Abstract:Isothermal titration calorimetry (ITC) is an important method for the thermodynamic measurement of solution reaction. Based on the quantitative evaluation of ITC basic error sources, an explicit ITC measurement model is formulated, which consists of both different thermal effects and main error components. By utilizing this ITC measurement model, BaCl 2 / 18-crown- 6, CaCl 2 / EDTA and 1-Propanol dilution are used to calibrate the calorimetric factor (f) and the effective volume of the reaction cell (Vcell) of the NanoITC (SV) calorimeter. Based on the concentration uncertainty of the titration solution, the basic error terms of the calorimeter and the uncertainty of the theoretical values of the reaction system, the calibration uncertainties can be calculated by the Monte Carlo method. Three groups of different calibration experiments are compared with each other. The results show that an 8. 1% deviation in the calorimetric factor and a 3. 4% deviation in the effective volume of the reaction cell. Supplementary materials are provided in terms of experimental data and Python source code.

Abstract:During the measurement of large space objects with the laser tracking multilateral method, the measurement accuracy is affected by a lot of factors. The accuracy of the self-calibration is one of key factors. In this paper, a self-calibration optimization method is proposed. Through introducing the cross spatial distribution model of the points for the optimized calibration, the data of the laser base stations obtained from the initial self-calibration are calibrated once again, and the optimization algorithm is used to get the more accurate data of the laser base stations, so as to reduce the error in the self-calibration process and improve the accuracy of laser tracking multilateral measurement. The distance measurement experiment on a groups of space points in the x-axis direction was carried out, the optimization effect of the data of base stations was verified through introducing calibration points. The experiment results indicate that the distances of the measured points obtained based on the data of the base stations after the optimized calibration are closer to the reference distances, and the distance difference is reduced from -2. 3 μm to -1. 6 μm after optimization at the position with the x-axis coordinate of 750 mm. So, the self-calibration optimization method proposed in this paper could improve the accuracy of the laser tracking multilateral measurements.

Abstract:To improve the vibration environment of the micro-nano space camera during launch and avoid damage, a passive vibration isolation system is designed for the mission requirements of micro-nano space camera. Firstly, the camera vibration isolation system model is formulated and the response characteristics of the system is studied. Then, the vibration isolation system is decoupled and compared with the traditional method. Based on the camera parameters and the vibration isolation system, the passive rubber vibration isolator structure is calculated and designed. The fundamental frequency is 120 Hz and the mass fraction of camera is 1. 7 % . Finally, the dynamic simulation of the vibration isolation system is implemented by the finite element analysis method. The vibration tests of the space camera and the vibration isolation system are carried out. Experimental results show that the isolation efficiency of the passive vibration isolation system for the sinusoidal vibration and random vibration response of the micro-nano space camera can reach 55 % and 81 % , respectively. The validity and rationality of the vibration isolation system are verified.

Abstract:The dual-material micro-cantilever structure is adopted to realize a quantitative detection method for the micro-quantity liquid. The method is based on the temperature sensitivity of bi-material micro-cantilever. When the micro-quantity liquid analyte absorbs the light of the characteristic wavelength of the photothermal spectrum, it′s temperature changes, the cantilever will be bent accordingly, and the amount of bending is measured with the optical lever method, so that the quantitative detection of the micro-quantity liquid is realized with the micro cantilever. In order to verify the applicability of measuring the micro-quantity liquid, the NaYF4 sample solution with a known light-heat absorption peak at 980 nm was used in the experiment, and multiple sets of 1 and 0. 5 μL solutions with known concentrations were taken for measurement. The relationship between the concentration and the measured voltage value was fitted by the least square method. The results show that the liquid concentration has a linear relationship with the measured voltage value, which is consistent with the theoretical analysis. And the linear correlation coefficients of the fitted curves of the sample solutions with the volume of 1 and 0. 5 μL are R1 = 0. 998 0, R2 = 0. 989 5, and the resolutions are 1. 25, 1. 40 mg / mL, respectively. Finally, multiple sets of different concentration solutions with a volume of 1 μL were tested, the errors of the concentration measurement results are all within 2 mg / mL, which indicates that the proposed method can effectively detect the concentration and can be used in the concentration detect of the volume determined micro-quantity liquid.

