Abstract:Aiming at the problems of premature convergence, poor ability in global optimization and low accuracy in local optimization for support vector machine ( SVM) based on parameter optimization, an adaptive particle swarm optimization / cuckoo ( APSO/ CS ) parameter optimization method is proposed, in which the optimization of kernel function parameters and penalty factors in SVM model is realized. The optimization performance of APSO/ CS, APSO and CS is compared and analyzed by test functions, which shows that APSO/ CS can accelerate the convergence speed of local and global optimization. The gesture recognition methods based on surface electromyography signal ( sEMG) by APSO/ CS, APSO and CS are compared. The experiment results show that the SVM method optimized by APSO/ CS can realize the highest recognition accuracy, which is about 94. 50%. The proposed method can provide a new way for the recognition and classification algorithm.

Abstract:The study of flow-acoustic coupling mechanism is crucial to the development of large diameter ultrasonic flowmeters applied in industrial piping systems. An ultrasonic flowmeter with an inner diameter of 500 mm is simulated, mainly focusing on the coupling mechanism between the flow field and acoustic field. The distribution characteristics of flow field in the ultrasonic flowmeter are analyzed. The effects on the flow field and acoustic field caused by the installing structure of transducers are revealed. The acoustic characteristic and the acoustic pressure distribution on the transducer receiver end face are shown. The flow velocity and the system deviation of the ultrasonic flowmeter are calculated by the difference in travel times between the downstream and the upstream propagations. The in-depth study of complex flow-acoustic coupling mechanism has significant meanings for the improvement of measurement performance of the ultrasonic flowmeter.

Abstract:Aiming at the problems of low reliability and sensitivity when evaluating the degradation of bearing performance with traditional characteristic indicators, a method for evaluating the degradation of rolling bearing performance based on dispersion entropy and cosine Euclidean distance is proposed. First, the vibration signal of the rolling bearing to be tested is divided into health data and test data, decomposed by EEMD respectively to obtain several Intrinsic Mode Functions (IMF). Calculatethe correlation coefficient between each IMF component and the original signal, and the IMF components are selected according to the correlation coefficient criterion to reconstruct signal. Then, the dispersion entropy of the reconstructed signal is calculated, and the Euclidean distance and the cosine distance are combined to obtain the degradation index cosine Euclidean distance between the health data and the test data dispersion entropy. Finally, the Chebyshev inequality is used to calculate the cosine Euclidean distance health threshold to evaluate the degradation of the bearing performance. The experimental result shows that the cosine Euclidean distance between dispersion entropy can effectively and timely judge the degradation state of the bearing performance, and compared with other indexes, its sensitivity and robustness are higher, which can better describe the degradation trend of the rolling bearing performance, and provide a new solution for the evaluation of the rolling bearing performance degradation.

Abstract:To reduce the energy consumption of wireless sensor network nodes and prolong the lifetime of the network, the optimal location of graph theory is applied to the clustering routing protocol. In order to reduce and balance the energy consumption in the cluster, the cluster head election is carried out according to the residual energy of the nodes and the center theory algorithm of graph theory. According to the energy of the next hop cluster head, Dijkstra algorithm is used to build the best route and determine the best hop number to save energy consumption of data transmission and balance the whole network energy consumption. The algorithm is applied to the airport single light monitoring system, and the simulation results show that compared with AL-CAME and ECOMP, the algorithm based on graph theory has the least network overhead, the residual energy value is far greater than the other two algorithms, the distribution of network energy consumption is more balanced, and the life cycle of the network is effectively extended.

