Abstract:Insulate gate bipolar transistor (IGBT) is extensively utilized in aerospace, weapon systems, modern industry, transportation, and power systems. Due to the complex environment, IGBTs are highly susceptible anomalies, leading to system failures and significant economic losses as well as casualties. Consequently, IGBT fault detection has garnered widespread attention and significant focus from researchers. However, systematic reviews of fault detection techniques for IGBTs are scarce, hindering practical engineers' deep understanding and knowledge of this field. Therefore, this review provides a systematic overview of research achievements in IGBT fault detection from a methodological perspective. Firstly, an overview of the basic structure, operation principles, and common failure mechanisms of IGBTs is presented. Secondly, IGBT fault detection technologies are categorized into three major classes based on detection methods, with a summary of the advantages and limitations of each class. Finally, considering the current technological advancements, a deep analysis of challenges and prospects of the field of IGBT fault detection is provided.

Abstract:In view of the low contrast and complex topological structure of pavement defect images, most of the currently proposed segmentation algorithms still have great shortcomings in capturing the receptive fields and extracting pavement defect features. Therefore, this article proposes an improved U-Net road defect image segmentation algorithm. First, the SN-Disout residual block is proposed in the classic U-Net convolution block to enhance the model’s robustness against overfitting. Secondly, a criss-cross module is introduced between the encoder and the decoder to enhance the model’s ability to capture features between different positions in the feature map and more accurately model the boundaries of defects. Finally, the spatial channel squeeze and excitation module is introduced in the decoder, which enables the network to focus more on important features while reducing the dependence on irrelevant or noisy features; position-aware multi-head attention is added to the neck of the model to further helps the model better understanding and utilizing the internal relationship of the input data, thereby improving the performance and performance capabilities of the model, and using the hybrid loss function Dice+BCE to replace a single loss function. The intersection ratio and F1 of this algorithm on the Crack500 image data set reached 60.13% and 75.22% respectively, both exceeding mainstream semantic segmentation networks such as U-Net, DeepLabV3+, PSPNet, TransU-Net, and UNet++. Experimental results show that this algorithm can effectively improve the prediction accuracy of the network and the segmentation results of small targets, and it also meets the real-time requirements while ensuring segmentation accuracy.

Abstract:Wind turbine blades, being the core component of wind power generation systems, have their health status directly impacting the overall power generation efficiency and operational safety. Addressing the challenges of blade defect detection, researchers delved into the YOLOv8n network and innovatively proposed the Efficient Multi-Scale Convolutional module (EMSConv). This module effectively replaces the convolutional layers in traditional residual blocks, significantly reducing the interference of redundant features on detection results through grouped convolution techniques, thereby enhancing detection accuracy. Furthermore, in the detection head, a diverse set of attention mechanisms from Dynamic Head are incorporated. These self-attention mechanisms work in concert, spanning across different feature layers, to achieve precise perception of target scales, spatial locations, and detection tasks, vastly strengthening the comprehensive capabilities of the target detection module. Moreover, researchers innovatively integrated Inner-IoU, Wise-IoU, and MPDIoU, creatively proposing Inner-Wise-MPDIoU to replace the traditional CIoU loss function. This move not only improved the network’s detection precision but also accelerated the convergence process. During testing on a self-made dataset of wind turbine blade defects, YOLOv8-EDI exhibited remarkable performance, achieving an mAP50 value of 81.0%, a 2.3% increase compared to the original YOLOv8n. The recall rate also reached 76.8%, marking a 3.7% improvement. While enhancing detection performance, this model managed to reduce computational requirements by 5.5%, fully meeting the need for efficient, accurate, and large-scale blade defect detection in industrial settings.

Abstract:The power generation efficiency and service life of wind turbines are related to their surface integrity. This study aims to address the issue of inaccurate detection results and long detection time in traditional surface defect detection methods for wind turbines. A surface defect detection model for wind turbines is designed. Firstly, lightweight convolution technology is integrated into the model, effectively enhancing the ability of information exchange between channels and improving the detection effect of small-sized defects through richer feature information. Secondly, the visual attention network module has been introduced into the backbone network, enriching the contextual information and improving the feature extraction ability of convolutional neural networks. Then, a coordinated attention mechanism is introduced into the neck network to capture the location information of defects through spatial orientation. Finally, modify the loss function of bounding box regression to WIoU to develop an appropriate gradient gain allocation strategy. The experimental results show that the improved detection model improves the detection accuracy by 4.14% compared to the original model, significantly enhancing the detection ability of small defects. At the same time, the parameter count of the improved model was reduced by 2.29 M, while the parameter count was reduced by 6.2 G. The detection speed was significantly improved, meeting the real-time detection requirements of the model.