Abstract:The health indicator (HI) extracted by the traditional phase space warping (PSW) method cannot track the degradation trend of rolling bearings under the variable working conditions. To solve this issue, the improved phase space warping (IPSW) is proposed in this study. From the perspective of ensuring the independence of PSW components and maximizing their information, the highdimensional reconstruction of the phase space phase trajectory distortion and entanglement can be avoided by the integrated average mutual information local minimization and mutual information entropy local minimization between the high-dimensional reconstruction components. The slow-changing damage trend item and the working condition change interference item are decoupled to realize the HI extraction that is independent of the working condition change and reflects the fault evolution trend. Simulation and experimental signal verification results show that the extracted HI with the IPSW method can effectively avoid the influence of speed change and track the bearing damage degradation trend effectively.

Abstract:The on-wing monitoring data of aircraft auxiliary power unit (APU) are difficult to characterize its performance states, which will lead to the difficulty that the performance evaluation and remaining useful life (RUL) prediction of the APU is difficult to carry out. To solve this problem, a performance evaluation and RUL prediction approach of APU is proposed based on the state space model ( SSM) and Kalman filter (KF). Firstly, a performance indicator (PI) containing noise is constructed from the on-wing monitoring data to characterize the performance state of the APU. The performance degradation process of APU is described by state equation, which is constructed with the help of the Wiener process and the constructed PI with noise contained. Then, the KF state estimation and prediction method is applied to the SSM. Through estimating the on-wing performance state of APU, the purpose of predicting RUL is achieved. Finally, the APU on-wing monitoring data from the operation of an airline company in China are adopted to conduct the comprehensive verification and evaluation of the proposed method. Experiment results show that compared with ELM and Optimized ELM, the prediction absolute percentage error of the proposed method is reduced by 72. 1% and 67. 9% , respectively. In addition, compared with the experiment results of other three kinds of methods, the prediction absolute percentage error of the proposed method is reduced by 69. 2% at least. The proposed method can effectively predict the RUL of an on-wing APU, which can provide a reference for the operation and maintenance personnel to plan maintenance and repair reasonably. More importantly, the method can improve the comfort of the passengers and the aircraft safety to a certain degree.

Abstract:In the case of a small amount of labeled sample data, the deep model trained by the traditional deep learning-based gearbox fault diagnosis method has poor generalization ability, which is prone to over-fitting. To address this issue, a semi-supervised fault diagnosis method for the gearbox is proposed, which is based on the augmentation of labeled sample in deep embedding relation space. In this method, a small number of labeled vibration signals are input into the relation network in pairs for supervised training. Then, the labeled vibration signals are used as references, and a large number of unlabeled vibration signals are input into the trained relation network to establish the embedding relation space between labeled signals and unlabeled signals. In the relation space, some of the most similar signals are selected, and their predicted labels are set as pseudo labels and added to the labeled vibration signals. The above steps are iterated to expand the labeled samples to improve the generalization ability of the relation network. After the relation network is trained, it is used for mechanical fault diagnosis to realize fault diagnosis and classification. Experimental results show that the proposed method successfully expands the number of labeled sample when it is used to process the gear vibration signals with only a small number of labeled samples. The gearbox fault identification effect is better than the traditional supervised and semi-supervised fault diagnosis methods.

Abstract:The two currently existing fault diagnosis methods for rolling bearings based on signal processing technology and big data processing technology have the disadvantages of over-reliance on signal processing, complicated model and weak interpretability. Aiming at the shortcomings of traditional fault diagnosis technologies, this paper introduces the time series classification method based on shapelets learning algorithm into the field of fault diagnosis, and establishes the unbalanced data set of the faults of the EMU axle box bearing through the EMU wheelset bench rolling vibration experiment. The diagnosis model is improved based on the idea of Dropout. Experiment results show that the method guarantees the accuracy of fault diagnosis while retaining the strong interpretability of the shapelets as “the most representative time series subsequences”. At the same time, the improvement of the model based on Dropout improves the generalization performance of the model. The diagnostic accuracy of 100% on the training set and test set of bearing fault data was achieved, which proves that the improved learning algorithm based on shapelets is a feasible method applied to the fault diagnosis of axle box bearing of electric multiple unit.

