The traditional eddy current displacement sensor is limited by the diameter of the probe coil, and the measuring range is relatively small. In order to improve the range of an eddy current displacement sensor, a differential compensated eddy current displacement sensor (DCECDS) is proposed in this paper. The DCECDS is designed with three coils and is characterized by a large measuring range and good linearity. Based on the analysis of the working principle of DCECDS, the corresponding equivalent circuit model and mathematical model were established in this study. The effects of excitation frequency, inner diameter and thickness of excitation coil on the sensor performance were investigated. Based on the simulation results, the structural parameters of the sensor were designed and verified by experiments. The experimental results show that the measuring range of the DCECDS designed in this study can reach 1.76 times the diameter of the probe coil, which is 3.5 times higher than that of the traditional eddy current displacement sensor. The measuring range of the eddy current displacement sensor is effectively improved, which provides an important reference and practical value for the design of eddy current displacement sensors.

Developing highly efficient and durable trifunctional electrocatalysts is essential for advancing technologies related to water splitting, oxygen reduction, and hydrogen storage, which are critical for energy conversion and storage. In this study, we encapsulated Co-MOFs on the surface of Mo2C hollow spheres and achieved the Co@NC/Mo2C composite material through high-temperature annealing. In the optimal catalyst, Co@NC/Mo2C-0.1, Co@NC nanosheets were uniformly distributed on the Mo2C hollow spheres, and the BET surface area was well maintained. Compared to Co nanoparticles, Co-Nx in Co@NC/Mo2C-0.1 exhibited a greater modulation effect on the electronic structure of Mo2C. Density functional theory (DFT) calculations showed that the Co@NC/Mo2C heterostructure effectively lowered the energy barrier for the hydrogen evolution reaction (HER) on Mo2C and enhanced the performance of the oxygen evolution reaction (OER) and oxygen reduction reaction (ORR) through electronic coupling. Specifically, for Co@NC/Mo2C-0.1, at a current density of 10 mA cm(-2), the half-wave potential for ORR was 0.86 V, the overpotential for OER was 349 mV, and the overpotential for HER was 158 mV. Additionally, this catalyst exhibited exceptional stability across all three reactions, making it an ideal candidate for sustainable energy technologies.

Conducting bench tests for tractors according to the load spectrum is essential for assessing their reliability and fatigue durability. However, achieving an accurate replication of field operating conditions poses challenges due to the disparity in response characteristics between the tractor bench loading system and the frequency data of the load spectrum. Consequently, discrepancies in test results arise. To mitigate this challenge, this study proposes a loading method for the tractor rotary tillage load spectrum based on extreme load retention resampling. Firstly, a tractor rotary tillage load acquisition test was conducted, and a one-time-extrapolated load spectrum was compiled based on the Peak Over Threshold model. Considering the characteristics of rotary tillage operations, the tractor rotary tillage load spectrum loading bench was developed, which primarily includes the Power Take-Off (PTO) loading system and the suspension loading system. On this basis, a rotary tillage load spectrum loading system based on a Fuzzy-PID controller was proposed to realize the dynamic loading of the rotary tillage load spectrum. The dynamic response characteristics of the loading bench were analyzed based on a simulation model, and the results showed that the loading bench can achieve dynamic loading of load spectra with frequencies below 25 Hz. To match this characteristic, a load spectrum resampling method based on extreme load retention was proposed to resample the rotary tillage load spectrum. Based on the retention of the load spectrum fatigue damage, a resampled rotary tillage load spectrum with a resampling ratio of 3 was obtained, with a loading frequency of 20.98 Hz. Finally, the tractor rotary tillage loading test was conducted based on the resampled rotary tillage load spectrum. The test results demonstrated that the loading bench effectively replicated the resampled rotary tillage load spectrum. For the PTO torque load spectrum, the average error is -7.53%, with a delay of 30 ms, and for the suspension load spectrum, the average error is -2.48%, with a delay of 22 ms. The result indicated that the resampled load spectrum can well match the dynamic characteristics of the loading bench. This research can serve as a practical reference for implementing tractor bench tests grounded in load spectra.

