The working principles and current instrument development status of Secondary Ion Mass Spectrometry （SIMS） and Atom Probe Tomography （APT） are elucidated， analyzing their applications across metals， semiconductors， biology， and geology. SIMS enables trace element characterization and interfacial segregation analysis， playing a pivotal role in semiconductor impurity detection， failure analysis， geological dating， and biomolecular imaging. APT enables atomic-level resolution of material composition and spatial distribution， providing critical support for studies on metal strengthening mechanisms， semiconductor doping distribution， biomineralization， and nano-geochemistry. The development of relevant standards and outlines future trends are summarized， including breakthroughs in instrument sensitivity and resolution， exploration of novel ion sources and in-situ analysis techniques， integration with artificial intelligence for data processing， and continuous refinement of the standards system. These techniques will further advance frontier fields such as advanced materials， life sciences， and earth sciences.

Existing aggregate gradation detection methods based on machine vision primarily rely on aggregate size information for gradation analysis. However， size information alone does not directly reflect aggregate mass， leading to significant detection errors. To address this issue， a gradation detection method based on shape feature tensors is proposed. In this method， aggregate size and shape features are extracted from two-dimensional images. The size features are used to determine the probability distribution of aggregate size intervals， while shape feature tensors are constructed. A convolutional neural network is then employed to predict aggregate shape factors. By integrating the size interval probability distribution and shape factors， aggregate quantity is converted into equivalent sphere quantity. Finally， equivalent sphere quantity serves as input data for an MCMC algorithm based on Bayesian inference to obtain gradation detection results. Experimental validation shows that the absolute detection error remains within ±2.5%， meeting the ±5% accuracy requirement for engineering applications and outperforming methods based solely on size features.

Mismatch and displacement induced by mechanical failures of aero-engine components under high-temperature conditions will lead to operational abnormalities， which seriously threaten flight safety. To realize the operational condition monitoring of rigid curved shafts in aero-engines under harsh environments， a wireless radial displacement measurement method for rigid curved shafts based on impedance matching and eddy current effect is proposed. The sensor probe is fabricated by screen printing and high-temperature sintering processes. The displacement signal is wirelessly transmitted to the network analyzer through electromagnetic coupling between the probe and the rigid curved shaft， where variations in displacement will result in corresponding changes in S ₁₁ （return loss）. Test results show that under normal temperature conditions， the displacement exhibits a strictly monotonic relationship with S ₁₁ within the range of 0~15 mm， which verifies the feasibility of this method for wireless displacement measurement of rigid curved shafts. Under the high-temperature environment of 200 °C， the displacement sensor achieves a resolution of 10 µm， with a maximum repeatability error of 1.16% and a maximum relative error of 1.19%.

3D reconstruction serves as a key foundation for precise minimally invasive surgery. However， poor texture and insufficient lighting in cavities pose significant challenges to feature matching， a crucial part of 3D reconstruction. a hierarchical structure-enhanced feature matching method is proposed. A C-SuperPoint feature extraction model， integrated with multi-scale attention mechanisms， enhances detection in weak texture areas， increasing the number of extracted feature points by 33.70% on average. Additionally， a P-LightGlue feature matching algorithm based on the transformer architecture performs initial rough matching and refines results through progressive consistency sampling， effectively filtering mismatches and fitting the model. This approach achieves up to 98.72% accuracy and a minimum average matching error of 1.065 1 pixels across multiple datasets. The results confirm that C-SuperPoint can effectively improve feature extraction performance， while P-LightGlue significantly enhances the accuracy and reliability of image processing in minimally invasive surgery.

To improve the measurement performance of domestic ocean temperature thermometers， a study on dynamic temperature measurement errors and correction methods was conducted based on a self-developed ocean temperature sensor. Stability testing and precise static water bath calibration of the self-developed thermometer were carried out， achieving an expanded uncertainty of 0.93 mK （k=2）. A large high-stability thermostatic bath was utilized to simulate actual sea trial environments， and dynamic temperature measurement experiments were conducted at 5 ℃ and 35 ℃. The results indicate that during the sensor's deployment， temperature changes in the measuring circuit caused a deviation between the measured temperature and the true temperature， with the maximum deviation reaching approximately 0.40 mK. By applying temperature compensation to the measuring circuit， the stability of the temperature measurement was effectively improved. The research findings provide technical references for the development of high-precision domestic ocean temperature thermometers.