Abstract:Aiming at the complex data distribution characteristics of industrial processes, this paper proposes a probabilistic principal component analysis fault detection method based on local neighborhood standardization ( LNSPPCA ) to solve the problem of unsatisfactory fault detection effect caused by multi-modal characteristics and uncertainty of process data. Firstly, LNS is used to solve the data multi-modal problem, so that the standardized data obey a single Gaussian distribution as much as possible. Then, the PPCA method is used to analyze the data from the perspective of probability, which can take into account the randomness of the data, so as to describe the data more realistically, extract more comprehensive and valuable information, and effectively detect faults in the complex data distribution process. Therefore, the LNSPPCA method can effectively improve the industrial process fault detection capability in multi-modal process complex data distribution. Numerical examples and TE process were used to conduct application experiments, and the test results are compared with those of principal component analysis (PCA) and PPCA methods, which verifies the effectiveness of the LNSPPCA method.

Abstract:When the electromagnetic ultrasonic transducer (EMAT) is applied to the castings and forgings with coarse grains for fast and online inspection, the signal-to-noise ratio ( SNR) and range resolution of the ultrasonic echoes are low due to the influence of harsh environments such as high temperature, large lift-off, and rough surface. To address the above problem, a finite element model for the testing process of a racetrack coil EMAT is formulated, which is based on the chirp signal excitation. The orthogonal test table is used to analyze the influence of EMAT design parameters, bandwidth and pulse width of the chirp signal on the peak and width of the main lobe width of the detected echo after pulse compression. In this way, the optimal combination of EMAT parameters referring to the peak and width of the main lobe are obtained, and they are evaluated by experiments. The detection capability of 0. 5 MHz tone burst signal as an excitation with different synchronization averages for Φ4 flat-bottomed hole is compared with that of chirp pulse compression with different lift-offs and no synchronization averages. Compared with the SNR of the tone burst excitation with 128 synchronization averages, results show that the SNR of the detected echo using the chirp pulse compression technology with no synchronous averaging from the Φ4 flat-bottomed hole is increased by 6. 6 dB. In addition, the SNR of the pulse compressed signal from the Φ4 flat-bottomed hole can reach 8. 0 dB with an EMAT′s lift-off of 3. 5 mm and no synchronization averages.

Abstract:In order to accurately determine the location of the defects in pipeline and reduce the complexity of echo signal, an electromagnetic acoustic transducer structure with unidirectional propagation in the pipeline is proposed. According to the theory of electromagnetic acoustic transducer, the mathematical models of the magnetic field, displacement and ultrasonic superposition were established, and the finite element simulation analysis of wave fluctuation displacement of unidirectional electromagnetic acoustic transducer was carried out. In the experiment, the double coil excitation method was used to control the propagation intensity of T(0,1) mode guided wave. The results show that when the distance between adjacent wires of the transducer is λ/ 4 and the delay time is T/ 4, the amplitude of echo signal on the weakened side is almost zero. The amplitude of the echo signal on the enhanced side is significantly enhanced, and the relative sensitivity to the intensity of the acoustic wave generated by the traditional transducer is 5. 88 dB. The unidirectional T(0,1) mode electromagnetic acoustic transducer realizes the control of propagation intensity and direction, improves the signal-to-noise ratio of detection signal, provides theoretical basis for accurate identification of defect orientation, and provides accurate defect location information for pipeline detection engineering application.