Channel pruning is a method to compress convolutional neural networks, which can significantly reduce the number of model parameters and the computational amount. Current methods that focus on the internal parameters of a model and feature mapping information rely on artificially set a priori criteria or reflect filter attributes by partial feature mapping, which lack the ability to analyze and discriminate the channel feature extraction and ignore the basic reasons for the similarity of the channels. This study developed a pruning method based on similar structural features of channels, called SSF. This method focuses on analysing the ability to extract similar features between channels and exploring the characteristics of channels producing similar feature mapping. First, adaptive threshold coding was introduced to numerically transform the channel characteristics into structural features, and channels with similar coding results could generate highly similar feature mapping. Secondly, the spatial distance was calculated for the structural features matrix to obtain the similarity between channels. Moreover, in order to keep rich channel classes in the pruned network, different class cuts were made on the basis of similarity to randomly remove some of the channels. Thirdly, considering the differences in the overall similarity of different layers, this study determined the appropriate pruning ratio for different layers on the basis of the channel dispersion degree reflected by the similarity. Finally, extensive experiments were conducted on image classification tasks, and the experimental results demonstrated the superiority of the SSF method over many existing techniques. On ILSVRC-2012, the SSF method reduced the floating-point operations (FLOPs) of the ResNet-50 model by 57.70

To enable the highly sensitive detection of acetylene (C2H2) dissolved in transformer oil, a high-power quartz-enhanced photoacoustic spectroscopy (QEPAS) sensing system is proposed. A standard 32.7 kHz quartz tuning fork (QTF) was employed as an acoustic transducer, coupled with an optimized acoustic resonator to enhance the acoustic signal. The laser power was boosted to 150 mW using a C-band erbium-doped fiber amplifier (EDFA), achieving a detection limit of 469 ppb for C2H2 with an integration time of 1 s. The headspace degassing method was utilized to extract dissolved gases from the transformer oil, and the equilibrium process for the release of dissolved C2H2 was successfully monitored using the developed high-power QEPAS system. This approach provides reliable technical support for the real-time monitoring of the operational safety of power transformers.

This study explored the optimization of control systems for atmospheric pipeline air-floating vehicles traveling at ground level by introducing a novel composite wheel-fan system that integrates both wheels and fans. To evaluate the control impedance, the system simulates road conditions like inclines, uneven surfaces, and obstacles by using fixed, random, and high torque settings. The hub motor of the wheel fan is managed through three distinct algorithms: PID, fuzzy PID, and the backpropagation neural network (BP). Each algorithm’s control strategy is outlined, and tracking experiments were conducted across straight, circular, and curved trajectories. Analysis of these experiments supports a hybrid control approach: initiating with fuzzy PID, employing the PID algorithm on straight paths, and utilizing the BP neural network for sinusoidal and circular paths. The adaptive capacity of the BP neural network suggests its potential to eventually supplant the PID algorithm in straight path scenarios over extended testing and operation, ensuring improved control performance.

Pruning convolutional neural networks offers a promising solution to mitigate the computational complexity challenges encountered during application deployment. However, prevalent pruning techniques primarily concentrate on model parameters or feature mapping analysis to devise static pruning strategies, often overlooking the underlying feature extraction capacity of convolutional kernels. To address this, the study first quantitatively expresses the feature extraction capability of convolutional channels from three aspects: global features, distribution metrics, and directional metrics. It explores the multi-dimensional information of the channels, calculates the overall expectation, variance, and cosine distance from the unit vector as the quantitative results of the channels. Subsequently, a clustering algorithm is employed to categorize the multidimensional information. This approach ensures that convolutional channels grouped within each cluster possess similar feature extraction capabilities. An enhanced differential evolutionary algorithm is utilized to optimize the number of clustering centers across all convolutional layers, ensuring optimal grouping. The final step involves achieving channel sparsification through the calculation of crowding distances for each sample within its designated cluster. This preserves a diverse subset of channels that are critical for maintaining model accuracy. Extensive empirical evaluations conducted on three benchmark image classification datasets demonstrate the efficacy of this method. For instance, on the ImageNet dataset, the ResNet-50 model experiences a substantial reduction in FLOPs by 58.43