Abstract:To address the challenges in detecting dispersed and fine defects on metal surfaces, convolutional neural network (CNN) often fall short due to their limited ability to capture global features, leading to missed detections and loss of detail in identifying defects such as oxidation particles, cracks, and scratches. Although Transformers can capture comprehensive global information, the extensive computation required can be costly. In pursuit of an efficient and accurate method for metal surface defect detection, this study introduces a novel network architecture, the DPG-Transformer, which synergistically combines the local feature extraction capabilities of CNNs with the global modeling strengths of Transformer. This integration is facilitated through the use of depthwise separable convolutions (DW-Conv) and pooling grid window attention mechanisms (PGW-Attention). The effectiveness of the DPG-Transformer was validated on both a proprietary metal defect dataset (ST-DET) and a public dataset (NEU-CLS), achieving defect detection accuracies of 99.3% and 99.6%, respectively, and outperforming several classic networks in terms of accuracy, computational efficiency, and floating-point operations. Additionally, visualization experiments demonstrated that the DPG-Transformer more comprehensively extracts defect features associated with corrosion and scaling compared to CNN models, and more precisely focuses on the global features of elongated cracks and scratches than Transformer models. The results indicate that the DPG-Transformer not only reduces computational load and complexity but also enhances the comprehensive and precise detection of metal surface defects, making it a highly suitable approach for practical applications in metal surface defect detection.

Abstract:The classification of wafer defect patterns plays a crucial role in the wafer manufacturing process. Accurate identification of wafer defects enables the determination of the root causes of defects, thereby pinpointing issues in the production process.However, existing deep learning-based wafer defect classification methods are designed solely from the spatial or frequency domain, failing to achieve mutual supplementation and integration of spatial and frequency information. This limitation constrains the improvement of wafer defect classification accuracy. To address this issue, a dual-stream wafer defect classification network based on the fusion of spatial and frequency domain features, named SFWD-Net, is proposed.The network utilizes the proposed multi-scale feature extraction convolution module and multi-view attention module to form the spatial stream branch, which extracts spatial information from wafer images. The frequency stream branch, utilizing discrete wavelet transform, extracts frequency information from wafer images. After integrating spatial and frequency information, defect classification is performed. Experiments on the large-scale semiconductor wafer image dataset WM-811K demonstrate that SFWD-Net, by simultaneously designing the network from both spatial and frequency domains, achieves a classification accuracy of 99.299 2%, outperforming five other state-of-the-art methods and significantly improving the accuracy of wafer defect classification.

Abstract:In order to solve the problem that the defect echo overlaps with the surface echo when measuring the defect depth near the surface of metal components under sub-Nyquist sampling condition which leads to inaccurate depth measurement, a high-precision measurement method for the near-surface defect depth of metal components based on ultrasonic echo resampling and time-frequency transformation is proposed. The ultrasonic echo signal is decomposed into a series of narrowband mode components by variational mode decomposition, and the useful mode components are clustered based on the correlation coefficient to obtain the preprocessed signal. The preprocessed signal is then processed by cubic spline interpolation to realize the data expansion of the sub-Nyquist sampling echo signal and obtain the reconstructed signal. Based on the synchro extracting transform, the timefrequency spectrum and amplitude distribution curve of the reconstructed signal are obtained. According to the principle of ultrasonic single transceiver, the high-precision depth measurement of the near-surface defect of metal components is finally realized. The experimental results show that the proposed method can accurately measure the near-surface defect depth of the metal component under the condition of sub-Nyquist sampling condition, and the measurement relative error is 2.161%, which is obviously less than that of the envelope method and the time-frequency transform method, whose relative error are 3.570% and 13.182%, respectively. Furthermore, the effects of the sub-Nyquist sampling rate, the signal noise and the defect location on the measurement accuracy of the near-surface defect depth are investigated. The results consistently demonstrate that the proposed method has better accuracy and robustness than the traditional methods for the near-surface defect depth measurement under different sampling rates, different noise and different defect depths.

