Measurement最新文献

Online condition monitoring methodologies for rotor dynamic systems are an emerging trend and an essential for safe system operation. During operation fatigue cracks in rotor shaft can lead to catastrophic failures if not identified at early stage. Therefore, these systems require a reliable crack fault detection methodology. The vibrations may induce in the system due to different type of faults, like shaft transverse cracks, oil whirl and whip. To make the detection methodologies robust and reliable, redundancy in crack detection methodologies is vital. The present research is a foolproof procedure to detect the travers fatigue crack in the rotor shaft. Rotor response, orbital pattern, and response FFT during transient (runup/cost down) and study state operations at 2X & 3X sub critical speeds are analyzed to identify the presence of crack in rotor bearing system. The rotor system is analyzed by numerical Finite Element Modeling (FEM) simulations and experimental means. Open Wedge Crack (OWC) and Breathing Wedge Crack (BWC) are modeled and their behavior is studied. The effectiveness of Newton Rapson method, Houbolt numerical method, and Harmonic Balance Method (HBM) in crack diagnosis are studied. Numerical FEM results are validated with experimental and published literature results. FEM simulated and experimental response FFT shows similar response behavior with presence of crack with 98 % correlation and less than 2 % error. FEM simulated and literature response orbital pattern at 2X sub critical backward speed are similar with less than 3 % error (97.76 % close) and at 2X sub critical forward speed are similar with less than 1 % error (99.33 %).

Aiming the fault diagnosis problem in aviation high-speed bearings with complex operating conditions, most domain adaptation methods primarily focus on eliminating the differences from operating conditions by utilizing domain labels and class labels. However, these methods ignore the fine-grained information and the discriminative information among classes. Therefore, this paper constructs a novel self-supervised domain adversarial generalization framework, called deep subdomain adversarial network (DSAN). Specifically, a one-class diagnosis model is constructed for each subdomain with source domain supervision information to extract fault diagnosis rules. The pseudo-labeling is generated with self-supervised method by calculating the distance from the target domain data to each diagnosis rule, and to realize the aligning the subdomain among the target and source domains through local maximum mean discrepancy to realize the transfer of diagnosis rules. Test results on two bearing datasets indicate that our method has a better diagnosis performance and generalization.

ASTM F382 and ISO 9585 utilize Four-point Bending Test Method (FBTM) to evaluate the most vulnerable loading performance of bone plates. However, the pin boundaries in FBTM are unable to reflect practical physiological loading patterns for bone plates, and unstable, which makes the test result hard to be repeated. Considering the above, this study proposes a test method designed based on practical physiological loading patterns with stable boundaries. All key indexes required by ASTM F382 and ISO 9585 can be equivalently obtained through the proposed test method/platform. The validity of the proposed test method/platform has been verified through experimental study and Finite Element analysis through the comparison between FBTM and the proposed test method. The stable boundaries guarantee stable test results, in which the relative deviation ratio is 2.625% for the test curve, and the relative standard deviation is 1.613% for rotational bending stiffness and 6.034% for yield bending moment.

For exposed combustion environments, multi-view optical imaging method is commonly used for tomographic radiation thermometry. However, in the case of closed combustion with limited optical testable space, it is ineffective due to the light obstruction by burner wall. To overcome this problem, this paper proposes a monocular multi-focal imager equipped with tomography channel. There exists difference in defocus among different imaging channels. In theory, the combustion flame is segmented in the form of parallel slices along the main optical axis of imager. Based on the Fourier optics, the radiation information of slices at different spatial axial positions is extracted by establishing the mapping relationship between object and image. Then, the temperature of combustion slice is further calculated utilizing the Planck law. In experiment, the monocular multi-focal imager is designed and processed by optoelectromechanical integration. Multiple imaging channels with different defocus are operated with common optical path structure and synchronous trigging to match the transient three-dimensional combustion flame. The measured results show that the temperature distribution of flame can be effectively revealed, which demonstrates the potential applicability of combustion diagnosis.

The measurement accuracy of the turbine flowmeter is severely affected under vibration conditions. To investigate the impact mechanism, a DN10 turbine flowmeter is taken as the research object. An experimental facility is built to simulate the vibration conditions, and the performance of the turbine flowmeter under vibration conditions is obtained. The average meter factor of the turbine flowmeter increases with vibration amplitude and vibration frequency, and the maximum average meter factor under vibration conditions is 1.24 % higher than that under non-vibration conditions. A simulation method is proposed to simulate the working status of the turbine flowmeter under vibration conditions. Through the analysis of the flow field in the turbine flowmeter, it is found that the change in axial velocity distribution caused by vibration is the main reason for the change in the performance of the turbine flowmeter. The average axial velocity under vibration conditions is greater than that under non-vibration conditions, and increases with the increase of amplitude and frequency, resulting in the increase of impeller speed and the increase of meter factor.