When metrology laboratories employ the comparative method in cryogenic calibration systems， the thermostat frequently utilizes cryogenic fluid as the cold source. This kind of thermostat has the defects of large consumption of cryogenic fluid， short durations of single constant temperature maintenance and inconvenience for use in non-fixed places. These factors result in high operating costs and fail to meet the needs of rapid and continuous calibration of thermometers in industrial production. To address these issues， this study introduces a system that uses a small pulse tube refrigerator as the cold source and establishes a heat transfer and control model to achieve rapid evaluation of PID parameters. The system's well thermostatic structure is used to optimize the adiabatic performance of the system. The system achieves continuous temperature control in a wide temperature range of 90~290 K. The temperature fluctuation of the system is less than ±0.030 K/30 min and the axial temperature uniformity is better than 0.010 K， fulfilling the traceability requirements of conventional industrial-grade cryogenic temperature sensors below 200 K.

To address the problems of large sample size， long test cycle and low efficiency in the model construction of existing machine learning-based in-service inspection models for ultrasonic flowmeters， backpropagation （BP） neural network and random forest models under different sample sizes were established based on experimental data from the national urban gas flow standard device （uncertainty 0.26%， k=2）. A small sample set was obtained by halving the initial sample set （time interval 6 s，30 s） using the arithmetic mean method. The results show that the prediction performance （evaluated by R² and root mean square erro） of both models under small sample size is lower than that of the initial sample size model， and the BP neural network has an overall better performance. To improve the performance of the small sample model， three feature optimization algorithms （ReliefF， Recursive Feature Elimination， and random forest feature algorithm） were adopted. The results indicate that the random forest feature algorithm achieves the optimal optimization effect， with the maximum improvement of prediction accuracy reaching 48.65% （at 884 m³/h flow point）. The optimized small sample model retains 9~14 key features， and its prediction performance is comparable to that of the initial sample size model， with significantly improved modeling efficiency. Field application verification shows that the predicted indication error of the model is stable within ±1%， which can meet the in-service inspection requirements of ultrasonic flowmeters in natural gas transmission stations.

An industrial robot operational status monitoring system is designed. The system adopts a joint monitoring device to acquire audio signals generated by the operation of robot joints. To address the difficulty in abnormal feature analysis of audio signals， an SVMD_IBWO_MCKD method is proposed. First， decomposes the audio signal into multiple intrinsic mode functions（IMFs） using the sequential variational mode decomposition（SVMD） method， and then screens out the optimal IMF through the Gaussian weighted kurtosis index. Secondly， the improve beluga whale optimization（IBWO） algorithm is utilized to adaptively select the parameters T， M and L of maximum correlation kurtosis deconvolution（MCKD）， and perform MCKD processing on the selected optimal IMF. Finally， fault features in the robot joint audio signals are extracted through envelope spectrum analysis. The experimental results show that neither the BWO_MCKD nor the WOA_MCKD method can extract the effective octave， and the SSA_MCKD method can only extract the 3-octave component. In contrast， the SVMD_IBWO_MCKD method can effectively extract the fourfold frequency of the periodic fault frequency in the joint audio signal.

By analyzing the water profiling and bottom track modes of the broadband acoustic Doppler current profiler（BBADCP） beam irradiation areas of the two modes， the scatterers within the area are replaced by the unit reverberation volume or the unit reverberation area. The volume and bottom reverberation theories are adopted to predict the echo acoustic intensity of the proposed model. Based on acoustic scattering theory， analytical expressions are derived for the volume backscattering model and the bottom backscattering model， respectively. Compared with the traditional modeling method based on superposition of scattered signals from discrete scatterers， all scatterers within the irradiation area are characterized by the volume reverberation concept in the proposed model， which effectively reduces the computational complexity. The model can accurately predict the acoustic intensity of echo signals. Furthermore， its analytical expression explicitly reflects the influence of actual environmental parameters， enabling the model-generated signals to be more consistent with real underwater conditions. Under the environmental conditions consistent with the experimental site， the simulated echo signals agree well with the measured data in terms of echo intensity， statistical distribution and time-domain waveform characteristics. The relative error of the predicted water-column backscattering intensity is less than 4%， and the absolute error of the bottom backscattering intensity is less than 3 dB. The instantaneous amplitude and envelope of simulated echoes follow the Gaussian distribution and Rayleigh distribution， respectively. To further verify the accuracy and stability of the model in flow velocity simulation， Doppler shifts are repeatedly estimated using multiple groups of simulated echo signal realizations. The results show that the relative error is lower than 0.5%， and the velocity estimation variance is less than 0.7 mm²/s². Additional tests under different signal-to-noise ratios confirm that the model maintains good applicability when the SNR is above 9 dB.