Abstract:Detecting the surface and subsurface micro weld defects is the key to ensure welding quality. A weld defect detection method with magneto-optical imaging based on deep convolutional network is proposed. On the basis of Faraday magneto-optic rotation effect, the principle of magneto optical imaging is analyzed. A deep convolutional network prediction model is established to study the influence of different model structure parameters on the training results. Through analyzing the intermediate mechanism of deep convolutional neural network, the model training process is studied and the optimal parameters of convolution kernel are found automatically. Experiment results show that the optimal prediction model can be achieved by selecting the size of the first layer convolution kernel as (7×7) and using the Relu activation function. The average training accuracy of magneto-optical imaging of weld defects is 98. 61% , and the prediction accuracies of 5 weld samples with pit, crack, incomplete penetration, incomplete fusion and non-defect are 84. 38% , 98. 05% , 84. 38% , 100% and 100% , respectively, and the average prediction accuracy is 93. 36% .

Abstract:In order to ensure the stable and reliable work of gastrointestinal (GI) micro-robots in human body, with the goal of expanding the working range of the wireless power transfer (WPT) system, improving the received power and its stability, studies a new type of combined solenoid WPT system. Through finite element simulation analysis, the best configuration and structure parameters of the power transmitting coil ( PTC) is determined. According to the minimum power transmitting requirement and the proposed performance evaluation index of the PTC, the number of turns of the PTC is optimized and determined. The size of the fabricated PTC is 50 cm× 50 cm×42 cm, which is 17 cm larger than that of the traditional Helmholtz coil in the axial direction, the PTC can cover the human GI region more completely. The experiment results show that using a three-dimensional orthogonal power receiving coil with a side length of 12 mm, the load received power exceeds 660 mW at any positiosn and orientations inside the PTC. In the non-edge area of the PTC, the position stability of the received power exceeds 80% . The system can meet the practical requirement of power supply for gastrointestinal micro robot.

Abstract:To apply the visual image analysis of speech spectrogram to human fatigue detection effectively, a human fatigue detection method based on speech spectrogram features is proposed. Firstly, the influence mechanism analysis of human fatigue on speech spectrogram is analyzed. The Mel frequency stretching transform of speech spectrogram based on the auditory perception theory is used to highlight the region of interest which is susceptible to fatigue. Secondly, the Mel frequency stretched spectrogram is divided into 24 overlapping critical frequency band sub-images, and 15 texture features are extracted from the gray level co-occurrence matrixes of each sub-image in 4 directions to quantitatively describe the fatigue information. Finally, a human fatigue detection model based on multi sub-bands fatigue information fusion is formulated by designing the feature-layer classifier for distribution detecting the features of each critical frequency band. In this way, the fatigue detection result can be achieved, which is based on the decision-level multi-classifiers fusion decision. Experimental results show that the extracted speech spectrogram features have stronger fatigue classification ability than traditional acoustic features. The fatigue detection effectiveness of this method is also better than the existing spectrogram feature recognition methods.

Abstract:Artificial anal sphincter (AAS) is an implantable medical device used to treat fecal incontinence ( FI). The existing AAS equipment has shortcomings in biological safety and intestinal content perception, which limits its clinical application. Firstly, a hyperelastic model of rectal tissue was established combined with the characteristics of intestinal dynamics, and ANSYS was used for finite element analysis to simulate the process of closing the intestine with the prosthesis. Then, based on the simulation results, a new connecting-rod artificial anal sphincter structure embedded with 10 sets of pressure sensors was proposed. Finally, in vitro experiments and live animal experiments were carried out. The experimental results confirm that the simulation model is established accurately, and the sphincter prosthesis can control 200 g feces and meet the normal life needs of human body within the blood supply safety pressure threshold. The 10 sets of pressure sensors are reasonably distributed, and the measurement results have good linearity. The accuracy of intestinal content quality prediction in vitro experiments is 89. 69% , and the accuracy of defecation warning in live animal experiments is 82% , which can preliminarily reconstruct the rectum perception of FI patients.