Abstract:In order to solve the problems of poor accuracy of the charge inversion method for thunderstorm clouds and the influence of the existing charge inversion model by the environmental error caused by multi-station network observation, a nonlinear equation system is derived to establish a charge inversion model based on the three-dimensional atmospheric electric field. On the basis of the assumption of equal thickness model of thunderstorm clouds, a set of nonlinear equations required for the charge inversion of thunderstorm clouds is derived, and a three-dimensional atmospheric electric field-based charge inversion model of thunderstorm clouds is established. The population initialisation of the sparrow search algorithm (SSA) is optimised by the sinusoidal chaotic mapping function to improve the nonlinearity of the distribution of the populations, and the Levy function and the inverse learning strategy are used to optimise the position updating of the algorithm’s discoverer way, a thunderstorm cloud charge inversion method based on improved sparrow search algorithm (ISSA) is proposed. A 3D atmospheric electric field instrument is used to observe the ground electric field data and analyse the electric field characteristics, and the improved SSA algorithm using the hybrid strategy is used to invert the thunderstorm cloud charging model parameters. The experimental results show that the inversion of the data obtained from the three-dimensional atmospheric electric field instrument (3DAEF) can effectively eliminate the errors caused by the multi-station network observation. Compared with the SSA, the deviation rate of the two-second neighboring charges of the thunderstorm cloud obtained by the improved ISSA algorithm is around 1%, and the fitness value reaches as low as 5.38, which is able to accurately invert the charging parameters of the thunderstorm cloud, and provide a certain reference to the study of its charging and discharging process.

Abstract:To consider increased difficulties of the GNSS signal acquisition due to a wider range of the frequency bandwidth in high dynamic environment, the transmission characteristics of the numeric intermediate frequency signal in the GNSS receiver was analyzed, as well as the correlation peak detection of the complex baseband signal processed by the FFT module, a fast detection method for high dynamic GNSS signal acquisition based on ML estimation is proposed. Firstly, by the construction of Binary hypothesis testing conditions based on the statistical theory of random signals, an acquisition threshold model was presented using the Neyman Pearson Criterion; Secondly, the variance of equivalent White Gaussian Noise based on ML is estimated by judgment statistical characteristics and an acquisition threshold is calculated from variance estimated value，in the meantime the estimation error caused by increased judgment samples was resolved by means of false alarm rate quantized amplification. Finally, the acquisition detection simulation experiment of Beidou B3I signal was conducted in different high dynamic conditions. The result showed the proposed method have a wider scope of high dynamic adaptive capacity, and the precision accuracy of the acquired doppler frequency shift was on a par with the SINS information aiding method, also increased by more than 28% compared to the sequential detection method, as well as a faster detection speed in the same acquisition condition.

Abstract:Partial discharge of electrical equipment is not only the main factor of insulation deterioration, but also an important parameter to effectively characterize insulation defects. The accurate detection of partial discharge can detect the potential faults that endanger the safety of equipment in time. UHF detection has the advantages of good real-time performance and strong anti-interference, and is widely used in discharge detection. However, the existing microstrip antenna sensor is limited by the structure size, and the working bandwidth is difficult to increase. In this paper, a new flexible microstrip antenna sensor based on polyimide is developed. The partial flooring technology combined with the beveled meandering technology is used to improve the structure. The nonlinear relationship between antenna size and working bandwidth is considered to optimize the size. The working bandwidth is expanded while keeping the antenna area unchanged. In order to solve the problem that the antenna performance is unstable due to the adjustment of single size parameters in the process of size optimization, a relationship model between multi-size and working bandwidth is proposed by using radial basis function (RBF) neural network, and an improved beluga whale optimization (IBWO) algorithm is used to optimize antenna size. The simulation results show that the size of the new flexible microstrip antenna is reduced by 59.59%. The operating bandwidth is increased from 0.598~0.6 GHz to 0.3~3 GHz, which fully covers the design requirements of partial discharge detection. By simulating partial discharge test and comparing with Archimedean spiral antenna and three-dimensional spiral antenna, the results show that the new flexible microstrip antenna has more reliable detection performance.

