Research On Data-driven3D Accident Simulation Technology

In the past decades, the rapid development of petrochemical industry and the swiftexpansion of production scale have led to a higher level of destructiveness ofpetroleum-related chemical accidents during the process of the production, reserve andtransportation with their specific nature of inflammability and explosion hazard. Theseaccidents take the forms of explosion, fire and toxic gas leakage, etc. taking a heavytoll of people’s lives and property. They generally have a feature of unexpectedness,swift diffusion, rescue hardship and mass destruction. Therefore, the research on theregular course of the diffusion and flow of these accidents will shed light on themanipulation and lessening the hazards. The research will be conducted from theperspectives of the model construction of atmospheric dispersion, process simulationof flaming and rescue simulation of fire artillery’s jetting trajectory.This thesis, on the basis of the above theoretical models, will attempt to apply theself-developed platform i4Factory by Marine Information Technology Laboratory ofInformation Science and Engineering (MITLOISE), Ocean University of China (OUC)to the research of simulative data-driven3D accident visualization. New methods,through the construction of fire drilling and planning models, hence, will be put forwardby the simulation of the atmospheric diffusion of toxic substances, that of fire ongoingand fire artillery’s jetting trajectory. According to the characteristics of the physicalmodel, the research will be implemented from the following three aspects:1. A brief introduction to the development of atmospheric dispersion model, and adescription of the Computation Fluid Dynamics, SIMPLE algorithm and the sources ofsimulative data applied in the research. Analyzing data structure of simulation data. Inaccordance with the simulation process, specifications will be given to the dataprocessing, visualizing and data offloading. The final visual results will be shown inpictures.2. The research will then be conducted according to the classification of chemicalaccident types (fire, explosion, toxic gas leakage and smoke smudge). The technical means of particles systems will be used to the type simulation of accidents performanceon the basis of visual patterns. Take smoke effect for an example, a specification willbe offered to the designing process of particle effect with the technical assistance ofMagic Particles software and the particle effects caused by the parameter adjusting. Theparticle effects will be integrated into i4Factory platform under the unifyingmanipulation.3D sound effects and global air control will finally be attached to increasethe sense of reality.3. A comparative analysis of the current fire jet trajectory simulation models will bemade to arrive at a proper physical model available for the research according to theparameter types and implementation methods. By setting a series of parameters, jettingtrajectory simulation data will be accumulated to make a further fitting of jettingtrajectories according to the performance patterns of particle system. The ultimatesimulative data-driven jetting trajectory model will intend to offer an alternative methodto the simulative manipulation of firefighting and rescue in the simulative3D settings.Through the study of the above three modules, the combination of virtual realitytechnology and physical model-generated simulation data attaches a physical meaningto the virtual effects in the virtual reality. By imitating the virtual reality scenes andapplying fire plans and drills, this research will provide a new approach to theprevention and mitigation of disasters in the chemical industry.