Abstract:This article aims to monitor and control the vertebral lamina′s milling depth of the orthopedic robot in real-time. First, based on the milling cutter′s geometric structure and the lamina′s forced vibration equation, the vertebral lamina milling process is modeled and analyzed. An acoustic signal-based monitoring method of milling depth is proposed. Then, a bone milling acoustic signal processing method based on the fast Fourier transform accurately extracts the harmonic amplitude whose frequency is an integer time of the surgery power device′s spindle rotation frequency when the frequency changes. The harmonic amplitude is used as a feedback signal. The robot′s planned axial feed rate is corrected based on the proportional-differential controller with a dead zone to control the milling depth. Finally, milling test experiments and milling depth control experiments are implemented on some artificial bone block based on a threedegree-of-freedom Cartesian robot system. Experimental results show that based on the milling acoustic signal′s harmonic amplitude, the milling depth is monitored with a resolution of 0. 1 mm within the safe measurement range of the milling depth of 0~ 1. 2 mm. Therefore, the method can effectively control the vertebral lamina′s milling depth during bone surface deformation and displacement. The proposed method can improve the accuracy and safety of the laminar milling operation of the orthopedic robot.

Abstract:Related studies show that about 8% of the global population suffer from chronic kidney disease (CKD). The prevalence is on the rise and the patients are getting younger. According to the principle of Lambert-Beer law, the urine albumin is detected by the immune-transmission turbidimetric method, aiming at the characteristic of antigen-antibody conjugate forming antigen-antibody conjugate precipitate from the liquid phase. The renal health of the patients is evaluated by measuring urinary creatinine and calculating the ratio of urinary albumin to urinary creatinine. The experimental device is set up to carry out clinical comparative experiments with HITACHI 7180 automatic biochemical analyzer. Experimental results show that the coefficient of determination for linear fitting of urinary albumin is r 2 = 0. 996, and the coefficient of repeatability variation is CV ≤4. 087% . Compared with the HITACHI 7180 automatic biochemical analyzer, the positive and negative albumin detection coincidence rates are 89. 36% and 92. 06% , respectively. The ratios of urinary albumin to urinary creatinine is 92. 96% positive and 87. 18% negative.

Abstract:The gesture recognition using electromyography ( EMG) has advantages of selective detail information and strong antiinterference ability. However, the adaptability and recognition accuracy of the existing methods are insufficient. By adding a long-term and short-term memory network layer on the basis of the convolutional neural network, a gesture recognition model is formulated. In this way, it can capture the EMG timing characteristics of the gesture, and the phenomenon of overfitting is reduced to a certain degree. The rich time-frequency domain information of EMG is utilized to extract the wavelet packet feature image of EMG. In addition, the input data of the recognition model are used with the EMG image to expand the category information of the EMG signal. Meanwhile, the attention mechanism is introduced between the time memory network processing layer and the convolutional neural network layer. Then, the model can indirectly increase the weights of the key gesture EMG channels. Compared with the method of ordinary convolutional neural network model using single EMG image, experimental results show that the recognition accuracy rate of the processing methods of EMG two feature inputs is improved by 4. 25% .

Abstract:The traditional rehabilitation robot mechanism has many kinematics solutions. In addition, the terminal control is complicated, and there is only audio-visual feedback and weak feedback. It fails to achieve the best rehabilitation effect. An interactive rehabilitation system based on the self-decoupling mechanism is developed. It contains parallel connecting rod mechanism to realize the decoupling of translational and rotational degrees of freedom and the decoupling of force and moment. Three control modes are designed based on the patient′s terminal position, including the human-led mode, robot-led mode and safe stop mode, which can realize high-precision trajectory tracking, dynamic adjustment of auxiliary force and safe stop function. The force feedback technology is combined with virtual reality and rehabilitation medical technology to realize multi-sensory interaction between the patient and the system. Experimental results show that the system can accurately track the rehabilitation trajectory in the 500 mm × 350 mm working space. The joint motion trajectory error is less than 5 mm, and the system can provide feedback force with an error less than 1 N. And it can stop in the danger zone. The effectiveness of the rehabilitation training system is evaluated.

