Hongli Li

In aero-engine altitude simulation tests, the inlet air flow rate is a key parameter for thrust calculation and inlet matching analysis, yet its measurement uncertainty is difficult to evaluate reliably. The input quantities exhibit significant synchronous variation and complex dependence, but the complete correlation structure is rarely obtainable in practice. Consequently, uncertainty evaluations based on the GUM often default to the independence assumption, and conventional Monte Carlo method (MCM) implementations sample each input variable independently when joint distribution information is unavailable. To address this, we propose a high-dimensional MCM approach within the GUM Supplement 1 framework that uses a vine copula to model and preserve input correlations during stochastic propagation, combined with a posterior predictive checking procedure for validation. Simulation results show that ignoring synchronous variation overestimates the standard uncertainty and coverage interval by approximately 16%, whereas the proposed method yields results nearly identical to the correlated reference values while maintaining a stable mean. Experimental data confirm that the method produces more convergent interval estimates and reveal that mean static pressure accounts for approximately 85% of the total uncertainty, identifying the primary target for optimization.

With the rapid development of flexible electronics and smart sensing technology, electronic skin has attracted increasing attention from academia and industry as a bionic sensing system. This paper presents a biomimetic skin system that integrates heterogeneous fiber Bragg grating (FBG) distribution with a bionic fingerprint structure to enable high-precision detection and identification of three-dimensional force vectors and slip states. The system eliminates the need for additional temperature-compensation components, while still effectively reducing the impact of environmental temperature variations. Experimental results demonstrate that the proposed system exhibits good linear response characteristics for both normal and shear force measurements. The normal force measurement range is 0-5 N, with a resolution of 0.042 N and a sensitivity of 1.54 V/N. The resolution of the X-direction shear force is 0.023 N with a sensitivity of 3.79 V/N, while the Y-direction shear force achieves a resolution of 0.012 N and a sensitivity of 7.78 V/N. Furthermore, a Random Forest (RF) algorithm is applied to train and classify slip signals, which enables prediction accuracies exceeding 97.2% for slip direction and velocity. These results provide new insights and methods for improving the performance and expanding the functionality of electronic skin technology, supporting its broader application in intelligent sensing.

The application of dynamic systems in modern industrial measurement is becoming increasingly prevalent, making the credibility analysis of their measurement results critically important. However, current uncertainty evaluation methods face significant challenges when dealing with nonlinear systems. To address this challenge, we propose a Physics-informed Evidential Network (PIENet) framework. This approach integrates governing differential equations with measurement data to impose physical constraints and guide the modeling of unknown system components, thereby enhancing predictive accuracy and robustness. It further introduces deep evidential regression and combines the uncertainty-aware loss function with physical constraints to enhance the model calibration capability and enable efficient quantification of prediction uncertainty. We further validate our proposed method on experimental datasets of nonlinear dynamic systems, and the results demonstrate that the method effectively models system dynamics and provides reliable uncertainty predictions, even in the presence of missing variables in the system equations. This approach makes a significant contribution to the use of deep learning for evaluating uncertainty in dynamic systems, which is crucial for enhancing the reliability of measurements.

The growing demand for three-dimensional (3D) ultra-precision measurement has driven the need for high-precision 3D measuring instruments. The micro-nano coordinate measuring machine (CMM), as a key instrument to meet this demand, has drawn attention to how to ensure its measurement accuracy. This paper proposes a comprehensive spatial error compensation model for a self-developed novel micro-nano CMM to enhance the measurement accuracy. The metrology system of the micro-nano CMM complies with the Abbe principle in the XYZ directions theoretically and is independent of the motion system, structurally avoiding the influence of motion errors. For the main error sources in the measuring machine, a comprehensive spatial error compensation model is established based on coordinate transformation, and error separation experiments are conducted. To validate the accuracy of the measuring machine after error compensation, a 100 mm calibration gauge block is measured in the spatial diagonal directions. The results indicate that over 72% of measurement errors are reduced, with the maximum indication error being <203 nm. The overall measurement uncertainty is better than 75.8 nm (k=1) with a measuring volume of 100 mm × 100 mm × 100 mm. The error compensation model presented in this paper significantly improves the accuracy of the measuring machine, which provides a reference for the development of ultra-precision 3D instruments.

