Qixing Zhang

Molten thermoplastic dripping behavior induced by flame spread over wire insulation under overload currents

The dripping behavior of the molten thermoplastic insulation of copper wire, induced by flame spread under overload currents, was investigated for a better understanding of energized electrical wire fires. Three types of sample wire, with the same polyethylene insulation thickness and different core diameters, were used in this study. First, overload current effects on the transient one-dimensional wire temperature profile were predicted using simplified theoretical analysis; the heating process and equilibrium temperature were obtained. Second, experiments on the melting characteristics were conducted in a laboratory environment, including drop formation and frequency, falling speed, and combustion on the steel base. Third, a relationship between molten mass loss and volume variation was proposed to evaluate the dripping time and frequency. A strong current was a prerequisite for the wire dripping behavior and the averaged dripping frequency was found to be proportional to the square of the current based on the theoretical and experimental results. Finally, the influence of dripping behavior on the flame propagation along the energized electrical wire was discussed. The flame width, bright flame height and flame spreading velocity presented different behaviors.

Deep domain adaptation based video smoke detection using synthetic smoke images

In this paper, a deep domain adaptation based method for video smoke detection is proposed to extract a powerful feature representation of smoke. Due to the smoke image samples limited in scale and diversity for deep CNN training, we systematically produced adequate synthetic smoke images with a wide variation in the smoke shape, background and lighting conditions. Considering that the appearance gap (dataset bias) between synthetic and real smoke images degrades significantly the performance of the trained model on the test set composed fully of real images, we build deep architectures based on domain adaptation to confuse the distributions of features extracted from synthetic and real smoke images. This approach expands the domain-invariant feature space for smoke image samples. With their approximate feature distribution separated from non-smoke images, the recognition rate of the trained model is improved significantly compared with the model trained directly on mixed dataset of synthetic and real images. Experimentally, several deep architectures with different design choices are applied to the smoke detector. The ultimate framework can get a satisfactory result on the test set. We believe that our method own strong robustness and may offer a new way for the video smoke detection.

A polarization-modulated multichannel Mueller-matrix scatterometer for smoke particle characterization

The polarization properties of scattered light are being exploited to determine the optical and physical information of small particles. In this paper, a scatterometer is developed for simultaneously measuring the Mueller scattering matrix elements as functions of the scattering angle. The scatterometer uses an electro-optic modulator to modulate the polarization state of the incident light, and uses two photomultipliers provided with different polarization optics to consist multichannel polarization-state detector. The instrument takes advantage of combination of the polarizationmodulation technique and division-of -amplitude photopolarimeter, which make for a compact design and substantial increase in measurement throughput and speed. The methods of calibration and alignment using the polarizationmodulated light are established, with which the instrument is calibrated precisely. The methods of data processing and error analysis of the measured Mueller matrix elements are developed. A hybrid experimental/theoretical approach to study the light scattering properties of smoke particles is also presented.

Sensitivity analysis of morphology on radiative properties of soot aerosols

Absorption cross section (Cabs), scattering cross section (Csca) and asymmetry parameter (ASY) of soot particles in different atmospheric aging status were investigated under fixed equivalent volume radius (RV) using the numerically exact multiple-sphere T-matrix method. The radiative properties of soot particles would be largely diverse in different aging status even RV is fixed. However, there are many insensitive parameters under different aging status. The Cabs and ASY is insensitive to monomers number (Ns) when Ns is larger than a threshold value. For bare and thinly coated soot aggregates, Cabs is insensitive to fractal dimension (Df) when the RV is small, where the relative errors of Cabs for different Df are within 2.5%. However, the effects of Df is obvious for large soot due to the shielding effects of large monomers, and the relative errors for different Df can reach to 18% for bare soot. For thinly coated soot, the changes of ASY with soot volume fraction (fsoot) is small due to the little changes of the fractal structure when the RV is fixed. In addition, for thickly coated soot, ASY is insensitive to Ns due to the unchanged overall spherical structure. Our results give a further understanding of the influences of morphology on radiative properties. It may be helpful for model selection and model simplification.

Scattering Matrix for Typical Urban Anthropogenic Origin Cement Dust and Discrimination of Representative Atmospheric Particulates

The complete scattering matrix for cement dust was measured as a function of scattering angle from 5° to 160° at a wavelength of 532 nm, as a representative of mineral dust of anthropogenic origin in urban areas. Other related characteristics of cement dust, such as particle size distribution, chemical composition, refractive index and micro morphology were also analyzed. For this objective, a newly improved apparatus was built and calibrated using water droplets. Measurements of water droplets were in good agreement with Lorenz-Mie calculations. To facilitate the direct applicability of measurements for cement dust in radiative transfer calculation, the synthetic scattering matrix was computed and defined over the full scattering angle range from 0° to 180°. The scattering matrices for cement dust and typical natural mineral dusts were found to be similar in trends and angular behaviors. Angular distributions of all matrix elements were confined to rather limited domains. To promote the application of light scattering matrix in atmospheric observation and remote sensing, discrimination methods for various atmospheric particulates (cement dust, soot, smolder smoke, and water droplets) based on the angular distributions of their scattering matrix elements are discussed. The ratio -F12/F11 proved to be the most effective discrimination method when a single matrix element is employed, aerosol identification can be achieved based on -F12/F11 values at 90° and 160°. Meanwhile, the combinations of -F12/F11 with F22/F11 (or (F11-F22)/(F11+F22)) or -F12/F11 with F44/F11 at 160° can be used when multiple matrix elements at the same scattering angle are selected.