Abstract:The surface damage during the use of parts will affect the interface bonding quality during remanufacturing and repair, so that reducing the bearing capacity. In order to obtain the effect of surface damage on the load carrying capacity of repaired parts, plasma arc welding was used to repair the matrix materials with different width of the surface prefabricated damage, and ultrasonic phased array was used to detect the repaired parts with three-point bending load. According to the test results, the sound path of the second bottom wave is extracted from the A-scan waveform, and the change curve of the sound path of the second bottom wave under different loads is obtained, which can get the load at the interface crack initiation of the repair part ( FA ) from the turning point of the curve. By quantifying the defects in the C-scan top view and calculating the area proportion, the change curve of the defect area proportion under different loads is obtained, which can get the load at the interface initial defect evolution of the repair part (FC ) from the turning point of the curve. The results show that both FA and FC can reflect the linear relationship between the load carrying capacity of the repaired parts and the surface damage of the matrix; on the premise of the error less than 5%, the load carrying capacity of the repaired parts decreases by 0. 1 kN for every 1 mm increase in the width of the surface damage. Moreover, compared with FA , the size of FC decreases by 11. 12%, which provides a timely and effective way to evaluate the remanufacturing quality of surface damaged parts.

Abstract:Aiming at the problem of insufficient accuracy of wave head time calibration and unstable wave velocity in traveling wave method, a single-ended fault traveling wave ranging method based on GST-TT transform is proposed. Firstly, the sampled voltage traveling wave signal is decoupled to extract the line mode voltage signal; Secondly, the line mode voltage signal is S-transformed in general; Thirdly, the appropriate high-frequency band is extracted as the fault characteristic band; Fourthly, Using high-frequency aggregation and low-frequency suppression characteristics of TT transform’s diagonal position elements, the arrival time of the wave head at the measuring end is accurately calibrated in the fault characteristic band; Lastly, the fault distance is obtained based on the fault ranging calculation method which does not depend on the specific traveling wave velocity. The method can improve the calibration accuracy of the wave head time and reduce the influence of the wave speed instability on the fault location. The PSCAD simulation proves that the method has high accuracy and the average relative error is 0. 25% in the absence of noise interference. It also has strong tolerance to transition resistance and anti-noise ability.

Abstract:Due to the complex and changeable geological environment, uneven thrust, uneven grouting and uneven shield clearance, etc. , There is often a deviation between the actual jacking trajectory of the pipe jacking machine and the expected trajectory during the construction process, and large deviations may cause construction accidents or serious economic losses. Aiming at the above problems, an attitude measurement system based on the combination of a photoelectric sensor array and a PLC is designed in the method of coordinate conversion. A related real-time display interface is developed, which is used for real-time display of the three angle attitudes of the pipe jacking machine. And the projection method is used to calculate the real-time display of the trajectory of the pipe jacking machine to assist the attitude control of the pipe jacking machine. In the experiment, artificially simulate the movement deviation of the pipe jacking machine head to verify the accuracy of the measurement system. Experiments show that the coordinate recognition accuracy of the point source of the laser measurement system is up to 3 mm, the angle measurement accuracy is 0. 1 degree, and the maximum error is about 0. 784 degrees, which meets the requirements of pipe jacking construction, facilitates the attitude measurement of pipe jacking, and improves the quality and efficiency of pipe jacking construction.

Abstract:In this paper, isolation forest algorithm is used to construct an abnormal parameter monitoring model for surge protectors. By making the model learn the distribution of normal SPD parameter data, the abnormal parameter data of SPD that cannot fit into the distribution will be identified, then the actual state and deterioration degree of the device can be evaluated and warned. The study first performs multi-parameter sampling on the SPD, pre-processes the obtained 390 sets of experimental data, and combines the algorithm model to calculate the abnormal value corresponding to each set of sampling data to identify the corresponding abnormal sampling data. Finally, according to the label corresponding to each set of data, Verify the accuracy of the algorithm. The test results show that the monitoring model based on the isolation forest algorithm can evaluate the abnormality of the SPD sampling data, and lock the parameter values corresponding to the SPD that may appear bad or degraded in the system. For the experimental data set, the performance index AUC value of the algorithm model under different parameters is not less than 96%, which provides a new idea for the monitoring research of SPD performance.

Abstract:Considering that the minimum entropy deconvolution (MED) was only sensitive to a single abnormal vibration signal, and the length of the filter needed to be adjusted manually, a maximum kurtosis entropy deconvolution (MKSED) method was proposed and applied to bearing fault diagnosis. Considering that kurtosis entropy has the advantage of continuous shock oscillation, kurtosis entropy was chosen as the objective function of deconvolution. At the same time, kurtosis entropy was used as the fitness function of the improved local particle swarm optimization algorithm (LPSO), and LPSO was used to optimize the filter length, so that MKSED can adaptively adjust the filter length when deconvolution, so as to accurately extract the continuous pulse signal. The experimental results show that the method can extract continuous pulse signal more effectively and improve the accuracy of fault diagnosis.