On the basis of calibrating hydrophone sensitivity by the free-field comparison method，a standard sound source method more suitable for in-situ calibration is proposed. This method enables the calibration of array element sensitivity with real-time compensation of environmental parameters， thereby obtaining sensitivity levels better adapted to actual environmental conditions. Compared with laboratory calibration， the proposed method exhibits higher environmental adaptability and practicality. Lake experiments are carried out on a 15-array vertical linear array to verify the performance of the standard sound source method. The experimental results show that the deviation of array element sensitivity levels obtained by this method from those measured by the free-field comparison method under laboratory conditions is mostly within ±0.8 dB， with a maximum deviation no more than ±1.2 dB， indicating a high consistency between the two methods. The main uncertainty sources are analyzed， and the expanded measurement uncertainty is evaluated as U=1.65 dB （ k = 2）.

To address the adverse effects of underwater multi-source noise interference， dynamic medium absorption， and three-dimensional acoustic path distortion on the localization accuracy of underwater wireless sensor networks， an adaptive localization algorithm for underwater nodes driven by dual-stage RSSI filtering （ALAUN-DSRF） is proposed.First， a two-stage filtering architecture is designed to systematically mitigate the interference of outliers and non-Gaussian noise on RSSI data quality in complex marine environments. Subsequently， based on the corrected RSSI data and environmental parameters， an underwater node ranging optimization method is proposed using the Newton's iterative method to improve the accuracy of underwater acoustic propagation distance estimation. Finally， by deeply integrating the high-accuracy ranging model with a spatial motion model， an environment-motion joint adaptive unscented Kalman filter-based node localization algorithm is designed to achieve dynamic tracking of node motion states. Experimental results demonstrate that compared with traditional algorithms， the proposed ALAUN-DSRF achieves a minimum accuracy improvement rate of 18.89%~45.65% in three-dimensional positioning under simulated dynamic marine environments.

The important information of power quality disturbance signals is easily obscured in the noise environment. In order to solve the problems of signal non-smoothing and feature information loss after noise removal， an improved block-matching and 3D Filtering （BM3D） based on orthogonal matching pursuit （OMP） for denoising power quality disturbance signals is proposed. The method first preprocesses the noisy signals by using a block-matching operation to obtain preparatory blocks and their corresponding similar matching blocks. Subsequently， these similar matching blocks are iteratively processed in the discrete cosine transform （DCT） domain using the OMP algorithm， which selects the features most relevant to the current residual and approximates the optimal solution. This process extracts the most relevant sparse coefficients and efficiently filters out noise. Finally， the processed similar matching blocks are reintegrated to obtain the final denoised signal.

A classification method based on multi-scale extreme point peak distance and improved Shapelet is proposed to address the issues of time-consuming search for subsequences， difficulty in determining the length range of candidate subsequences， and the need for extensive computation and comparison when using Shapelet for time series feature extraction. First， envelope kurtosis is employed as the fitness function. The black-winged kite algorithm （BKA） is used for parameter optimization of the FMD， decomposing the power quality disturbance signal into multiple modal components （IMFs）. Permutation and combination entropy is then applied for modal optimization screening， yielding IMF peak-to-peak distances across multiple scales to determine the optimal sub-sequence length range. Subsequently， the Shapelet method is enhanced. Improvements in sub-sequence blocks， multi-loss functions， and Euclidean distance weight initialization enable direct learning of near-optimal sub-sequence fragments， achieving rapid classification. The proposed method was validated using simulated power quality disturbance signals and multi-condition bearing signals. It achieved 98.75% accuracy in identifying power quality disturbances across 16 simulated environments and 98.00% classification accuracy on the multi-condition bearing dataset， demonstrating the algorithm's capability for rapid and precise signal classification.