A gas analyzer has been developed using cavity ring-down spectroscopy (CRDS) technology, designed to simultaneously measure the concentrations of CO2 and CO. These gases are critical indicators of transformer insulation degradation. The analyzer covers a CO2 concentration range from a few ppm to several thousand ppm and can measure CO/CO2 concentration ratios from 0.076 to 0.8. The selected absorption lines for CO and CO2 are centered at 6337.99 and 6338.59 cm-1, respectively. The two candidate lines are proximate to each other and nonoverlapping. The experimental results indicate that the system accurately fits the measured and calculated signals, enabling concentration measurements at the ppm level. This confirms its ability to measure the gas concentrations and ratios critical for assessing the condition of transformer insulation. This CRDS-based system offers a reliable and sensitive method for early detection of potential transformer faults.

Pruning is an efficient method for deep neural network model compression and acceleration. However, existing pruning strategies, both at the filter level and at the channel level, often introduce a large amount of computation and adopt complex methods for finding sub-networks. It is found that there is a linear relationship between the sum of matrix elements of the channels in convolutional neural networks (CNNs) and the expectation scaling ratio of the image pixel distribution, which is reflects the relationship between the expectation change of the pixel distribution between the feature mapping and the input data. This implies that channels with similar expectation scaling factors ($\delta _{E}$) cause similar expectation changes to the input data, thus producing redundant feature mappings. Thus, this article proposes a new structured pruning method called EXP. In the proposed method, the channels with similar $\delta _{E}$ are randomly removed in each convolutional layer, and thus the whole network achieves random sparsity to obtain non-redundant and non-unique sub-networks. Experiments on pruning various networks show that EXP can achieve a significant reduction of FLOPs. For example, on the CIFAR-10 dataset, EXP reduces the FLOPs of the ResNet-56 model by 71.9% with a 0.23% loss in Top-1 accuracy. On ILSVRC-2012, it reduces the FLOPs of the ResNet-50 model by 60.0% with a 1.13% loss of Top-1 accuracy. Our code is available at: https://github.com/EXP-Pruning/EXP_Pruning and DOI: 10.5281/zenodo.8141065.

Multi-kilojoule, multi-picosecond short-pulse lasers, such as the National Ignition Facility-Advanced Radiographic Capability laser and the OMEGA-Extended Performance laser, which have been constructed over the last two decades, enable exciting opportunities to produce high-brightness, high-energy laser-driven proton sources for applications in high-energy-density science like proton fast ignition for inertial fusion energy, particle radiography, and materials science studies. Results on these platforms have demonstrated enhanced accelerated proton energies and electron temperatures when compared to established scaling laws. Recent work has developed a new scaling for proton TNSA in the multi-ps regime. However, this new physics in the multi-ps regime motivates the need to understand the origin of the enhancement in proton energies. Toward this goal, this work presents the first measurements of the TNSA accelerating sheath field in the multi-ps regime for pulse durations of 0.6, 5, and 10 ps. This measurement was achieved by using a separate TNSA proton source to radiograph the spatiotemporal profile of the accelerating sheath that is responsible for proton acceleration. The use of stacked radiochromic film detectors allows for a discrete time profile of the radiographs, thus enabling the measurement of the temporal and spatial evolution of the accelerating field. In performing this measurement, we extract quantities such as the sheath strength as a function of time and pulse duration, which shows that longer pulse durations sustain a stronger electric field for a longer duration when compared to sub-ps laser pulses, which may enable the observed boosted proton energies and proton conversion efficiencies.

In order to understand how close current layered implosions in indirect-drive inertial confinement fusion are to ignition, it is necessary to measure the level of alpha heating present. To this end, pairs of experiments were performed that consisted of a low-yield tritium-hydrogen-deuterium (THD) layered implosion and a high yield deuterium-tritium (DT) layered implosion to validate experimentally current simulation-based methods of determining yield amplification. The THD capsules were designed to reduce simultaneously DT neutron yield (alpha heating) and maintain hydrodynamic similarity with the higher yield DT capsules. The ratio of the yields measured in these experiments then allowed the alpha heating level of the DT layered implosions to be determined. The level of alpha heating inferred is consistent with fits to simulations expressed in terms of experimentally measurable quantities and enables us to infer the level of alpha heating in recent high-performing implosions.