Abstract:To address the issue of weak out-of-band suppression in traditional artificial surface plasmon (SSPPs) microwave filters, a miniaturized SSPPs wide-stopband microstrip lowpass filter featuring a novel butterfly structure is devised and fabricated. Firstly, by integrating the fan structure in the conventional microstrip filter with the SSPPs theory, the butterfly structure of SSPPs is designed to obtain a broader stopband. To enhance the out-of-band suppression level of the filter, an anti-directional arrowhead resonant patch is introduced to suppress the parasitic passband at 25 GHz within the stopband. The optimal design of the SSPPs unit is accomplished by simulating the dispersion characteristics of the unit structure. A complete SSPPs filter structure is designed, and the impact of key parameters on the performance of the filter is simulated. The metal-medium-metal (MIM) structure of the filter is fabricated using F4B substrate, and its performance is measured. The results indicate that the 3-dB cut-off frequency of the filter is 10.08 GHz,with the highest stopband frequency reaching 33 GHz. Return loss within the passband consistently exceeds 15 dB, while maximum insertion loss within this range is measured at 2.6 dB.Notably, minimum attenuation within the stopband measured at 22.98 dB between 11~35 GHz, with return loss generally exceeding 3 dB across the range of 10~30 GHz and peaking at 4 dB from 10~34 GHz. The test results of this design meet the design index, the out-of-band suppression effect is ideal, and it can be integrated with mixers, signal generators, and other devices to exert a superior circuit performance.

Abstract:In the context of time-varying operating conditions, fault diagnosis often exhibits high dynamism, while the limited model learning under small samples makes the issue more challenging. For the above situation, a fault diagnosis method based on contrastive deep convolutional networks is proposed. Firstly, considering the characteristic of small data samples, take advantage of differences in vibration data distribution caused by speed changes, and naturally realize data enhancement without manual operation. Subsequently, in the process of data processing, the vibration data of the same healthy state at different rotational speeds are used as positive samples, while the vibration data from different health states are used as negative samples. The key features are extracted by comparing the similarity between the samples so as to reduce the distance between the positive samples while increasing the distance between the negative samples. Finally, the feature extractor is trained and optimized by comparative training method, where a weighted combination of contrastive loss and cross-entropy loss is used as the composite loss function, enabling the model to effectively perform classification tasks while learning feature representations. The method is applied to two different bearing failure datasets at different time-varying rotational speeds for case studies respectively. The experimental results show that the proposed model not only performs well in the feature extraction and classification tasks, but also realizes high accuracy fault diagnosis under both data scarcity and time-varying speed conditions. It is verified that the proposed model shows high feasibility and effectiveness in dealing with time-varying small-sample data, and outperforms other advanced diagnostic methods.

Abstract:The microseismic signals generated during the deformation and rupture of loaded composite coal rock contain information about the rupture of the internal structure of the coal rock, and the microseismic signals collected by traditional equipment cannot be analyzed directly because of the presence of a large amount of environmental noise. In order to effectively extract the change characteristics of the microseismic signals during the deformation and rupture of loaded coal rock, a new CEEMD-IDWT joint denoising algorithm is proposed by integrating the complementary ensemble empirical modal decomposition algorithm (CEEMD) with the improved dmey wavelet (IDWT) algorithm. The algorithm firstly utilizes the CEEMD algorithm to decompose the original signal, then applies the IDWT algorithm to denoise the IMF components obtained from the decomposition, and finally reconstructs the processed components to obtain the denoised signal. The effectiveness of the algorithm is verified using simulation analysis and uniaxial compression experiments, and the results show that: the CEEMD-IDWT joint algorithm improves the signal-to-noise ratio by a maximum of 204.5% compared with the traditional algorithm in simulation analysis, and increases the signal-to-noise ratio of other improved denoising algorithms by a minimum of 11.8%, which is an obvious advantage in denoising ability; the microseismic voltage obtained by embedding the algorithm into the self-researched microseismic voltage acquisition equipment is significantly higher than that obtained by the conventional algorithm in the uniaxial compression experiments on the composite coal rock. The noise-to-noise ratio of the microseismic voltage signal obtained in the compression experiment is only 0.089 75, and the actual denoising effect is obvious; the microseismic voltage after denoising by the joint CEEMD-IDWT algorithm has obvious change characteristics, which significantly improves the signal denoising effect and effectively avoids the distortion of the microseismic voltage signal, and can be used as an ideal algorithm for the denoising of deformation and rupture of the microseismic voltage signal of the loaded coal rock and provides an ideal algorithm to accurately predict the coal rock dynamics and disaster. It can be used as an ideal algorithm for de-noising the microseismic voltage signal of loaded coal rock deformation and rupture, which provides a reliable and advanced technical reference for the accurate prediction of coal-rock power disasters.