Pulse-driven AC quantum voltage superconducting chip is the core device in Josephson Arbitrary Waveform Synthesizer （JAWS）. Tapered coplanar waveguide （CPW） is adopted to ensure the uniformity of current pulse excitation and improve the operating margin. However， fabricating size errors Δ in different layers and interlayer alignment error δ are introduced during the manufacturing process. The above errors result in deviations of the characteristic impedance of tapered CPW， affecting the actual effect of JAWS. To quantify the impact， a simulation based on a 3D full-wave electromagnetic field solver is conducted. The results indicate that the end of waveguide is more sensitive to fabricating errors； WR layer fabricating size error Δ _WR has a stronger impact on the characteristic impedance deviation than BE layer preparation size error Δ _BE and δ， with a maximum value exceeding 1.5 Ω； The characteristic impedance of tapered CPW varies approximately linearly with Δ， and is a quadratic function of δ.

The anionic surfactant sodium dodecyl sulfate （SDS） and the zwitterionic surfactant cocamidopropyl betaine （CAPB）， which do not contain fluorine， are widely used in the daily chemical industry and have moderate prices. When compounded， they are suitable for application in foam extinguishing agents. The surface tension， critical micelle concentration （CMC） and foam properties of the compounding systems with different molar ratios were measured. The interaction mechanism between the two surfactants was discussed with thermodynamic models. The results show that the mixed system of SDS and CAPB exhibits significant synergistic effects. The experimentally measured CMC values are significantly lower than the predicted values of the ideal mixing model， and the interaction parameters are negative values， with activity coefficients less than 1， confirming the mutual attraction between two molecules in the mixed micelles. In terms of foam properties， compounding significantly enhances the foam stability， especially when the molar fraction of CAPB is 0.3， the 50% liquid drainage time is the longest and the foam comprehensive performance is the best.

The inversion method based on concentration observations faces challenges in uncertainty evaluation due to its high-order nonlinear terms and multivariable coupling， which exceed the capabilities of the traditional guide to the expression of uncertainty in measurement （GUM） approach. To address this issue， a measurement model was developed based on the quantitative relationship between posterior emissions and grid-level posterior fluxes. By combining the joint distribution of grid posterior fluxes with the Monte Carlo method， the uncertainty assessment of CO₂ emission was achieved. An inversion experiment was conducted for Zhengzhou in August 2023. The results demonstrated that compared to the conventional bottom-up inventory approach， the simulated concentrations derived from the inversion exhibit smaller deviations from observations at both inversion sites and validation sites. Furthermore， the Monte Carlo method was applied to quantify the uncertainty of total CO₂ emissions， followed by further verification using adaptive Monte Carlo method to ensure the uncertainty results reached convergence. The final best estimate for total CO₂ emissions was determined to be 4.50 Tg， with a standard uncertainty of 0.31 Tg and the probabilistically symmetric 95 % coverage interval ［3.76， 4.95］ Tg.

To predict hand forces and analyze neural control， we propose a meta-adaptive temporal distillation model. This utilizes a LiteSFT-Transformer teacher to extract time-frequency-spatial features and a ProtoMixer student to capture temporal-channel features， using conditional modulation and meta-learning to enhance cross-task transfer. Using A-mode ultrasound and distributed pressure data from fist-clenching， bottle-lifting， and pinching tasks， the model achieved R ² scores of 0.935， 0.944， and 0.954， respectively. Comparative analysis against LSTM， PatchTST， TimeMixer， and TimesNet， along with ablation studies， validated the model's effectiveness.

BNP and NT-proBNP are essential heart failure markers， yet significant cross-platform discrepancies—stemming from in vivo metabolism and inconsistent antibody/calibration strategies—impair clinical thresholding and comparability. Standardization regarding reference procedures， materials， and quality assessment is advancing. Notably， the National Institute of Metrology （China） developed the world's first SI-traceable reference materials and LC-MS candidate methods， establishing a foundation for accuracy. By reviewing molecular traits， assay variations， and standardization pathways， this work provides a reference for the metrological framework of heart failure marker detection.