Manufacturing scalability and cost efficiency of Lentiviral vector (LV) production for gene therapy are hampered by slow offline detection of critical parameters and viral titer. To enable real-time monitoring, this study proposes a process analytical technology (PAT) based on Raman spectroscopy and an innovative deep learning algorithm to achieve real-time prediction of glucose, lactate, viable cell density, and viral titer during LV production. To address the limitations of existing Raman spectral analysis methods, this study integrated data augmentation, self-supervised learning, and time-series modeling strategies. A regression model integrating a one-dimensional Convolutional Neural Network (1D-CNN) with Squeeze-and-Excitation (SE) attention, enhanced by relative positional encoding (RPE), multi-query attention (MQA), and Top-K feature selection (termed ACNN), was constructed. This approach utilized vast amounts of unlabeled spectral data to precisely extract feature information from complex components, including viral particles. Experimental results demonstrated that the ACNN model significantly outperformed the traditional Partial Least Squares (PLS) model in predicting viral titer, cell density, glucose, and lactate concentrations (paired t-test, p < 0.05 for all parameters). Notably, titer prediction error was substantially reduced. Statistical analysis showed that the RMSE of ACNN for titer prediction was significantly lower than that of PLS (t = 3.82, p < 0.01), and the R2 was significantly higher (t = 4.15, p < 0.001). This study overcomes technical bottlenecks in real-time quality monitoring and process optimization for LV production, providing a feasible transition from "black box operations" to transparent and intelligent monitoring throughout the entire process and establishing a robust foundation for large-scale intelligent production of other gene therapy vectors.

The Rotating Fourier Transform Spectrometer (R-FTS) has garnered significant attention in the field of dynamic fine analysis due to its advantages of high throughput, high spectral resolution, high spectral scanning speed, and ease of miniaturization. However, studies have shown that stray light reflected from the surfaces of its rotating refractor can induce parasitic interference. This leads to frequency aliasing in the interferogram and generates spurious peaks in the spectrum, thereby distorting the true spectral information and compromising the accuracy of quantitative analysis and feature identification. Based on the optical system structure of the R-FTS, this paper conducts an in-depth analysis of the optical path difference between the stray light and the main beam. A multi-beam interference intensity distribution model incorporating the influence of stray light is established, systematically revealing the mechanism of stray light parasitic interference. On this basis, an optimized structure specifically for the R-FTS optical system is proposed, which can effectively suppress the highly detrimental stray light. Experimental results are consistent with theoretical predictions. This research provides a clear and feasible technical pathway to address the critical challenge of stray light parasitic interference in R-FTS.

In recent years, the rotating Fourier transform spectrometer (R-FTS) has shown significant engineering potential in the field of dynamic fine analysis due to its advantages of 4 cm(-1) high spectral resolution and 100 Hz high spectral scanning speed. However, it has been found that the parallelism error of the rotating refractor in R-FTS can significantly degrade the interference modulation depth, resulting in the loss of weak spectral features or a reduction in resolution. Based on the analysis of the R-FTS optical system incorporating non-ideal rotating refractor, this paper derives a computable mathematical model relating the wedge angle and rotation angle of the non-ideal parallel rotating refractor to the modulation depth, thereby revealing the transfer mechanism between parallelism error and modulation depth. Accordingly, a new assembly process for the rotating refractor is proposed, which relaxes the parallelism error tolerance from 6 '' to 86 ''. Experimental results are in good agreement with the model predictions. This study provides a transferable design basis for tolerance allocation and alignment technology in rotating interferometers.

The geometric accuracy of aero-engine blades is critical to engine performance and operational safety, imposing stringent requirements on blade profile measurement precision. This article presents a 3-D blade measurement system based on a four-axis platform equipped with a spectral confocal sensor. To model system-induced geometric deviations, a multitilt geometric error model (MT-GEM) is established to jointly account for guide-rail inclination errors, sensor installation errors, and rotary-axis deviations. The MT-GEM is solved in a two-stage manner: initial estimates of the guide-rail inclination parameters are first obtained via ellipsoid fitting combined with linear least-squares estimation, based on which the initial kinematic parameters of the rotary table, including the rotation axis and center, are derived. Subsequently, a joint nonlinear optimization is performed to refine all MT-GEM parameters. To enable efficient calibration with minimal artifacts, a three-parameter constrained sphere pair model (TPCSPM) is introduced, allowing system calibration using only a pair of standard spheres. Experimental results demonstrate that the proposed framework achieves robust and accurate parameter estimation and significantly reduces measurement deviations. After calibration, the radius error of a standard sphere is reduced to 0.018 mm, validating the effectiveness of the proposed method for high-precision, noncontact blade inspection.