Abstract:During the long-term use of lithium battery packs, there is a problem of inconsistent voltages among the series-connected individual batteries. To solve this problem, a cascade bidirectional Cuk equalization circuit system based on variable domain fuzzy PID control is proposed in this paper. In this system, a cascade bidirectional Cuk equalization circuit is employed to achieve equalization between non-adjacent cells, and the variable domain fuzzy PID algorithm is exploited to enhance the voltage equalization speed. To verify the feasibility and superiority of the system, simulation models of the traditional bidirectional Cuk equalization circuit and the cascaded bidirectional Cuk equalization circuit were designed in MATLAB/Simulink. The comparison of simulation results under the control of the fuzzy PID algorithm and the variable domain fuzzy PID algorithm reveals that the balancing time of the cascade bidirectional Cuk balancing topology based on the variable domain fuzzy PID algorithm is decreased by 64.29% compared with the traditional bidirectional Cuk balancing topology without a control algorithm. The equalization time of the cascade bidirectional Cuk equalization topology without a control algorithm is reduced by 50.82%. The balancing time of the cascade bidirectional Cuk balancing topology based on the fuzzy PID algorithm is reduced by 14.29%. According to the analysis of the simulation results, it is demonstrated that the cascade bidirectional Cuk equalization circuit system based on fuzzy PID control in the variable theory domain can augment the voltage equalization rate of the battery.

Abstract:Coal and gangue have the characteristics of dense targets and small feature differences, and the recognition methods based on image processing generally have the problems of slow detection speed and low accuracy. To further improve the speed and accuracy of coal gangue detection, a GE-YOLOv5s coal gangue detection model is proposed. Firstly, Ghost Conv is introduced based on YOLOv5s instead of convolution operation, and a new module GhostCSP is designed to improve the detection speed of the model while realizing the lightweight of the network; secondly, the GC self-attention mechanism is added in the prediction layer, which integrates the lightweight of SENet and the advantage of global capture of long-distance information of NLNet, to enable the network to memorize and magnify the Then in the Neck part, a bidirectional feature pyramid network (BiFPN) structure is adopted, and BiFPN is used to fuse the features of three different dimensions to improve the computational efficiency of the model through the weighted feature fusion mechanism to further enhance the speed of coal gangue detection; finally, a new type of activation function is designed to replace the activation function of SiLU, which can improve the utilization rate and accelerate the convergence of the model. Finally, a new activation function Eswish is designed to replace the SiLU activation function, which improves the parameter utilization rate, accelerates the convergence speed of the model and improves the robustness. The experimental data show that compared with the YOLOv5s model, the number of parameters is reduced by 34.1% the amount of floating-point operations is reduced by 38.6%, and the mAP 0.5:0.95 index is improved by 1.9%. Comparison experiments show that the mAP 0.5:0.95 metric is improved by 16.6%, 4.8%, 13.6% and 3.8% compared to YOLOv3, SSD, FasterR-CNN, and YOLOv5-scSE, respectively. Applying the GE-YOLOv5s model to the gangue target detection process has better recognition performance, robustness, and network generalization ability, and can effectively avoid the phenomena of leakage, misdetection and overlapping.

Abstract:In order to meet the high-precision measurement requirements of titanium alloy nanofilm pressure sensors, based on the influences of thin film thickness, convex islands, resistance shape and arrangement position to the nonlinearity and sensitivity of sensors, two range pressure sensors are designed and optimized using titanium alloy thin films as sensitive elements. The results indicate that whether there are convex islands or not, the maximum stress of the sensor occurs at the edge of the diaphragm. The maximum stress decreases and the position remains basically unchanged after adding convex islands. Adding convex islands or as the diameter of the convex island increases, the nonlinearity decreases and the sensitivity increases accordingly. To ensure that the sensitivity is not less than 2.5 mV/V in theory, 2 MPa sensor increasing convex islands with Φ2 mm, the nonlinearity decreases to 0.05% and the sensitivity is about 2.67 mV/V. 4 MPa sensor increasing convex islands with Φ3 mm, the nonlinearity decreases to 0.02% and the sensitivity is about 2.89 mV/V. Two ranges of sensors are prepared and tested, the results show that at a range of 2 MPa, the maximum deviations of the sensitivity and nonlinearity between actual values and theoretical values are 0.01 mV/V and 0.01%, respectively. At a range of 4 MPa, the maximum deviations of the sensitivity and nonlinearity between actual values and theoretical values are 0.16 mV/V and 0.02%, respectively. This study provides an important basis for the design of titanium alloy thin film pressure sensors.

