Research On The Wind Fields Of Different Height Levels In A Ventilated Room Based On Numerical Simulation

Due to the fact that industrial hazards frequently occur in urban areas, the moresophisticated investigation is needed to realize the rapid and precise detection ofpoisonous gas leak source, which is benefit to the development of economy andsociety. Considering that the toxic environment is not fit for human or animals todetect on-site, a few researchers attempted to locate the odor source using a mobilerobot in the early1990s, which is called robot-based gas leak source detection.In recent decades, although researchers have spared no effort to solve theproblem, coming up with a series of tracking strategies however, there is notsatisfying breakthrough. Reason is that there are still some key problems remainingunsolved: To begin with, there is no high simulation plume model. The numericalsimulation environment for testing the gas leak source localization algorithmsoutperforms the real plume environment for the reasons that it costs less and we caneasily control the parameters of the flow. In addition, the same simulationenvironment can be used many times to compare the performance of differentsearching algorithms. However, there is no universally accepted simulation plumewhich can perfectly match the real plume. Therefore the credibility of test results inthe simulation environment needs to be improved. Moreover, in existing plumetracking algorithms most parameters, such as time-step, angle with the wind direction,are fixed, owing to the fact that we know little about the change pattern of the airflowwhich can affect the diffusion of the gas molecular. Apparently, such a fixed settingcan’t adapt to the variation of environment.In response to those issues, this thesis presents a study of the following twoaspects:Firstly, in order to learn more about the regular pattern of the three-dimensionalflow field, a tool of computation fluid dynamics (CFD) software, FLUENT, is appliedto numerically simulate airflow in the ventilated room. Based on the simulation data aresearch of the variation of the wind in different height levels is carried out using thenonlinear implement Hilbert-Huang Transform. This study can provide more clues forcomparing the two kinds of test environment, which can help to establish morerealistic simulation plume environment. Meanwhile, this study can help to decide onwhich height to fix the wind sensor. Secondly, in order to enhance the environmental adaptability of trackingalgorithms, it is necessary to learn more about the differences between different areasof the wind field at typical heights. However, images representing the wind fieldgenerated by FLUENT are too complex to use mature algorithms to analyze.Therefore, a new concept of wind fingerprint, including wind direction fingerprint andwind velocity fingerprint, is presented in this paper. By extracting and analyzefeatures of wind fingerprint, we research the change intensity of wind direction andwind velocity in different areas, providing clues for further improvement of trackingalgorithms.