As advanced nanoactuators, nano flexible scanning platforms can quickly and stably achieve multi-axis scanning positioning at the nanoscale, and they are widely used in advanced fields such as chemical detection, micro–nano measurement, microelectromechanical control, and optical positioning. Such platforms usually include horizontal and vertical scanners, whose mechanical structure and combined multi-axis scanning will directly affect the response speed and displacement performance. The circuit and measurement system and nonlinear control algorithm are also crucial, since small changes to these can affect platform resolution and stability. Therefore, in this paper, a nano scanning platform with flexible hinge structure and a dedicated circuit and measurement system is designed. The platform has a three-axis structure combining series and parallel connections. An improved sliding mode closed-loop controller is proposed for nonlinear errors, which further improves the scanning accuracy of the platform through subdivision calculation and segmented calibration methods. Experiments prove the effectiveness of the nano flexible scanning platform and its control algorithm.

Fiber optic gyroscopes are widely used in high-precision angle measurement and inertial sensing because of their all-solid-state structure, high reliability, and fast dynamic response. However, under varying temperature conditions, their performance is often degraded by temperature-induced scale factor error and zero-bias drift. Existing studies usually compensate these two errors separately, which limits the exploitation of their correlation and the overall compensation performance. Moreover, the nonuniform thermal field inside the gyroscope makes single-point temperature information insufficient to characterize the thermal state accurately. To address these issues, this paper proposed a joint temperature compensation method based on Grey Wolf Optimization and Quasi Multi-Task Support Vector Regression (GWO-QMTSVR). A multidimensional temperature feature set was constructed from multi-point temperature measurements, temperature rates, and discrete temperature-gradient information, and the two error components were modeled simultaneously within a unified framework. Grey Wolf Optimization was further introduced to optimize the key parameters of QMTSVR. The results demonstrated that the proposed method effectively improved the joint compensation performance and provided an effective solution for temperature error compensation of fiber optic gyroscopes.

Magnesium and its rare-earth alloys are extensively studied for their lightweight properties and high specific strength, making them attractive for aerospace, automotive, and biomedical applications. However, their hexagonal close-packed structure leads to a strong basal texture, limiting plasticity and formability at room temperature. Considerable research has been devoted to texture control strategies, including alloying, thermomechanical processing, and recrystallization mechanisms, yet a comprehensive understanding of their effects remains an ongoing research focus. This review summarizes recent advances in texture regulation of rare-earth magnesium alloys, focusing on the role of RE elements (Gd, Y, Nd, Ce) and non-RE elements (Zn, Ca) in modifying basal texture and enhancing mechanical properties. The influence of key processing techniques, such as extrusion, rolling, equal channel angular pressing, and rotary shear extrusion, is discussed in relation to their effects on recrystallization behavior. Additionally, the mechanisms governing texture evolution, including continuous dynamic recrystallization, discontinuous dynamic recrystallization (DDRX), and particle-stimulated nucleation, are critically examined. By integrating recent findings, this review provides a systematic perspective on alloying strategies, processing conditions, and recrystallization pathways, offering valuable insights for the development of high-performance magnesium alloys with improved formability and mechanical properties.

The long-term monitoring of respiratory status is crucial for the prevention and diagnosis of respiratory diseases. However, existing continuous respiratory monitoring devices are typically bulky and require either chest strapping or proximity to the nasal area, which compromises user comfort and may disrupt the monitoring process. To overcome these challenges, we have developed a flexible, attachable, lightweight, and miniaturized system designed for extended wear on the wrist. This system incorporates signal acquisition circuitry, a mobile client, and a deep neural network, facilitating long-term respiratory monitoring. Specifically, we fabricated a highly sensitive (11,847.24 kPa-1) flexible pressure sensor using a screen printing process, which is capable of functioning beyond 70,000 cycles. Additionally, we engineered a bidirectional long short-term memory (BiLSTM) neural network, enhanced with a residual module, to classify various respiratory states including slow, normal, fast, and simulated breathing. The system achieved a dataset classification accuracy exceeding 99.5%. We have successfully demonstrated a stable, cost-effective, and durable respiratory sensor system that can quantitatively collect and store respiratory data for individuals and groups. This system holds potential for everyday monitoring of physiological signals and healthcare applications.

