基于PAR的算法形式化开发

形式化方法是构建可信软件的重要途径.基于对算法问题的分析,针对形式化方法PAR开发算法的特征,刻划了问题分划、递推关系构造方面的规律.从一类问题的形式化功能规约出发,可机械地完成问题的分划及规约的变换,自然地揭示出求解问题的算法思想,在相关工具的支持下自动生成算法程序.研究结果将算法设计中尽可能多的创造性劳动转化为非创造性劳动,降低了形式化求解算法问题的难度,提高了算法程序的可靠性和形式化开发效率.

PAR方法和循环不变式的范畴语义

范畴论对理解程序规约及程序设计和正确性证明十分有用.PAR方法则是建立在严格的数学基础之上的一种统一的算法程序设计方法.循环不变式在循环算法程序的设计中至关重要.使用格理论和范畴论作为工具对PAR方法建立一个理论框架,并对其用范畴论的概念加以解释,从而使得PAR有更强的理论基础.在此基础上引入不动点原理深入刻划循环不变式的含义,循环不变式可以表示为谓词泛函的最小不动点,并从范畴论的角度解释该过程.

Effects of convection and solid wall on the diffusion in microscale convection flows

The diffusive transport properties in microscale convection flows are studied by using the direct simulation Monte Carlo method. The effective diffusion coefficient D is computed from the mean square displacements of simulated molecules based on the Einstein diffusion equation D = x2 t /2t. Two typical convection flows, namely, thermal creep convection and Rayleigh– Bénard convection, are investigated. The thermal creep convection in our simulation is in the noncontinuum regime, with the characteristic scale of the vortex varying from 1 to 100 molecular mean free paths. The diffusion is shown to be enhanced only when the vortex scale exceeds a certain critical value, while the diffusion is reduced when the vortex scale is less than the critical value. The reason for phenomenon of diffusion reduction in the noncontinuum regime is that the reduction effect due to solid wall is dominant while the enhancement effect due to convection is negligible. A molecule will lose its memory of macroscopic velocity when it collides with the walls, and thus molecules are hard to diffuse away if they are confined between very close walls. The Rayleigh– Bénard convection in our simulation is in the continuum regime, with the characteristic length of 1000 molecular mean free paths. Under such condition, the effect of solid wall on diffusion is negligible. The diffusion enhancement due to convection is shown to scale as the square root of the Péclet number in the steady convection regime, which is in agreement with previous theoretical and experimental results. In the oscillation convection regime, the diffusion is more strongly enhanced because the molecules can easily advect from one roll to its neighbor due to an oscillation mechanism. © 2010 American Institute of Physics. doi:10.1063/1.3528310

Numerical study on NOx/CO emissions in the diffusion flames of high-temperature off-gas of steelmaking converter

The combustion of high-temperature off-gas of steelmaking converter with periodical change of temperature and CO concentration always leads to CO and NOx over-standard emissions. In the paper, high-temperature off-gas combustion is simulated by adopting counterflow diffusion flame model, and some influencing factors of CO and NOx emissions are investigated by adopting a detailed chemistry GRI 3.0 mechanism. The emission index of NOx (EINOx) decreases 1.7–4.6% when air stoichiometric ratio (SR) increase from 0.6 to 1.4, and it dramatically increases with off-gas temperature at a given SR when the off-gas temperature is above 1500 K. High-concentration CO in off-gas can result in high NOx emissions, and NOx levels increase dramatically with CO concentration when off-gas temperature is above 1700 K. Both SR and off-gas temperature are important for the increase of CO burnout index (BICO) when SR is less than 1.0, but BICO increase about 1% when off-gas temperature increases from 1100 K to 1900 K at SR > 1.0. BICO increases with CO concentration in off-gas, and the influence of off-gas temperature on BICO is marginal. BICO increases with the relative humidity (RH) in air supplied, but it increases about 0.5% when RH is larger than 30%.