Numerical Simulations Of The Lake Water Environment And Parallel Computations Of The Non-hydrostatic Flows

With the acceleration of urbanization,problems like lakes eutrophication have become more obvious,which bring serious threats to economic and social sustainable development of our country.Lake environmental mathematic models can accurately and rapidly predict the variation of flow,water temperature and water quality in lakes,thus providing scientific and technical support to carry out water environment treatment and water network connection engineering in lakes.Due to the theoretical shortcomings,traditional water environmental mathematic models and methods are hard to satisfy the research demand.The main problems include:(1)The hydrostatic hydrodynamic models can't simulate the flow movement in some discontinuous terrains;(2)Water temperature models perform poor simulations due to the ignorance of some important thermal effects;(3)Water quality models are unable to accurately simulate the process of pollutants migration and transformation;(4)obvious numerical oscillation caused by Low accuracy schemes;(5)low computational efficiency.To address these problems,the dissertation explored the evolution of lake hydro-ecological system.Based on fluid mechanics,eco-hydrology,computational mathematics and computer technique,the dissertation analyzed the mechanisms of flow movement,material transportation,energy exchange and pollutants migration and transformation.On this basis,lakes environmental mathematic models were established to make a deep research on some frontier issues in the field of lakes environment numerical simulation,such as non-hydrostatic flow,lake stratification,simulation of pollutants migration and transformation,high accuracy schemes and parallel computing method.The main work and innovations are as follows.(1)A 3D hydrodynamic model and a 3D water quality model which can reflect the flow and water quality changing processes were established.Based on N-S equations,the 3D hydrodynamic model considers factors such as complex terrain,density contrast,wind shear force,bottom friction and so on.A two equation turbulence model was used to close the governing equations.An internal-external mode splitting procedure was adopted to solve the governing equations.Based on materials transportation and reaction kinetics principle,the 3D water quality model reveals the migration and transformation mechanism of nitrogen,phosphorus and oxygen quantitatively.The established 3D models were used for simulation of lakes of Dadong Lake Basin in Wuhan city.The Results showed that,the proposed models could effectively reveal the changing process of wind driven current and water quality,thus providing an efficient method to lake pollution prevention and water network connection engineering.(2)To overcome the shortage of simulation distortion of water temperature caused by the ignorance of wind sheltering effect,an Urban Lake Temperature Model which was based on a wind sheltering model was proposed.Based on the relationship between wind directions,the location of the obstacle,and the characteristics of wind variation downwind from the obstacle,the wind sheltering model could calculate a time-dependent wind sheltering coefficient to dynamically adjust the input wind data,thus improving the calculation accuracy of heat flux.The results of Guozheng Lake showed that the proposed water temperature model gained a more accurate simulation of lake temperature distributions in the vertical direction.The model could effectively analyze the influences of wind and wind sheltering on vertivcal temperature distribution in lakes.What's more,the wind sheltering model can also be used to develop the theoretical research system of wind driven current.(3)To overcome the shortage of the numerical oscillation caused by low accuracy schemes,a splitting method by using a high-order scheme was proposed.The method divides the all the operators of the water temperature model to the source sink term operator and the diffusion term operator.Using the undetermined coefficient method,a high-order difference scheme was designed to discretize the source sink term;the diffusion term was discretized using a scheme which has second-order accuracy in time and space.The effectiveness and capability of the splitting method was assessed by performing numerical tests.The test results showed the proposed method could decrease the truncation errors,thus improving the computational accuracy of water temperature model.(4)To solve the problem that the hydrostatic hydrodynamic model can't accurately simulate flows in some discontinuous terrains,a non-hydrostatic hydrodynamic model was proposed.Based on N-S equations,the total pressure was divided into dynamic pressure and static pressure,the vertical momentum equation was approximated by the Keller-box scheme.A predictor corrector method was proposed to solve the model equations to obtain the pressure and velocity distributions.The results of the solitary and sinusoidal experiments showed that,the model was able to strictly meet the zero dynamic pressure boundary conditions at the free water surface and accurately reflect variation of the non-hydrostatic flow.(5)To solve the problem of the low computational efficiency of the non-hydrostatic hydrodynamic model,a multithreading parallel computing method was proposed.Based on the parallel strategy of block calculation,the original computational domain was divided into several subdomains that were physically connected via a virtual boundary technique.Several sub-threads were created and tasked with the computation of the subdomains.A threading communication method by using the thread lock technique was developed to achieve synchronous communication between sub-threads.The results showed that,the proposed method could increase the computational efficiency of the non-hydrostatic model effectively,thus promoting the applications of non-hydrostatic model to research on lake hydrodynamic.