Fourier Transform Infrared (FTIR) spectrometers are the mainstay of modern infrared spectral analysis. While traditional FTIR spectrometers and new static FTIR spectrometers each have their strengths, they struggle to balance key performance metrics such as spectral measurement frequency, radiant flux, resolution, and coverage range. Our study has developed a novel FTIR spectrometers based on a rotating interferometer, capable of meeting the demands for high-sensitivity, real-time dynamic analysis of material composition in fields like precise biochemical reaction analysis. To address the non-linearity in optical path difference changes within interferometers, we've designed a new type of rotating interferometer that incorporates a single-frequency laser, enabling equal optical path difference interval sampling and thus preventing spectral errors. The system's effectiveness and feasibility were validated through simulations and alcohol flame tests. This instrument boasts a high radiant flux, achieving a 100 Hz spectral measurement frequency, 4 cm-1 spectral resolution, and a 2-15 mu m spectral coverage range.

Accurate identification of vertical wheel-rail forces is critical for structural safety assessment and the reliable operation of high-speed railways. However, direct measurements are often infeasible, and traditional inverse methods suffer from ill-posedness and sensitivity to operational variability. To overcome these limitations, this study proposes a physics-informed Laplace neural operator-CNN hybrid model (PhysiLNO-CNN) that reconstructs wheel-rail forces from bridge displacement responses. By incorporating structural vibration principles via a Laplace-domain pole-residue formulation, the model captures essential modal dynamics across transient and steady-state regimes. A multiscale CNN encoder-decoder further enhances spatiotemporal feature extraction under sparse sensor layouts. Validation is conducted on a high-speed railway box girder bridge using both traintrack-bridge interaction (TTBI) simulations and field measurements. Compared to LSTM, CNN, Transformer, and DeepONet, PhysiLNO-CNN reduces prediction errors by 25.00 %, 93.75 %, 85.71 %, and 96.90 %, respectively, while requiring over 85 % fewer trainable parameters (excluding DeepONet). These improvements originate from the integration between physics-based constraints and data-driven learning. Additionally, a gradient-based interpretability framework is introduced to evaluate sensor importance and guide optimal sensor placement. The proposed framework provides a lightweight and physically consistent solution for high-fidelity force identification in complex environments.

Color deviation in sunglass lenses can cause various issues when pairing sunglasses, including visual fatigue, discomfort, and imbalances in pupil adjustment. The high color saturation of lenses, the minimal color variation between them, and the sensitivity of color measurement to environmental brightness restrict the effectiveness of traditional machine learning and deep learning methods in detecting and classifying lens colors. In this study, we developed a high signal-to-noise ratio image acquisition system that utilizes a point light source to illuminate the lens and capture the reflected light image, thereby minimizing the impact of ambient brightness on image acquisition. A YCrCb Multi-stage Spectral-wise Transformer (YCrCb-MST) is employed to reconstruct the hyperspectral image of the reflected light, addressing the limitations in classification ability caused by the stacking of the RGB spectrum and the minimal color differences. To improve the accuracy of color classification, this study utilizes the Particle Swarm Optimization-Extreme Learning Machine (PSO-ELM) algorithm. The PSOELM optimizes the input weights and hidden layer bias of the ELM through PSO. Through a specific learning and training process, the optimized ELM achieves satisfactory classification, and the reconstructed hyperspectral images of the lenses are efficiently categorized. The results confirm that the proposed YCrCb-MST-PSO-ELM method achieves an accuracy of 98.69 +/- 0.59 % in the classification of sunglasses lens colors. We propose a low-cost, high-accuracy lens color classification system that offers a novel technical solution for classifying sunglasses lens colors. Additionally, it serves as a valuable reference for addressing color classification challenges in other fields.