Abstract:In general, vibration signals generated by the switching operation of a power distribution switch contains important information to reflect its mechanical status. A novel type of fault diagnosis method for a power distribution switch based on vibration signals analyses is proposed in this study. Firstly, the Choi-Williams distribution ( CWD) for the vibration signal is calculated to obtain a twodimensional time-frequency matrix. Then, the block singular value decomposition (BSVD) is performed on the two-dimensional timefrequency matrix, which is used to characterize the time-frequency characteristics of different mechanical states. Finally, the extreme learning machine (ELM) classification algorithm is adopted to train and test the feature vectors of four mechanical states of measured vibration signals. The advantages of the proposed method are that the time domain and frequency domain characteristics of vibration signals inside the power distribution switch are effectively extracted, and the diagnostic model can be trained without many samples. Experiments based on measured data show that the proposed method has a higher recognition accuracy with a faster convergent speed. Keywords:distribution switch; mechanical state recognition; Choi-Williams distribution (CWD); block singular value decomposition

Abstract:As the grounding electrode of HVDC is in single electrode blocking condition, the large stray current into the ground may cause the electromagnetic valve misoperation risk, affecting the normal operation of the pipeline. The mathematical model of the influence of conducting interference on the solenoid valve coil is established, the mechanism of misoperation of electromagnetic valve caused by dc stray current is studied. By means of numerical simulation, a simulation model of grounding electrode interference to nearby pipelines in a 1 100 kV UHVDC transmission project in China was established, and a real-time data acquisition system of ground potential was designed. Simulation and field test results show that the dc interference of the electromagnetic valve in winter is 3. 43 times that in summer. The simulation results are basically consistent with the measured results, and the error is within 10%. After adding 4 and 6 insulated joints along the pipeline, the ground potential difference of the pipeline is reduced by 62. 1% and 71. 4%. In order to reliably prevent abnormal operation of electromagnetic valve, inspection of valve chamber should be strengthened in winter, and insulation joint should be installed along the pipeline upstream and downstream.

Abstract:In order to solve the calibration problem of air negative ion measuring instruments, based on the working principle of air negative ion monitor, a method for tracing the source values of instruments and a method of developing the calibration equipment are proposed. The developed negative ion generator and standard negative anion measuring instrument are tested and verified. The verification results show that the anion field produced by the anion generator has the characteristics of continuously adjustable anion concentration, good stability, fast response speed and small particle size, the relative error of the standard negative ion measuring instrument is not more than ± 5%, and the measurement uncertainty of the standard device is 6%, which can be used for the calibration and test of the negative ion measuring instrument. Using a calibration device composed of the negative ion generating device and the standard ion measuring instrument, a calibration test is performed on two different negative ion measuring instruments, and the measurement errors are analyzed to verify the scientific rationality and feasibility of the calibration method.

Abstract:In view of the low accuracy of the traditional pattern recognition algorithm for qualitative and quantitative detection of mixed gases, a novel algorithm of hybrid gas qualitative identification and concentration quantitative detection based on machine learning is proposed. The algorithm constructs the feature map of sensor array data first, then uses the convolutional neural network ( CNN) to extract features from feature maps. According to the feature map after feature extraction, different branches are used to identify different gases, then the species of gases and their concentration range were obtained; based on the results of gas identification, the kernel principal component analysis (KPCA) and gradient boosting decision tree (GBDT) were used to estimate the composition of the mixed gases quantitatively. Finally, this paper used the dataset of sensor array of mixed gases of Machine Learning Database of the University of California to verify the results. Experimental results show that the accuracy of the algorithm in the qualitative recognition of ethylene and methane reaches 98. 7% and the average relative error of quantitative detection was less than 4. 1%. Compared with the traditional pattern recognition algorithm, the machine learning based mixed gas detection algorithm that proposed has higher accuracy and stronger generalization ability.