Abstract:In the task of epilepsy detection, the selection of EEG channel directly affects the detection performance. To solve the problem of weak detection ability in some periods of detection method using static channels, a dynamic channel selection method is proposed. The channel set is determined according to the channel position and power spectral density (PSD) of EEG. The channel with the strongest epileptic detection ability is selected as the feature extraction channel, which can enhance the overall detection ability by improving the local detection ability of epilepsy. Experimental results show that the dynamic channel selection method can detect epilepsy with 98. 99% accuracy, 98. 52% sensitivity and 99. 52% specificity. Compared with multi-channel, the detection performance is similar. However, the feature extraction channel is the least, and the time complexity was reduced to O(1). Compared with single channel, the accuracy, sensitivity and specificity are improved more than 4. 93% .

Abstract:The cerebral blood flow autoregulation (CAR) is the brain blood supply regulation process of brain vessel under the control of nerve. The nerve and physiological states of the brain can be judged through evaluating the CAR. In this paper, using bioimpedance technique to evaluate CAR is studied. The brain blood flow circulating resistance parameter and artery blood vessel elasticity parameter are introduced in the CAR evaluation. In the experiment, the intervene method was used to make the brain blood supply system of the subjects in normal and abnormal states. The brain blood flow impedance signals were measured in the two states, and the resistance parameter and elasticity parameter were extracted. Comparative analysis shows that the change ratios of the resistance and elasticity parameters are 8. 53% and 20. 89% , respectively. Statistical analysis shows that these two parameters have significant difference (P< 0. 001) before and after the intervention. The experiment results verify that the brain blood flow bioimpedance technique can track the change process of CAR, and the proposed method has great significance for the further study of brain physiological activity and brain nerve function using bioimpedance detection.

Abstract:In view of the dense deployment of bluetooth mesh network nodes and the influence of metal materials on the wireless signal transmission of the bluetooth antenna, improves the gain of the bluetooth antenna through changing the shape of inverted-F antenna radiation arm, so as to achieve the purpose of using bluetooth mesh network node to transmit wireless signals in long-distance, while ensuring that the surrounding metal materials will not affect the transmission of bluetooth wireless signals, reducing signal energy loss and realizing the long-distance deployment of bluetooth mesh network nodes. These two bluetooth antennas have the characteristics of miniaturization, high gain and high bandwidth, and cover the entire ISM frequency band of bluetooth communication. The HFSS was used to simulate and analyze the characteristics of the bluetooth antennas, the optimized structural parameters of the bluetooth antennas were obtained. The real bluetooth antennas were manufactured, and the experiment tests were conducted. The experiment results show that the return losses of the folded antenna and spiral antenna are both less than - 10 dB, and the bandwidths completely cover the bluetooth frequency band; moreover, the performance of the spiral antenna is better, its gain reaches 4. 6 dB, and the farthest transmission distance in a metal environment is 20 m.

Abstract:To solve the drift problem in the electronic nose sensor array, a long-term drift suppression method with the augmented convolutional neural network is proposed. First, by combining historical data to expand the database, it has the effectiveness of data enhancement. Then, the incremental compensation module is used for network training to enhance the entire network performance. Finally, public dataset and measured dataset are utilized to evaluate the drift suppression performance, respectively. Compared with the traditional convolutional neural network and machine learning algorithms, experimental results show that the proposed augmented convolutional neural network ( ACNN) has great accuracy increase about 10% - 20% , and the accuracy fluctuation of 1% is good robustness, which verified that the augmented convolutional neural network is robust and effective in the suppression of electron nose drift, at the same time, also provides ideas for the drift suppression of electronic nose from the algorithm level.