Microplastics(MPs)are emerging contaminants in aquatic environments characterized by their polar structure,small particle size(Typically less than 5 mm),large surface area,good stability,and resistance to biodegradation.They pose adverse effects on the normal physiological activities of aquatic organisms and can accumulate in biota,including humans.Therefore,there is an urgent need for rapid and accurate quantitative analysis of MPs in water environments.In this study,Raman spectroscopy combined with partial least squares(PLS)was employed for rapid and accurate quantitative analysis of polyethylene(PE)and polystyrene(PS)MPs in real water samples.Initially,33 simulated water samples containing different concentrations of MPs were prepared,and their Raman spectra were collected.Six spectral preprocessing methods(Normalization,multiplicative scatter correction,standard normal variate transformation,first derivative,second derivative,and wavelet transform)were investigated for their impact on the predictive performance of PLS calibration models.Subsequently,three variable selection methods including synergy interval partial least squares(SiPLS),variable importance in projection(VIP)and mutual information(MI)were employed to optimize the input variables of the PLS calibration model.The predictive capability of the PLS calibration model was evaluated and validated using leave-one-out cross-validation.Under the optimal conditions of spectral preprocessing,variable selection,input variables and latent variables,the wavelet transform-partial least squares(WT-PLS)calibration model based on distilled water was established,and the contents of PE and PS in real water samples were predicted with prediction correlation coefficients(R2p)of 0.9540 and 0.8472 for PE and PS,respectively,and prediction errors(Errorp)of 0.0690 and 0.1126,respectively.Furthermore,a mixed sample MI-PLS calibration model was developed,demonstrating the best predictive performance in real water samples(With R2p values of 0.9776 and 0.9755 for PE and PS,respectively,and Errorp values of 0.0360 and 0.0392,respectively).This method provided a novel approach and new methodology for quantitative analysis of MPs and other organic pollutants in real water samples.

This paper builds a portable digital ion trap mass spectrometer prototype by designing the vacuum system, control circuit system, and host computer of the ion trap mass spectrometer. Dextromethorphan and ketamine samples were tested. At a scanning rate of 5000Th/s, the resolution of the dextromethorphan sample reached 0.26Th, and the resolution of the ketamine sample reached 0.33Th. The actual mass spectrum collected by the prototype is basically consistent with the theoretical mass spectrum.

The screw, a critical element in a variety of transmission mechanisms, significantly influences the performance of the transmission process. Accurate measurement of screw lead is crucial for ensuring the quality of transmission equipment. However, the measurement process can be affected by the precision limits of the measuring instruments and the challenges of manual fine adjustments. This can lead to the screw being misplaced, introducing errors due to the off-center positioning of the workpiece. Such errors can hinder the achievement of high-precision measurements. This research aims to reduce the time needed to adjust for workpiece misalignment and to improve the accuracy of screw lead measurement through error compensation. This study starts by examining two specific scenarios that can cause workpiece misalignment in screw lead accuracy measurements: the tilting of the workpiece and the misalignment of the workpiece axis relative to the circular grating axis. Then, a mathematical model to quantify this misalignment and measure the associated parameters is developed. Based on the measured parameters, a computational model is established to compensate for the bias error under these conditions. This method allows for efficient and precise measurement of the screw lead even when the workpiece is not perfectly aligned. Calibrated screws and a digital micrometer are used to conduct experiments on workpiece misalignment. By comparing measurements with and without error compensation, the effectiveness of the compensation method in enhancing measurement accuracy is demonstrated.

Principal Component Analysis (PCA) aims to acquire the principal component space containing the essential structure of data, instead of being used for mining and extracting the essential structure of data. In other words, the principal component space contains not only information related to the essential structure of data but also some unrelated information. This frequently occurs when the intrinsic dimensionality of data is unknown or when it has complex distribution characteristics such as multi-modalities, manifolds, etc. Therefore, it is unreasonable to identify noise and useful information based solely on reconstruction error. For this reason, PCA is unsuitable as a preprocessing technique for most applications, especially in noisy environment. To solve this problem, this paper proposes robust PCA based on fuzzy local information reservation (FLIPCA). By analyzing the impact of reconstruction error on sample discriminability, FLIPCA provides a theoretical basis for noise identification and processing. This not only greatly improves its robustness but also extends its applicability and effectiveness as a data preprocessing technique. Meanwhile, FLIPCA maintains consistent mathematical descriptions with traditional PCA while having few adjustable hyperparameters and low algorithmic complexity. Finally, we conducted comprehensive experiments on synthetic and real-world datasets, which substantiated the superiority of our proposed algorithm.