Abstract:In view of the existing algorithms, traditional video emotion-based facial expression recognition method only pays attention to the spatial features of a single video frame, and ignores the hidden temporal features between adjacent frames. Therefore, this paper proposes a novel method to extract features using edge detection and improved recurrent neural network. Gradient edge detection can extract texture information of video frame in a more accurate way,at the same time, a kind of overlapping LSTM structure is proposed, and the recurrent neural network can acquire the hidden spatio-temporal information from the input frames. The experiments in this paper are carried out on the CK + and MMI video database. the result of 88. 4% and 69. 7% are obtained in the OCNN-RNN network respectively, and the outcome of 89. 8% and 73. 6% are acquired in the GCNN-RNN network from each database. and finally the random search is used to weight the fusion of the results of the GCNN-RNN network and the OCNN-RNN network. After the two networks are finally merged, the average recognition rate of the integrated model is 94. 6% and 79. 9% respectively, and the accuracy is better than other algorithms, the effectiveness of the proposed algorithm is proved.

Abstract:According to the amplitude deviation and its rate of change, the vibration process is divided into eight characteristic states, and the corresponding control modes are designed by using Heuristic and intuitionistic reasoning respectively. In order to solve the problem that the vibration characteristic state identification parameters and the control modal parameters are too many and not easy to set, the quantum genetic algorithm is used to optimize the controller parameters. The optimization principle, optimization flow chart and realization steps are given, and the experimental verification is carried out. The results show that the HSIC controller with optimized parameters can reduce the starting time by 0. 1 and 0. 6 s compared with PID controller, and its amplitude is stable, which proves the performance of the method.

Abstract:Aiming at the problems that the single-shot and single-acquisition water-acoustic ranging method in the shallow sea environment has high computational complexity and is susceptible to the special underwater environment, a low-complexity ranging method suitable for underwater acoustic multipath channels is proposed. Firstly, the composite code is applied to the underwater acoustic positioning system as a synchronous ranging signal, which supports fast and accurate signal synchronization and multipath delay detection. Then, an improved weighting function generalized cross correlation (GCC) algorithm is proposed to extract the multipath delay. Finally, a matching function is introduced to multiply the multipath delay. The path delay is correctly matched to the propagation path, and the geometrical methods such as hyperbola and triangle are used to calculate the distance between the transmitting and receiving ends. Simulation results show that in multipath and low signal-to-noise ratio ( - 10 dB) environments, the synchronization accuracy of the composite code is 100% and the computational complexity is low. Compared with the phase transformation-generalized cross correlation (PHAT-GCC) algorithm, the multipath of the GCC algorithm with an improved weighting function is improved. The delay detection performance is improved by nearly 6 dB, and based on this, a single receiving node can achieve ranging of underwater targets.

Abstract:To solve the problem of metal artifacts in reconstructed CT images, a method based on multi-threshold segmentation and B-spline interpolation is proposed to remove metal artifacts in CT images. Firstly, the causes of metal artifacts in CT images are introduced. Secondly, the implementation process of metal artifacts reduction method of CT image based on multi-threshold segmentation and B-spline interpolation is analyzed in detail. Finally, this approach is tested on clinical CT and simulated phantom images. Experiments show that the proposed metal artifact removal approach can effectively reduce or remove the metal artifacts in CT images, which provides clear CT images for the diagnosis of lesions in medical work.

Abstract:A method based on a comparison with the coordinates of the peaks in gray scale was proposed to detect the assembly defects of thermal battery. The stack of thermal battery was extracted by template matching. Then image preprocessing methods such as affine transformation, gamma correction and image pyramid were used to improve the contrast of the gray scale. Five defects included the deficiency of monomer thermal battery, the missing of negative electrode, the missing of current collector, wrong assembly order and the overall flip-chip of thermal battery were analyzed. Characteristics in defective battery such as the number of peaks and valleys in gray histogram, distances between the peak and valley were compared with corresponding ones in standard battery. Verified by 400 test images, the experimental results showed that the proposed method possesses an accuracy of 95. 5% and proved that this method can quickly and efficiently detect the defects in thermal battery.

