE latex powder mortar is only 55.1% of that of ordinary cement mortar,the permeability pressure is 78.6% higher than that of ordinary cement mortar,and the 28 day carbonation depth and chloride ion penetration height are 36.8% and 53.1% lower than that of ordinary cement mortar respectively.VAE emulsion powder and the aggregate in the mortar form a network membrane structure with high adhesion inside the slurry,fill the pores of the modified mortar,make the interior of the slurry more dense,and then enhance the durability of the modified mortar.