This paper presents a novel approach for detecting damage in high-speed railway standard box girders by leveraging the time–frequency characteristics of train-induced strain. Based on the mechanical and deformation characteristics of high-speed railway box girders, the method involves segmenting the box girder into distinct components based on plate element analysis to identify their damage separately. It utilizes coefficients derived from wavelet transforms as indicators sensitive to damage, and employs the particle swarm optimization algorithm (PSO) to determine the ideal frequency intervals in terms of number and position. Within these optimal frequency intervals, the sum of wavelet coefficients is only sensitive to damage features while remaining unaffected by environmental and operational variations. A convolutional denoising autoencoder is employed to remove noise from the strain time–frequency image, improving the robustness of the proposed method. By utilizing the Gaussian inverse cumulative distribution function to estimate confidence boundary (CB) for damage features (DF), outliers are detected, allowing for precise damage localization and quantification. A case study for the high-speed railway box girder show that the proposed method can effectively identify, locate and quantify damage across all components in the critical key section by using data acquired from just 4 strain sensors. This method holds promise for facilitating the development of effective maintenance strategies for high-speed railway box girders.

A new self-supporting material contained nanorods alpha-FeOOH and three-dimensional skeleton graphene foam (3D Graphene Foam, 3DGF) was prepared by hydrothermal method. The growth mechanism and phase change rules of nanorods alpha-FeOOH@3DGF were studied to obtain the electrochemical lithium storage performance. The results showed that growth mechanism of nanorods alpha-FeOOH was accompanied by the consumption of nanoparticles (3-FeOOH. If the hydrothermal time reached to 12 h, the nanoparticle (3-FeOOH phase completely disappeared with only rod-shaped alpha-FeOOH@3DGF single phase obtained. Rod-shaped alpha-FeOOH@3DGF was used as anode material of lithium-ion batteries and the reversible specific capacity remained at about 305 mAh center dot g- 1 after charging and discharging 250 times at 500 mA center dot g- 1, with significantly cycling stability compared with that of the monomaterial alpha-FeOOH.

Piezoelectric ceramic, an indispensable micro-nano actuator, has advantages of small size, high sensitivity and strong controllability, which is widely applied in various fields like micro-nano processing, medicine, chemistry and materials. However, its nonlinear and asymmetric hysteresis features directly affect its positioning accuracy. The classical and generalized Prandtl-Ishlinskii models with linear symmetry operators are commonly used to deal with these issues, but they still face difficulties in asymmetric systems. Therefore, a piecewise modified Prandtl-Ishlinskii (PMPI) model is proposed to compensate nonlinear and asymmetric hysteresis in this paper. The new nonlinear operator and piecewise nonlinear envelope function are introduced into the PMPI model. Independent right and left parts are adopted to describe hysteresis increasing and decreasing stages, while the threshold and weight vectors correspond to them. Weight vectors are extended and a new diagonal matrix is constructed for avoiding excessive parameters and computation. Simulations and experiments are shown that the PMPI model has good compensation performance and effectiveness, which has better fitting, higher accuracy, fewer parameters and computation, and can be well applied to the precise nano-positioning systems with less nonlinear and asymmetric errors.

With the increasingly widespread application of freeform lenses, the detection of lens dioptric power presents a challenge. To measure the surface dioptric power of freeform lenses, a wavefront detection system for eyeglass lenses based on transmissive phase measuring deflectometry has been developed. The system consists only of a monitor, an industrial camera, and the lens under test. In response to the phase errors introduced by nonlinear devices, a Frequency Domain Gamma Error Correction Algorithm based on Linear Regression (FDLR) is proposed to optimize the system detection process and enhance detection accuracy. In the experiments, the proposed system was used to measure the wavefront and surface dioptric power of both spherical lens and progressive addition lens. The results were compared with those obtained from a focimeter to verify the accuracy and reliability of the system. Furthermore, a comparison of the detection results for progressive addition lens before and after phase error correction was conducted to demonstrate the necessity and effectiveness of the proposed phase error correction algorithm.