Abstract:The discharge sound signal has been widely utilized to detect the partial discharge of the ring network cabinet, which has the characteristics of rich information and can accurately reflect the discharge fault. However, the effective detection of the discharge sound signal is a difficult point. The two-channel recording has unique advantages in the field of noise cancellation, which can effectively eliminate non-stationary noise. According to the characteristics of long-distance non-stationary noise measured on-site, the main noise interference has long-distance non-stationary feature. But, the partial discharge signal is a short-distance sound source. The noise elimination method of the discharge signal of the ring main unit is based on the dual-channel energy difference (power level difference, PLD). Compared with the single-channel denoising method based on spectral subtraction and Wiener filtering, the simulation experiment results show that the signal-to-noise ratio of the proposed denoising method under the two non-stationary noises is 14. 8 dB and 9. 1 dB higher on average, respectively. The square errors ( 1 × 10 -4 ) is reduced by 19. 34 and 15. 50 on average, respectively, and the denoising performance is better than that of the single-channel denoising algorithm. In the field experiment, the discharge sound waveform of the ring network cabinet is enhanced, which can effectively remove non-stationary noise of the surrounding environment. Thus, the effective discharge signal is retained, which can provide effective data support for the partial discharge diagnosis of the ring main unit.

Abstract:To improve the property of measuring accuracy and noise suppression of the current ultrasonic anemoscope, a measurement method of wind vector is proposed, which is based on the structure of ultrasonic receiving array with dual sensors. Firstly, a structure of wind speed and direction measurement system is designed by using the principle of ultrasonic wind measurement. It consists of an ultrasonic transmitting sensor and two ultrasonic receiving sensors. Then, an ultrasonic time delay estimation algorithm based on the correlation detection is proposed. The wind speed and direction can be obtained directly via the relation between ultrasonic transmission time and wind vector. Finally, the feasibility and effectiveness of the proposed method are evaluated by several simulation experiments. The actual measured data are also utilized through establishing a wind measurement system with dual-sensor ultrasonic array. Experimental results show that the proposed method has the advantages of simple structure, stable algorithm, low computational complexity and good noise suppression performance. The practical test experiments show that the maximum relative error of wind speed measurement is 2. 3% , and the maximum error of wind direction angle measurement is -1. 5°. The technical requirements of wind vector measuring instrument can be meet basically.

Abstract:Electrical impedance tomography ( EIT) is a kind of imaging technology to realize the image reconstruction of electric conductivity distribution in the practical field. Traditional electrical impedance imaging algorithms have the problem of low imaging accuracy. To address this issue, a new electrical impedance image reconstruction method based on the U 2 -Net deep learning model is proposed in this paper. First, based on the U 2 -Net model, this paper innovatively proposes the concept of concatenate (CAT) for data extension, which makes the input layer of U 2 -Net simple in structure and fast in operation speed. Secondly, the simulation data set is used to train the network, and the validation set is used to select the optimal model parameters. Experimental results show that the proposed algorithm has high measurement accuracy and good robustness. This method performs better than other algorithms in the simulation data set. Finally, a new EIT imaging quality evaluation index is proposed to evaluate the performance of the algorithm, which is named as center and area error (CAE). Experimental results show that the CAE of the proposed algorithm is 4. 975, which is more accurate for the prediction of the center and area of the target object. And the imaging effectiveness is better than other comparison algorithms.

Abstract:In non-linear dynamic batch processes, the measured variables have different serial correlations, and the cross correlation among the variables could be reflected at different sampling moments, however, traditional detection methods do not consider the correlation among the variables, the relationships among all variables are usually regarded as independent or correlative for feature extraction, and the features of fault information are not fully extracted, so the monitoring effect is bad. Therefore, a batch process fault detection algorithm based on the kernel dynamic latent variable-dynamically weighted support vector data description (KDLV-DWSVDD) of variable blocks is proposed. Firstly, the variables are divided into related and independent variable sub-blocks through obtaining mutual information (MI) values among the variables. Then, KDLV algorithm is used to divide the related variable sub-block into a dynamic part and a static part, the vector auto-regressive model is established to monitor the dynamic part and the neighborhood preserving embedding (NPE) algorithm is used to monitor the static part. In the independent variable sub-block, DWSVDD algorithm can be used to extract the dynamic information of independent variables. Finally, the monitoring statistics are established for fault detection through KDLV-DWSVDD algorithm. The average fault detection rate of the proposed algorithm in the penicillin fermentation simulation process reaches 90. 38% , which is nearly improved by 15% compared with that of the comparison algorithms. The actual semiconductor industry process also proves the feasibility and superiority of the proposed algorithm for the fault detection of batch processes.