Abstract:Power transformers are critical components of the electrical grid, and the insulation status of their windings is directly linked to the operational safety and reliability of power supply. Short circuits at the low-voltage side outlets can easily damage the inter-turn insulation, leading to inter-turn short circuit faults in the windings. To further study the transient processes of inter-turn short circuits and low-voltage side outlet short circuits in transformer windings, this research investigates the weak links in heat generation and mechanical stresses, and elucidates the mechanisms of insulation failure development. An electromagnetic-force coupled model, congruent with the actual structural dimensions of the transformer, was developed. Utilizing finite element simulation software, the electromagnetic transient processes of the windings under various operational conditions were examined. A comparative analysis of the electromagnetic force distribution characteristics was conducted to explore the influence of different fault conditions on winding insulation. The results indicate that during inter-turn short circuit faults in transformer windings, the current in the short-circuited turn significantly exceeds that in three-phase outlet short circuits. Compared to normal load conditions, the peak currents through the short-circuited turns and overall windings increased by 5 318% and 3 314%, respectively; the maximum magnetic field intensity rose by 1 511% and 2 111%, and the peak electromagnetic force density on the winding turns surged by 5 210% and 11 489%. Inter-turn short circuit faults in transformer windings are likely to severely damage the insulation, whereas three-phase outlet short circuits can lead to unstable deformations and degradation of insulation.

Abstract:Aiming at the problem of poor effect of X-ray and ultrasonic technologies in non-destructive detection of glass fiber reinforced plastic (GFRP), characteristics of high resolution of Xray images and high contrast of ultrasound images are used for complementary imaging fusion, and by integrating the detail information of the defect edge of X-ray images and the high contrast outline information of ultrasonic images, new images are formed to improve the defect display effect. The frequency domain algorithm based on cross sector filter is used to remove the horizontal and vertical fringe noise of X-ray images, the morphological filtering algorithm is used to remove the salt and pepper noise of ultrasonic images, and the image fusion algorithm based on region segmentation and static wavelet transform is proposed to fuse X-ray and ultrasonic image traits. The test results show that the standard deviation SD of the fused images is increased by 154.1% on average, the entropy H is decreased by 92.2% on average, and the contrast of defect detection images is higher and the edge details are clear. The algorithm can effectively remove the fringe noise and pepper and salt noise in the two kinds of images, and can effectively improve the weakness of low contrast of X-ray images and poor resolution of ultrasonic images, and provide a new idea for the defect damage detection of composite materials.

Abstract:The two-dimensional angular displacement table often consists of two twodimensional angular displacement table. It is difficult to avoid installation errors when mounting a biaxial tilt sensor on a two-dimensional angular displacement table, which can lead to inaccuracies in the measurement of two-dimensional tilt angles. Meanwhile, there is a two-dimensional angular coupling between the rotating axes of the off-axis angular displacement table. And the tilt of the bottom angular displacement table will cause the angle measurement deviation of the upper angular displacement table. Based on the existing angle accuracy of the angular displacement table and the tilt sensor, we hope that a method can be proposed to correct the above linear and nonlinear angle errors through the system model establishment and error analysis. Our work consists of the following parts. First, the systematic angle errors of the two-dimensional off-axis tilting system are analyzed and modeled in this paper, which are categorized into quasi-linear errors and nonlinear errors. The quasi-linear errors include vertical and horizontal tilt sensor installation errors. The nonlinear error is only caused by the off-axis layout of the angular displacement table. Second, a correction method combining the linear homography matrix and the nonlinear off-axis inclination model is proposed for compensating the two types of errors. Furthermore, the automatic correction process is designed. Finally, the calibration accuracy of the proposed method was verified by experiments. In order to verify the calibration accuracy of the proposed method, the off-axis angle control experiment was carried out in the range of ±12° to verify the calibration accuracy of the proposed method. The experimental results proved that the measurement accuracy of the two-dimensional angular displacement table tilt angle is improved from 0.559° and -0.216° before correction to 0.025° and 0.013°, which reduces the error by an order of magnitude. The method can meet the requirements of precise measurement and control of off-axis tilt angles. Furthermore, in order to verify the effectiveness of the proposed method, a comparative experiment was conducted with two existing methods. The experimental results show that the proposed method is superior to the other two methods in terms of accuracy and ease of operation. The errors of the biaxial tilt sensor mounted on a two-dimensional angular displacement table are analyzed and modeled, which are classified into quasi-linear errors and nonlinear errors. The corresponding correction method has been established by combining linear homography matrix and nonlinear off-axis inclination model. The experimental results proved that the proposed method effectively improve the measurement accuracy of the two-dimensional tilt angles. Compared with two existing methods, the accuracy of proposed method is higher, while the operation is much simpler.