In this paper, a mid-infrared thermal imager and a Complementary Metal Oxide Semiconductor (CMOS) imaging device were employed to measure the slagging thickness on the heating surface in a 330 MW subcritical drum boiler. A way to establish the relationship between the slagging temperature and emissivity was proposed based on the Boom emissivity model and the Hadley thermal conductivity model, and then the slagging thickness was calculated by the Newtonian iterative method. In the superheater area, the average slagging thickness of the lower section was 8.6 mm, and it was 1.71 mm higher than that of the upper section. In the water-cooled wall area, the average slagging thickness of the right wall was 10.8 % larger than that of the left wall. To verify the accuracy of the calculated slagging thickness, the slagging image was measured by the direct visual probe, and the relative error was 7.69 %.

Researchers and industry professionals often use surface metrology techniques to analyze and quantify these surface characteristics, enabling them to make informed decisions about machining strategies and process improvements. This paper presents a comprehensive exploration of surface characteristics of two different materials i.e., C45 steel and brass using electromagnetic measurement techniques. The research used techniques based on: Coherence Scanning Interferometry, Confocal, Focus Variation and Confocal Fusion. Additionally, different light colours were used for these techniques and measurement methods to assess their impact on the quality of mapping the electromagnetic surface. The results demonstrate that the regardless of the colour of the scanned surface (steel or brass), only approximately 50 % of the profile points achieved a 40 % match with the base profile points by scanning with blue light.

Iron powder is recognized as a promising clean energy carrier and a recyclable metal fuel due to its high energy density and carbon-free nature. Research focusing on single iron particles is essential for understanding the combustion characteristics and mechanisms of iron powder, as well as for designing reliable iron-fueled combustors. Temperature and particle size, as key parameters of burning iron particles, are typically measured using optical diagnostic methods. However, low-cost simultaneous measurement methods for the temperature and particle size of burning iron particles are scarce. This study introduces a method employing a single RGB camera to simultaneously measure the temperature and particle size of burning iron particles, validated by a heated thermocouple. Particle size is measured by the intensity of Blue channel based on light attenuation approach, and temperature is measured by the intensity ratio of Red/Green channels based on black-body radiation law. Statistical analyses of the temperature and particle size of iron particles at four different heights above the burner were conducted. Along the heights, the average particle size and temperature of the iron particles generally increase, except at a height of 50 mm, where more iron particles entering the cooling phase lead to a slight decrease in average temperature. The conditional probability density functions of temperature and particle size indicate a statistical dependence between the particle size and temperature of iron particles that larger particle sizes typically correspond to higher temperatures. The results confirm that the simultaneous measurement method is a cost-effective, high-accuracy tool for characterizing the combustion behavior of iron particles.

We present a novel method of particle-monolayer parametrization, based on least-squares fitting to the power spectral density of the monolayer image. An important advantage of this technique is avoiding the identification of individual particles. Using two images of random particle assemblies, found in the literature, we demonstrate the implementation of the method. Specifically, we determine the dimensionless particle radius and surface coverage, with their standard errors. For homogeneous monolayers of monodisperse particles, the errors are on the order of 1%. The errors increase with particle polydispersity and monolayer heterogeneity. We also show that the method is not sensitive to distortion in apparent particle size, frequently occurring in optical microscopy and AFM measurements.

Continuous and accurate monitoring of core body temperature (CBT) is crucial for personal healthcare, clinical disease diagnosis, and enhancing athletic performance. Current CBT measurement methods typically ignore contact interface variation at the sensor-body interface, resulting in insufficient CBT measurement accuracy under daily wearable scenarios with varying sensor contact pressure. This study proposed a core temperature estimation model with thermal contact resistance (TCR) compensation, introducing the TCR factor into the computational equation to improve estimation accuracy in daily wearable scenarios. A numerical simulation of the temperature field between the human forehead and sensor was developed for model performance evaluation. Given the inconvenience of TCR measurement, this study proposed a method based on pressure/optical photoplethysmography (PPG) signals for updating thermal contact resistance. The mean absolute error of CBTER model under varying thermal contact resistance conditions was 0.26 °C, which was 27 % compared to traditional model errors in simulation. Using pressure updating methods, the result of human trials was 0.01±0.23 °C (with 95 % limits of agreement). Overall, the proposed method effectively compensated for errors in core body temperature estimation due to thermal contact resistance, holding potential for continuous core body temperature monitoring to meet healthcare needs scenarios.