Researches On Real-time Water Dynamics Simulation Modeling And Spatial Coefficient’s Self-adaptive Correction

In this era of technological development,the computer simulation technology on the real world has become more and more important for resolving the practical engineering issues and scientific issues.With the global environment changing,the devastating events about water have occurred in nature,such as dam-break,urban flooding,flash floods and debris flow events.In order to avoid the harm of such events to the public’s life and property,the relevant practitioners are eager for a system that can predict the movement of water bodies.The core of water movement simulation is the water dynamic model.Although the researchers have made great achievements in the related fields for several decades,this paper puts forward a new idea and design for the water dynamic modeling from different view,offering advice and suggestions for the relevant areas.In this paper,the Water Dynamic Simulation Modeling and Spatial Coefficient’s Self-adaptive Correction are studied theoretically,firstly.And then,using the GPGPU algorithm based on the Direct Compute,the technical realization of GPGPU algorithm is studied.Finally,the theoretical work and technical work are carried out through the experiments.The main work is as follows:1,in the first section,the background and significance of the thesis selection are expounded,including the summarization about the research status of water dynamic simulation model at home and abroad.And the research scheme of the thesis is made at the end of this section.2,a new fluid dynamics simulation model is proposed:The Short-lived Water Cuboid Particle(SWCP)model.Based on the core idea of grid-based method and particle-based method,the theoretical study of The SWCP model is carried out by combining two methods’ advantages.In the SWCP model,particles are used to describe the movement of water during one calculation cycle,which can make the continuous and the non-continuous water motion events described by a unified motion model;after a calculation cycle,the particles update the motion properties into the attribute grid,and the particle life ends.In order to achieve the simulation requirements for the whole computational domain with fewer particles,the paper uses the idea of vertical integral used in the shallow water equation to reduce the amount of data to simulate the water movement.The main contents include how the particle attributes are transformed into the attribute grid,the derivation of the particle motion equation,the modeling of the particle motion consistency,and so on.3,the spatial coefficient’s self-adaptive correction model is proposed and studied in the water dynamic simulation model.Based on the study of the spatial coefficient of the water dynamic simulation model,it is found that there is little research on the self-adaptive correction on the spatial coefficient.A large number of scientific research and practice calculate the spatial coefficient through studying the terrain surface coverage type and the calculation formula.This type of work can be effective for the assessment work,but the coefficients derived from this method are not suitable for the simulation system generally,and they have to be further adjusted to achieve a good simulation results in the simulation system.Therefore,this paper directly studies the coefficient from the point of view of the simulation system,utilizing the self-adaptive correction system to achieve that the simulation results consistent with the real values.It not only improves the efficiency to make simulation system match the real situation for the target simulation area,but also provides a method to obtain the spatial distribution of spatial coefficients.The main contents of this part include the basic idea of adaptive control theory,the iterative formula of spatial coefficient’s self-adaptive correction and the adaptive correction logic based on the characteristics of water movement.4,the parallel algorithm based on Direct Compute is designed for the SWCP model and coefficient’s self-adaptive correction model.In order to make the SWCP model and self-adaptive correction model could directly simulate the water motion in real-time according to the initial data such as the terrain data and the initial state data of the water body,all simulation calculations need to be calculated efficiently for the interpretation of the water movement.Therefore,it is necessary to use the GPGPU parallel calculation with reasonable design so that the efficiency of each calculation cycle could meet the requirements of real-time display.In order to improve the efficiency of the calculation,firstly we must find the time-consuming stages\modules in the simulation system:speed calculation,CFL condition calculation,and control points’ cost distance calculation.During the speed calculation stage,if putting the speed calculation and attribute integration together,there will be more repetitivecalculation,which is not conducive to upgrade the efficiency.Therefore,We divide the speed calculation and attribute integration into two-step calculation,using the idea of space for time to improve simulation efficiency which makes the system uses more memory to reduce repetitive calculation.For the CFL condition calculation,we study how to design the GPU parallel computing for get maximum value through the reduction algorithm separately using the shared memory and the UAV resources.As control points’cost distance calculation is a cumulative process,which depends on the order logic,its parallel computing in the GPU is hard to meet the real-time requirements.However,the cost distance value is a static variable in the calculation process,so using the pre-calculation method,we can avoid real-time calculation,which could save calculation cost.Through the data organization and algorithm design above,the SWCP model and self-adaptive correction model can be used into a real-time simulation system.5,verification experiments are designed and the experimental results are analyzed.It mainly includes the verification of the SWCP model’s correctness,verification of spatial coefficient’s self-adaptive correction model and the efficiency comparison of the GPGPU parallel design.