We report on the first experiment dedicated to the study of nuclear reactions on dopants in a cryogenic capsule at the National Ignition Facility (NIF). This was accomplished using bromine doping in the inner layers of the CH ablator of a capsule identical to that used in the NIF shot N140520. The capsule was doped with 3$\times$10$^{16}$ bromine atoms. The doped capsule shot, N170730, resulted in a DT yield that was 2.6 times lower than the undoped equivalent. The Radiochemical Analysis of Gaseous Samples (RAGS) system was used to collect and detect $^{79}$Kr atoms resulting from energetic deuteron and proton ion reactions on $^{79}$Br. RAGS was also used to detect $^{13}$N produced dominantly by knock-on deuteron reactions on the $^{12}$C in the ablator. High-energy reaction-in-flight neutrons were detected via the $^{209}$Bi(n,4n)$^{206}$Bi reaction, using bismuth activation foils located 50 cm outside of the target capsule. The robustness of the RAGS signals suggest that the use of nuclear reactions on dopants as diagnostics is quite feasible.

We report on experimental results from a high-intensity laser interaction with cone targets that increase the number (×3) and temperature (×3) of the measured hot electrons over a traditional planar target. This increase is caused by a substantial increase in the plasma density within the cone target geometry, which was induced by 17 ± 9 mJ prepulse that arrived 1.5 ns prior to the main high intensity (>1019 W/cm2). Three-dimensional hydrodynamic simulations are conducted using hydra which show that the cone targets create substantially longer and denser plasma than planar targets due to the geometric confinement of the expanding plasma. The density within the cone is a several hundred-micron plasma “shelf” with a density of approximately 1020 ne/cc. The hydra simulated plasma densities are used as the initial conditions for two-dimensional particle-in-cell simulations using EPOCH. These simulations show that the main acceleration mechanism is direct-laser-acceleration, with close agreement between experimentally measured and simulated electron temperatures. Further analysis is conducted to investigate the acceleration of the electrons within the long plasma generated within a compound parabolic concentrator by the prepulse.

In this paper, a new vehicle distance detection technology based on image processing is proposed to solve the traditional ultrasonic vehicle distance detection errors, many blind areas and other problems. The binocular vision sensing system was used to collect real-time images in front of vehicles. The vehicle in front of the collected image is identified through template matching, and the parallax principle is used to calculate the vehicle distance after confirming the presence of the vehicle. Experimental results show that this method can detect the distance between different angles and different positions with high accuracy.

Aiming at the special nature of the torque testing environment for special vehicle transmission shafts, the torque testing method based on embedded capacitive grid sensors can solve the problems of small installation space, difficult leads, and easy signal interference. To solve the Edge effects problem of capacitive grid torque sensor, Ansoft Maxwell electromagnetic field finite element analysis software is used for simulation analysis. By changing the gate thickness and plate spacing with a fixed number of capacitor gate pairs, the trend of capacitance values with the two parameters can be obtained. Experimental research was conducted on this issue through a simulated rotating shaft experimental platform, and the experimental scheme was set by changing the gate thickness and plate spacing of the capacitive grid torque sensor. The results show that the change trend of the experimental data is the same as that of the simulation data. In the actual measurement experiment, the method of reducing the grid thickness and plate spacing can be used under the conditions of manufacturing technology and cost constraints to reduce the impact of Edge effects and improve the test accuracy.

In order to reduce the impact of the environment on the accuracy and sensitivity of detection, and to meet the requirements of concealment from detection and being lightweight, a technology for detecting flying metal objects based on photoelectric composite sensors is proposed. The method first analyzes the target’s characteristics and detection environment, and then compares and analyzes the methods for detecting typical flying metal objects. On the basis of the traditional eddy current model, the photoelectric composite detection model that meets the requirements of detecting flying metal objects was studied and designed. For the problems of the short detection distance and the long response time of the traditional eddy current model, the performance of the eddy current sensor was improved to meet the requirements of detection through optimizing the detection circuit and coil parameter model. Meanwhile, to meet the goal of being lightweight, an infrared detection array model applicable to flying metal bodies was designed, and simulation experiments of composite detection based on the model were conducted. The results show that the flying metal body detection model based on photoelectric composite sensors met the requirements of distance and response time for detecting flying metal bodies and may provide an avenue for exploring the composite detection of flying metal bodies.