Abstract:Magneto-inertial navigation system (MINS) is widely used in navigation control of unmanned aerial vehicle. It can provide attitude and heading information via data fusion of accelerometer, magnetometer and gyroscope. Complementary filter is a concise and efficient solution for data fusion in frequency domain. Therefore, it is also called attitude and heading reference system (AHRS). A generalized complementary filter (GCF) is introduced to consolidate various complementary attitude filters that based on different transfer functions and orientation representations. In this paper, the multiplicative attitude error of this kind of algorithm is analyzed, and compensation algorithm for motional acceleration is also proposed, in order to enhance attitude accuracy under dynamic condition. Numerical simulation and experiment results show that GCF can achieve the same performance of extended Kalman filter, but its execution time is approximately 1 / 20 of the latter's. Moreover, GCF has ideal numerical stability, and it can eliminate the impacts of linear acceleration with the proposed motion compensation algorithm. Therefore, it is suitable for miniature and low-cost UAV applications.

Abstract:In view of the vibration problem caused by the convergence of the planetary planets in the integrated transmission of the tracked vehicle, the paper designs and builds the vibration signal of the load-carrying movement of the converging planetary platoon in different positions of the busbar vibration test rig. It analyzes the variation of the time domain and spectral characteristics of the confluent planetary platoons under different gear positions. The results show that under the same gear position, the higher the load power, the greater the vibration intensity, and the z-direction vibration is stronger than x, y; the spectrum results show that the y-direction and zdirection vibration energy are mainly concentrated in the meshing frequency and frequency multiplication, z The frequency amplitude is slightly higher than the y direction, In the high-end position, the amplitude of the meshing frequency is concentrated in the 1st and 3rd steps, and its amplitude changes obviously. It has certain reference value for studying the vibration characteristics of the confluent planetary row under complex conditions.

Abstract:In different stages of product life cycle, including design, simulation, manufacture, test and operation and maintenance, bag filter generates a large amount of data. It excavates the complex, non-linear and coupling internal relationship between big data of product and its operation characteristics, and provides a new way to solve the common problems of design innovation and operation and maintenance optimization in bag filter industry. Aiming at the characteristics of large data of bag filter, a large data mining XGBoost model for on-line monitoring of bag breakage of bag filter is proposed, and the parameter optimization method of XGBoost model based on ant colony algorithm is studied. Compared with Stochastic Forest and BP network mining models, the results show that the XGBoost optimization model method has high accuracy and strong explanability.

Abstract:Aiming at the problem of the accuracy of the recognition of the operating state caused by the unbalanced data acquisition in the normal and blocked fault state of the drainage pipeline, a method for a pipeline clogging state recognition based on cost-sensitive support vector machine based on fruit fly optimization algorithm is proposed. According to the unbalanced data set collected under various operating conditions in the drainage pipeline, the wavelet packet decomposition is first performed on the unbalanced data set. Secondly, the energy entropy of each decomposition coefficient and the approximate entropy index are used to construct the feature vector set. The fruit fly optimization algorithm is adopted. (FOA) optimizes the penalty factor Cm and the kernel function parameter g, that is, the costsensitive support vector machine ( CS-SVM) model optimization, and inputs the feature set into the optimized CS-SVM model to normalize the drainage pipe. Blocking state recognition, by increasing the penalty cost of misclassification of a few types of samples, the results show that the recognition accuracy of a few classes is improved.

Abstract:The region of China was divided into four climate zones. The method of double weighted standard deviation and correlation coefficient threshold was adopted to take the sounding data of 120 stations in China as reference, the COSMIC, Metop-A, and Metop-B occultation inversion temperatures in 2014 were used as samples for quality control. It is also compared with the traditional standard deviation method, and the quality control results are tested. The results show that the three occultations are similar, and the temperature in different climatic regions has different characteristics, so that different threshold intervals are divided to make the quality control more accurate. The error data identified by the traditional standard deviation method can be identified by the double weight standard deviation method, and the double weight method can also detect error points that the traditional standard deviation method cannot detect, and the double weight standard deviation method is more suitable. Through statistical calculation, the correlation coefficient threshold of GNSS occultation inversion temperature and sounding temperature is determined to be 0. 860 9, so as to determine whether the suspicious data is erroneous data. From the quality control results, the distribution of erroneous data is mostly the difference between the occultation temperature and the sounding temperature. After eliminating the erroneous data from the quality control, the correlation between the occultation inversion temperature and the sounding temperature is obvious improvement, data quality can be better improved, this method is applicable to the quality control of GNSS occultation inversion temperature, and provides more accurate observation data for weather analysis and numerical prediction.

