Measurement最新文献

Azimuthal mode identification in aero-engines is essential for understanding the mechanism of fan/compressor noise and rotating instability (RI). The compressed sensing (CS) methods enable the identification of azimuthal mode signals at high order with a limited number of sensors, breaking the Nyquist-Shannon sampling theorem. However, the CS method based on ℓ₁ regularization introduces the problem of amplitude underestimation. For the accuracy aim, ℓ_1/2 regularization is an effective substitution despite the difficulty of solving due to its non-convexity. This paper proposes a Dual Iterative Half Thresholding (D-Half) algorithm to solve the ℓ_1/2 regularization problem with high efficiency and accuracy. In the inner iterations, parameter k is introduced to regulate the sparsity of solutions, referred from the Iterative Half Thresholding (Half) algorithm. The outer iterations of k serve to provide an appropriate initialization for each Half problem. The dual iteration structure helps achieve higher reliability and lower dependence on prior information. The applicability of the D-Half algorithm for signal identification with single and multiple dominant modes is validated by a series of numerical experiments and actual tests on a compressor rig. Both results indicate that the D-Half algorithm effectively mitigates the prevalent underestimation of dominant modes, enabling azimuthal mode identification with high accuracy. Furthermore, the D-Half algorithm surpasses the classical CS methods in noise immunity, robustness to sensor failures, and time efficiency.

Fire prevention in industrial settings are paramount for ensuring human safety and economic stability. However, current mainstream DETR detectors face significant challenges in applications due to the necessity for extensive memory accesses and inference delays. To tackle these complexities, we propose an Industrial Fire Smoke Detector based on an end-to-end structured framework. In a series of innovative optimizations, we firstly adopt a more lightweight backbone network, LeanNet, for feature extraction. Combined with the optimized Transformer architecture, this approach enhances the model’s detection speed to address real-time challenges effectively. Secondly, we introduce a Feature Fusion Network based on alignment mechanisms to enhance the DETR model’s multi-scale object representation capabilities without significantly increasing latency. Subsequently, to facilitate easier training and optimization of IFS-DETR, we introduce IoU-aware query selection and an aspect ratio-based denoising training strategy, and enhance the localization loss function using Inner-SIoU. Finally, we deploy IFS-DETR on NVIDIA Jetson Orin Nano. The dataset is available at https://github.com/Sonnenb1ume/IFS-DETR.

This paper explores the application of a liquid circular angular accelerometer (LCAA) in incipient bearing fault diagnosis. First, a wireless instantaneous angular acceleration (IAA) signal acquisition system is designed to collect motor IAA under various bearing fault conditions. Then, the IAA characteristics of the motor with both healthy bearings and incipient bearing faults are analyzed, which provides valuable insights into fault diagnosis method design. The proposed method implements an advanced signal preprocessing technique, which is developed based on self-adaptive noise cancellation (separates discrete frequency noises), minimum entropy deconvolution (enhances the fault-related components), and a novel approach of sliding time-window analysis to improve reliability. Hereafter, IAA-based estimated fault characteristic frequencies are identified in the envelope spectra of the post-processed data, which finalizes the bearing fault diagnosis. Simulation and experimental results substantiate the effectiveness of the proposed approach for early fault detection, even under the conditions of low sampling rates.

Thermal transport characterization of thermal interface materials (TIMs), especially in their application scenario is so far challenging. Herein, we demonstrate a low-frequency frequency domain thermoreflectance technique to measure the semi-in-situ thermal transport properties of TIMs in a sandwiched structure. The measurement ability is first demonstrated by accurate measurement on several well-known materials. Then the thermal conductivity and interface thermal resistance of a thermal gel, a thermal pad composed of phase change materials and an Indium foil, sandwiched in two silicon slides are investigated. Sensitivity analysis emphasized the importance of pump laser spot size and modulation frequency in the measurement. For TIMs with low thermal conductivity, the sensitivity to interface thermal resistance is highly suppressed, while for TIMs with high thermal conductivity, the sensitivity to interface thermal resistance is prominent, which makes it easier to separately determine the TIM thermal conductivity and interface thermal resistance.

Accurately reconstructing the unknown dynamic loading is crucial for controlling the vibration of mechanical systems. Many methods have been proposed for load identification of time-invariant systems, but in some mechanical systems, their structural parameters exhibit time-varying dynamic characteristics during their operation. Therefore, tracking the time-varying structural parameters to update the transfer function is necessary, which is the most important prerequisite for load reconstruction. However, some existing methods are unstable because of the nonlinear conditions. To overcome this issue, an improved dual-nested Kalman filter framework has been proposed. This method allocates unknown variables to two Kalman filters, allowing them to iterate and loop with each other. As a result, it achieves linear reconstruction of the dynamic loading for a time-varying system. The proposed method has been validated in two numerical examples: one is a spring-mass system with abruptly changing mass parameters, and the other is a milling system with slowly changing mass parameters. The results show that the proposed method can reconstruct loading stably and is not susceptible to noise interference.