Computer models of intense, laser-driven ion acceleration require expensive particle-in-cell simulations that may struggle to capture all the multi-scale, multi-dimensional physics involved at reasonable costs. Explored is an approach to ameliorate this deficiency using a multi-fidelity framework that can incorporate physical trends and phenomena at different levels. As the basis for this study, an ensemble of approximately 8000 1D simulations was generated to buttress separate ensembles of hundreds of higher fidelity 1D and 2D simulations. Using transfer learning with deep neural networks, one can reproduce the results of more complex physics at a much lower cost. The networks trained in this fashion can, in turn, act as surrogate models for the simulations themselves, allowing for quick and efficient exploration of the parameter space of interest. Standard figures-of-merit were used as benchmarks such as the hot electron temperature, peak ion energy, conversion efficiency, and so on. We can rapidly identify and explore under what conditions differing fidelities become an important effect and search for outliers in feature space.

We present the development of a compact Thomson parabola ion spectrometer capable of characterizing the energy spectra of various ion species of multi-MeV ion beams from >1020W/cm2 laser produced plasmas at rates commensurate with the highest available from any of the current and near-future PW-class laser facilities. This diagnostic makes use of a polyvinyl toluene based fast plastic scintillator (EJ-260), and the emitted light is collected using an optical imaging system coupled to a thermoelectrically cooled scientific complementary metal-oxide-semiconductor camera. This offers a robust solution for data acquisition at a high repetition rate, while avoiding the added complications and nonlinearities of micro-channel plate based systems. Different ion energy ranges can be probed using a modular magnet setup, a variable electric field, and a varying drift-distance. We have demonstrated operation and data collection with this system at up to 0.2 Hz from plasmas created by irradiating a solid target, limited only by the targeting system. With the appropriate software, on-the-fly ion spectral analysis will be possible, enabling real-time experimental control at multi-Hz repetition rates.

We present the development of a flexible tape-drive target system to generate and control secondary high-intensity laser-plasma sources. Its adjustable design permits the generation of relativistic MeV particles and x rays at high-intensity (i.e., ≥1 × 1018 W cm-2) laser facilities, at high repetition rates (>1 Hz). The compact and robust structure shows good mechanical stability and a high target placement accuracy (<4 μm RMS). Its compact and flexible design allows for mounting in both the horizontal and vertical planes, which makes it practical for use in cluttered laser-plasma experimental setups. The design permits ∼170° of access on the laser-driver side and 120° of diagnostic access at the rear. A range of adapted apertures have been designed and tested to be easily implemented to the targetry system. The design and performance testing of the tape-drive system in the context of two experiments performed at the COMET laser facility at the Lawrence Livermore National Laboratory and at the Advanced Lasers and Extreme Photonics (ALEPH) facility at Colorado State University are discussed. Experimental data showing that the designed prototype is also able to both generate and focus high-intensity laser-driven protons at high repetition rates are also presented.

Abstract Convolutional neural networks are limited by the large number of floating-point operations (FLOPs) and cannot be easily ported to edge devices. Popular pruning methods formulate a fixed pruning strategy with the assistance of the intrinsic attributes of the model, ignoring of the effect of the dataset on the pruning results. This study introduces the expectation scaling factor (δE) as the zero-order attribute of the channel, which reflects the expectation changing of the channel on the pixel distribution of the image dataset. This attribute establishes a connection between the input and output of the convolutional layer and reflects the overall information of the channels. Then, during the fine-tuning, we formulate a pruning strategy by observing the changes in δE to improve the adaptability of the pruned model to a specific dataset, and call this method DExp. Extensive experiments on representative image classification tasks reveal that DExp outper-forms mainstream methods. For example, with DExp, on CIFAR-10, the FLOPs of ResNet-56 are reduce by 53.99%, and the accuracy improved by 0.38%; on ILSVRC-2012, the FLOPs of ResNet-50 are reduced by 58.43%, and the Top-1 accuracy only loses by 1.15%. The code is available at: https://github.com/EXP-Pruning/DExp pruning.