Abstract:In order to improve the output torque and speed, and reduce the permanent magnet (PM) eddy loss for magnetic gear, a novel magnetic gear with stationary PM ring is proposed. The operating principle and torque performance are analyzed, and the results showed that the novel magnetic gear has high efficiency, high torque density and high speed compared with traditional magnetic gear. Taking high output torque and low cogging torque as the optimization goals, a multi-objective optimization method based on the parameter sensitivity and response surface method is introduced, which aims to accurately obtain the magnetic gear structure parameters. Then, the PM ring, outer rotor and inner rotor are optimized. By analyzing the electromagnetic performances for novel magnetic gear between initial and optimal magnetic gear, the results show that the output torque is improved, and it lays a foundation for its application and optimization design.

Abstract:In the process of spheroidization audio detection, the selection of exciting position directly affects the amplitude of the modal frequency response of each order directly. For similar modal frequency responses, it is likely to affect the measurement results. In order to extract the modal frequencies of interest and suppress the modal frequencies of no-concern, a finite element-based optimization method for exciting position is proposed. Taking the car’s part flywheel as an example, modal analysis is performed on the spheroidal graphite cast iron test piece through ANSYS, combining the characteristics of each order mode, natural frequencies and modal excitation theory to determine the optimal excitation point, and the audio signal of the exciting positions is analyzed by MATLAB GUI verification. The experimental results show that this method can effectively stimulate the modal shapes required for the observation of the flywheel structure. It is verified that the close modal frequency response is closely related to the choice of the exciting position. Effectively distinguishing modal shapes of similar frequencies has strong practicability.

Abstract:In the process of micro and nano scale, the probability of an open circuit fault in the embedded memory increased, which resulting in dynamic faults. When static faults and dynamic faults coexisted, the traditional pause-and-export design of built-in self-test designs could correctly output fault diagnosis data, but there was a problem of diagnostic data redundancy. Therefore, a built-in self-test design for dynamic fault diagnosis data compression was proposed. Under the premise that it was not affecting the integrity of the diagnostic data, the problem of redundancy of diagnostic data was solved by identifying the fault modes as line faults, column faults and unit faults, and compressing the diagnostic data. The simulation results show that the design can correctly compress the dynamic fault diagnosis data, greatly improve the output efficiency, reduce the output time, and the area overhead is small. The area overhead of the 8 K×16 memory is 3. 16%, and the diagnosis data compression ratio is 3. 96% under 20% row failure and 5% dynamic failure

Abstract:Based on the traveling wave theory, the distributed parameters equivalent model of 27. 5 kV cable was built and verified by simulating the speed of lightning wave in different segments. Make sure the necessity of the fitted protector on metal sheath and the reasonableness of protection provided in the first end by theoretical calculations. On this basis, when the lighting current intruded upon the cable core along the catenary, one software was adopted to simulate and study the influence of the cable length, grounding resistance of the metal end and impedance of the core end on the lighting induced voltage of cable metal sheath. The results show that the average of the maximum value of sheath lightning induced voltage may be the smallest when the cable length is 800 m; when the grounding resistance value is in the range of 2. 5~ 3 Ω, the average of the maximum value of sheath lightning induced voltage may be the smallest, and they are lower than 25 kV, if the grounding resistance value is more than 5 Ω, the maximum value of sheath lightning induced voltage may exceed the lightning impulse withstanding voltage of outer sheath, leading to the breakdown of outer sheath; when the cable end is connected to a short circuit or a larger capacitive impedance, the sheath lightning induced voltage may be very large and may go far beyond the lightning impulse withstanding voltage of outer sheath, while the cable end is connected to an open circuit or an inductive impedance, the sheath lightning induced voltage may be smaller.