Abstract:The near-field acoustic holography is an acoustic image method, which can be used to localize sound source effectively in technical practice. The equivalent source method can be applied to fix the ℓ2 norm regularized inverse problem and the ℓ1 norm regularized inverse problem for different numerical models of sound sources. To enhance the resolution quality of traditional ESM and narrow the sidelobe, we propose an improved fast iterative shrinkage / thresholding algorithm for sound source identification, which is based on the equivalent source method (mIFISTESM-v). This method is a combined application of the modified improved fast iterative shrinkage thresholding equivalent source method and functional beamforming. Its practicability is evaluated by the numerical simulation of the resolution quality analysis, and the experiments of the single sound source location and the coherent sound source location. Results show that the proposed method can achieve excellent sound source identification and location, and the main lobe area obtained by this method is smaller than 0. 002.

Abstract:The received signal of traditional four elements ultra-short baseline positioning system is affected by noise seriously. The location result may have phase ambiguity and low reliability. To address this issue, a virtual element location method is proposed, which is based on flat array beamforming. Firstly, the method divides the uniform square array into four sub-arrays symmetrically. By sub-array beamforming, virtual elements are projected to form a four-element cross array respectively. Then, the direction of arrival estimation method of four-element cross array is implemented to calculate the target direction. The achieved result that has a large deviation from the previous results is discarded. Otherwise, it is used as the reference direction of the next ping data to form a feedback loop and improve the positioning accuracy. The complexity of the algorithm is significantly lower than that of traditional eigen decomposition positioning algorithm. The simulation shows the positioning performance is obviously better than the traditional four-element positioning algorithm. Using the sea trial data to analyze the feasibility of the algorithm, it can be seen from the result that the algorithm can effectively improve the accuracy by increasing the SNR of the received signal.

Abstract:The filter wheel-based spectral imaging system is widely used in spectral imaging, its spatial resolution is high, however, its spectral resolution is low. Aiming at this problem, a high spectral resolution imaging based on filter wheel dual camera system is introduced in this paper, meanwhile, a multi-spectral calculation and reconstruction method based on interpolation compensation is designed to achieve high spectral resolution and high spatial resolution imaging of the system. Firstly, the filter wheel dual camera imaging system is used to acquire the multi-spectral images and RGB images, and then the discrete spectral response curves are obtained from the multi-spectral images. Finally, according to the mapping relationship between the RGB three channel data and the spectral high-dimensional data, and the theorem of conservation of energy, the interpolation compensation of the spectral response curve is performed and the high spectral resolution imaging is achieved. Experiment results show that the proposed method can efficiently achieve the imaging with a spectral resolution of 5 nm even higher while maintaining the spatial resolution. The root mean square error between the reconstruction result and the true value is 0. 017 1, the proposed method has high accuracy and robustness.

Abstract:The effective identification of micro cracks is of great significance to the early fault diagnosis of structures. The image segmentation method and other methods are difficult to achieve satisfied results in the detection of micro cracks with complex shapes and broken area. Therefore, transforms the problem of micro cracks identification into a series of dense and continuous central point prediction. A feature extractor is established by using the refined layered residual module, and the feature reuse attention module is also utilized to propose a micro cracks detection method. Firstly, the same rectangular bounding box is used to label the crack track densely and continuously. Secondly, the ablation experiments are implemented on the different refined hierarchical residual module to obtain the backbone network which is conducive to the feature extraction of micro cracks. Finally, six different feature reuse methods are compared by combining the attention module with feature reuse and backbone network. Experimental results show that the highest and average accuracy of the proposed method are 61. 0% and 54. 7% , respectively, which are 4. 9% and 6. 3% higher than the original model. The proposed method successfully identifies the micro cracks and their local broken areas, and suppresses background interference in practical application.