In the field of advanced manufacturing, with the rapid development of micromachining and microelectromechanical systems, the demand for 3D high-precision measurement is increasingly urgent. To address the contradiction between large-stroke motion and high-precision measurement, a new cross-scale micro-nano coordinate measuring machine (CSMN-CMM) is developed in this paper. The measuring machine’s measurement method is with the probe fixed and the worktable(workpiece) moving. The metrology system consists of three-axis laser interferometers and an angle measurement system, which can measure the worktable’s spatial coordinates and posture in real-time. The entire metrology system adheres to the Abbe principle and is also independent of the motion system, which avoids the influence of the motion system. The motion system comprises a 3D large-stroke macro-motion stage and a six-degree-of-freedom (6-DOF) micro-motion stage, which can achieve high-precision and rapid driving of the worktable within a motion range of 100 mm × 100 mm × 100 mm. The developed 6-DOF micro-motion stage is driven through piezoelectric ceramics. It can achieve more than 20 μm linear displacement for each axis with a resolution better than 0.56 nm and over 27″ angular displacement with a resolution better than 0.09″. It ensures precise driving and spatial posture adjustment of the worktable. In terms of the control of the CSMN-CMM, the paper proposes macro-micro collaborative positioning and secondary triggering strategies to enhance measurement performances. These strategies achieve positioning accuracy better than 6.4 nm and the repeatability of the probe better than 20 nm (k=1) in the 3D direction. The maximum permissible error of the CSMN-CMM is evaluated following ISO 10360-2, which is better than ±(250 + 3.6 × 10^-6L) nm. A series of experimental results shows that the developed CSMN-CMM has outstanding measurement performance in 3D space. Furthermore, its innovative structural design and key technologies are important for enhancing the current micro-nano measurement technology.

In microstructure geometry measurement, the lateral resolution of optical methods is limited by diffraction, which is generally in the micrometer and sub-micrometer levels. To achieve the measurement with nanometer resolution required for ultra-precision manufacturing, this paper innovatively extends the astigmatism method to lateral measurement. By analyzing the morphological changes of the astigmatic spot when the edge passing through the beam, a model is developed that relates focus error signal to edge position. Based on this model, a lateral measurement method with nanometer resolution is proposed. Performance tests show that the probe based on this method can achieve a lateral resolution of 2 nm in the X- and Y-directions, with repeatability of 15 and 20 nm, respectively. In microstructure measurement experiments, high-precision measurement for microslit width and edge angle on single crystal silicon wafers is achieved in the developed experimental system.

Evaluating measurement uncertainty is crucial for ensuring the reliability of piezoelectric drive systems. However, existing international standards are insufficient for dynamic measurement uncertainty evaluation, primarily due to the complexity of dynamic systems and the challenges of establishing uncertainty propagation models with limited samples. To address this issue, we propose a temporal evidential regression network (T-ENet) for developing dynamic models and evaluating uncertainty in piezoelectric impulse drive systems with small-sample nonlinear characteristics. We combine an evidential regression model with gated recurrent units (GRUs) to create a robust modeling framework. This framework integrates gradient-updated meta-learning algorithms, allowing it to perform effectively with minimal training data and gradient updates, accurately capturing the temporal features of dynamic systems and estimating and predicting the distribution parameters of system dynamic uncertainty. Experimental results validate the effectiveness of our method, and comparisons with traditional long short-term memory (LSTM) and GRU networks demonstrate its superiority in dynamic prediction. The correlation between prediction uncertainty and actual error confirms the effectiveness of our method in estimating uncertainty in dynamic measurements and provides a key reference for analyzing the reliability of actual measurement results.

In multi-dimensional nanopositioning and nanomeasuring devices, interference measurement is widely used. The three-dimensional (3D) target mirror serves as the spatial reference plane to achieve multidimensional interference measurements. However, the surface shape errors of the target mirror are superimposed on the geometric dimensions of the measured workpiece, which limits the system’s overall measurement accuracy. This paper proposes a method for on-machine separation and compensation of the target mirror’s surface shape errors based on the micro–nano-coordinate measuring machine (MNCMM) that employs interference measurement. This method provides the model for the separation and compensation of the surface shape errors. The MNCMM employs a home-made resonant probe and a reference flat crystal to achieve the separation experiment. Subsequently, an interpolation algorithm is used to compensate for the surface shape errors at any point in space according to the compensation model. By comparing the flatness measurement results of a standard flat crystal with a flatness of 50 nm before and after compensation, the flatness is reduced from 175 to 77 nm. It demonstrates the reliability of the method. This method can be widely applied to on-machine compensation for surface shape errors in multidimensional interference measurement systems.