Abstract:To address the issue of output saturation caused by higher-order term constraints, we propose a novel unsaturated tri-stable second-order stochastic resonance (UTSOSR) system that leverages the excellent anti-saturation properties of piecewise potential function. First, simulation experiments verified that this system can significantly mitigate the output saturation problem of classical tri-stable second-order stochastic resonance system. Next, based on the adiabatic approximation theory, we derived the steady-state probability density, mean first-passage time, and spectral amplification factor (SA) of the UTSOSR system. By analyzing the influence of various system parameters on these performance metrics, we can further explore the system’s dynamic behavior in greater depth. Subsequently, using the SA and the signal-to-noise ratio gain (Gsnr) as evaluation metrics, numerical simulations were conducted to verify the superior signal enhancement and noise robustness performance of the UTSOSR system. Additionally, to achieve superior output performance, we combined maximum correlated kurtosis deconvolution (MCKD) with the UTSOSR system, proposing the MCKD-UTSOSR method for extracting target signal features. Finally, a combined approach using genetic algorithm and variable step-size grid optimization algorithm is employed to identify the optimal parameters for the MCKD-UTSOSR method in bearing fault diagnosis. The data analysis results indicate that compared to other methods, the MCKD-UTSOSR method improved the signal-to-noise ratio by 1.128 9~23.585 4 dB and the spectral peak value by 88.423~7 488.118 133. This provides an innovative and reliable solution for efficient signal processing and fault detection in practical engineering applications.

Abstract:In the crucible of public safety, the imperative to guard against the scourge of fire is non-negotiable, yet conventional detection methodologies often falter when confronted with the complexities of specific environments. Herein lies the promise of computer vision technology, which offers the capability to monitor expansive territories in real-time and to identify the telltale signs of impending fires, most notably smoke. However, the intricate morphologies, textural variations, and chromatic subtleties of smoke present a significant challenge to the precision of its detection through machine vision.Addressing this exigency, we have conceived and developed an innovative smoke classification algorithm, seamlessly integrating a lightweight neural network and the convolutional block attention module (CBAM) within the YOLOv8 framework. This approach is designed to augment the accuracy and efficiency of smoke classification. Our algorithm leverages the GhostNet architecture, ingeniously replacing standard convolutional layers with a more efficient alternative, thereby maintaining high performance while drastically reducing the computational load on the model.Furthermore, the integration of CBAM imbues the algorithm with the ability to dynamically adjust its focus across different regions of the image, ensuring that salient smoke features are prioritized for detailed analysis. This feature enhances the model’s robustness and adaptability to diverse scenarios.To validate the efficacy of our algorithm, we conducted extensive experiments using both a publicly available smoke dataset and a custom dataset augmented with challenging samples. Empirical results have demonstrated that our algorithm achieves a smoke recognition accuracy of 99.9% on the public dataset and 99.2% on the custom dataset, outperforming existing methods. On our experimental machine, the algorithm exhibited a frame rate of 833 fps under GPU-accelerated conditions and 28 fps under CPU-only operation, affirming its potential for rapid and accurate early fire detection.

Abstract:Statically deployed marine ranching WSN is prone to the energy hole in the region close to the sink node during multi-hop data transmission. A reliable communication path planning method for the mobile sink node based on improved ant colony algorithm is proposed. First, for the traditional ant colony algorithm, non-uniform distribution of initial pheromone concentration is used to solve the problem of blind search at the initial stage of the algorithm. The heuristic function value in the transition probability function is modified and the crowding influence factor is added to avoid deadlock and speed up the convergence. In order to ensure the convergence ability of the algorithm at a later stage, an improved update rule for the pheromone is used. Secondly, LEACH protocol is used to cluster the network. According to the location of the cluster head node and the communication coverage area, the ergodic point set of the mobile sink node is constructed. Finally,the path planning problem of the mobile sink node is regarded as the traveling salesman problem, and the optimal path of the mobile sink node is obtained by the improving ant colony algorithm and the constructing backbone node set. The simulation results show that, at the scale of 275 network nodes, compared with other algorithms, the path length of PMRM is reduced by 41.9%, 20.3% and 30.4%, the data transmission delay is reduced by 42%, 38.5% and 46.7%, and the network throughput is increased by 10%, 10.6% and 16.4%. The superiority of the method is verified. The proposed method can effectively optimize the reliability of marine ranching WSN data reception, the energy consumption characteristics of the network and the life cycle of the network by introducing the mobile sink node and exploiting the rationality of their mobile paths.

Abstract:To meet the requirements of the multi-hole balance orifice plate structure parameters design, a general idea for the design and calculation of structural parameters was proposed, and take the plate containing a central hole and one circle of function holes as an example to give the specific design method and calculation example. Firstly, determine the thickness and chamfer angle of the orifice plate. Secondly, according to the relationship between the permanent pressure loss ratio and the equivalent hole opening diameter ratio (Δω/ΔP-β), and the relationship between the pressure loss coefficient and the equivalent hole opening diameter ratio (ζ-β), the β value is determined by iterative calculation. Then, determine the diameter of the hole opening circle Db and the hole number N, according to the pipe diameter. After that, according to the equivalent hole opening diameter ratio definition and supplementary geometric relationship, establish an equation group about the central hole diameter d0 and the equilibrium hole diameter db, then solving the equation group to obtain values of them. Finally, calculate the outflow coefficient. The general idea, process of the method and a design calculation example are given in the paper. In this paper, the structure parameter calculation process of a DN250 balance orifice plate is taken as an example to show this method. The results show that deviations between the calculated and the calibrated data of the discharge coefficient and full range differential pressure are within 5% and 10%, respectively, which meet the needs of engineering design. The example proves the practicability of this method.

Abstract:Crack detection is of great practical significance for steel structures and concrete buildings, nuclear power, aviation and other fields. In this paper, a resonant coaxial probe is used to construct a microwave microscope to carry out the study of non-destructive testing methods for metal and non-metallic cracks. First of all, the influence of key parameters such as coupling gap, lifting height and probe material on the detection performance is studied through electromagnetic simulation. Then, the experimental verification was carried out with machined brass, aluminum alloy samples and 3D printed non-metallic samples. The results show that the S11 amplitude of the coaxial probe at the resonant frequency can be used for non-destructive detection of cracks on the surface of the material. The coupling gap and lifting height are the main design parameters that affect the detection sensitivity. Compared with the traditional reflective coaxial probe, the detection sensitivity of the resonant probe (S11 at the observation frequency point when there is a crack or not) The amplitude difference) has been increased by about 20~30 times; the simulation of the 3.9 mm crack in the width of the metal sample is in good agreement with the measured results. The relative error of the crack width detection is < 7.7%, and for 3D printed non-metallic crack samples with a crack width of 1.7 mm, the measured relative error of the crack width is Within 6%. This research has certain reference significance for the non-destructive detection of metal and non-metallic cracks.

Abstract:It is needed to achieve very high multipath delay accuracy to better approximate the real communication scenario in the process of channel simulation, which puts higher requirements on the simulation ability of the channel emulator. Channel models is processed and loaded usually within the digital baseband, and so the clock resolution is limited. It is necessary to make use of Farrow structured fractional filters to achieve higher delay accuracy. In according to the characteristics of channel simulation algorithms, the Farrow structure fractional filter was designed and optimized by mixing DSP and distributed multiplication to achieve ultra-high precision delay simulation. The design scheme was validated and tested on the TRANSCOM Pathrot X80 channel emulator. The results show that,the improved Farrow filter′s structural design significantly reduces the consumption of FPGA computing resources, enabling the fractional delay algorithm to achieve a balance between high delay accuracy and low resource overhead; the multipath delay accuracy is consistent with theoretical calculations in the low frequency range, meeting the expected 0.1 ns requirement for channel simulation; the delay accuracy differs significantly from theoretical calculations in the high-frequency range. In order to achieve better performance or reduce signal distortion, it can be considered to increase the order of the filter or find a better coefficient calculation algorithm.

Abstract:The vector network analyzer is an essential instrument in RF and microwave testing. The traditional methods of detection necessitate the presence of onsite personnel for operation and control, which is inefficient and inflexible, failing to meet the contemporary requirements for intelligent and convenient electronic measurement. To address these challenges, an online software system based on a cloud server was designed for vector network analyzer. The system connects the vector network analyzer to a local computer via the SCPI protocol. The local computer uploads the Sparameters, comprising frequency points, real and imaginary components, to the cloud server’s MySQL database via SSH. The Djangodeveloped website allows users to access realtime visual data of Sparameters, including frequency domain, time domain, and standing wave ratio graphs, from any device. The system was tested on the E5071C vector network analyzer for dualport Sparameter detection of rectangular stainless steel tubes from 10~12 GHz. Results showed no data loss during multiple transmissions between the instrument and the local computer. The error between Sparameter data and the instrument’s detection results was less than 001%. Users could access realtime Sparameter waveforms simultaneously from multiple devices, proving the system’s practicality and providing a reliable solution for vector network analyzer detection.