基于数值模拟仿真优化技术的冒口设计新方法研究

发布时间：2018-09-07 11:41

【摘要】：冒口设计常用方法主要有模数法、比例法、三次方程法等。以数值模拟技术为基础,结合仿真优化方法是一种更加科学的冒口设计方式,并且是未来冒口的主流设计方法。以冒口为研究对象,以数值模拟软件为基础,结合优化算法进行了冒口设计新方法的研究。这种方法是以金属凝固理论、铸件形成理论、传热学、流体力学等学科为基础,科学的铸造工艺设计方法,研究处于初始阶段。研究结果能够为冒口自动化与集成化的新方法研究工作打下一定基础,摆脱冒口设计对经验的依赖。对仿真优化方法在冒口设计中的应用进行了研究。将模拟用于策略验证的优化方法用于冒口设计。对冒口的候选策略逐一输入仿真模型,采用枚举的方法来对比每个策略的效果,完成冒口的设计。这种设计方法将仿真技术和优化技术结合在一起,具有简单易行快速得到优化结果的特点,适用于规律性问题的研究。在冒口形状优化的研究中应用了这种方法。针对冒口优化设计中多参数优化效率低下、优化过程复杂的问题,对冒口形状优化进行了基础性的研究工作,确定具有最佳补缩效果的冒口形状。详细地给出了冒口高度、直径对补缩效果的影响,将冒口的多参数优化问题转化为以直径为参考的单变量函数优化问题。对于冒口形状优化,研究了相同模数、不同形状冒口的补缩能力。在冒口刚好完成对铸件补缩的情况下,相同模数的球形冒口的补缩能力比等边圆柱和正方体的补缩能力低。结果表明模数相同、由于形状因素对冒口体积的影响,使冒口的补缩能力有很大的不同。研究了不同形状冒口的补缩能力。分别用球形、等边圆柱和正方体形状冒口对相同铸件进行补缩,安全高度为零,通过模拟用于策略验证的优化设计方法,得到球形、等边圆柱、正方体的体积比为:1:1.08:1.0825。球形冒口的补缩能力最好,正方体形状最差,但与等边圆柱的差别并不大。说明在具有相同公切球的情况下,不同形状对冒口的补缩能力影响区别不大。但是球形冒口在制作工艺上比较麻烦,所以,生产中常用圆柱形冒口进行补缩。针对生产中常用的圆柱形冒口,研究了冒口高度对补缩能力的影响。当冒口的高径比K≤0.7时,随着冒口高度的增加,缩孔向上移动的速度逐渐降低。当K1后,由于冒口顶部距离铸件距离变大,冒口顶部对铸件温度场的影响变小,所以增加冒口高度,基本不能提高冒口的补缩能力。研究了冒口直径对补缩能力的影响。在K≤1范围内,随着冒口直径的增大,缩孔高度逐渐升高,增加冒口直径比增加高度效果要好。在0.7K≤1范围内,可以通过提高冒口高度或增大冒口直径来对冒口的补缩效果进行微调。研究了具有最佳补缩形状的冒口高径比。在冒口体积相同情况下,比较不同K值的冒口补缩效率。当冒口高径比0.4≤K≤0.7时,冒口具有很好的补缩效率,当K=0.7时,冒口的补缩效果最好。取K=0.7为冒口的最佳高径比,将冒口的多参数优化转化为以直径为参考的单变量函数优化问题。将求冒口最小体积的优化问题转变为求冒口最小直径的一维搜索问题后,冒口优化设计过程变得简单且设计效率明显提高。对单变量函数的优化问题进行了研究。从单变量函数优化的基本思想、搜索区间的确定、优化问题的解决方法等几方面对单变量冒口优化问题进行了研究。将二分法和爬山算法相结合,提出了“贪吃蛇算法”,并详细给出了算法的求解数学模型优化算法具有收敛速度快、计算精度高、结果稳定的特点。随着优化次数增加,精度能够进一步提高,但是优化效率明显降低,所以优化过程中要选择合适的优化精度以提高优化效率、减少优化次数。采用这种方法,进行了冒口优化设计对比实验,优化过程从原来的110次减少到4次。结果表明将冒口的多参数优化问题转变为单变量函数优化后,冒口优化设计过程变得简单且设计效率明显提高。经过分析后,找到了贪吃蛇算法存在的缺陷:搜索范围为初始值加上2倍步长。如果初始值过小,当初始值小于目标值的1/2时,则导致不能得到优化结果。为此,对冒口的优化模型进行修正,提出了向上翻倍长大算法,并将长大算法和优化算法结合,建立了冒口“生长”数学模型,实现了冒口的自动化设计。这是一种全新的冒口设计方法,将冒口设计从静态设计变成动态生长过程。在铸件需要冒口的位置,放上冒口“种子”,在优化迭代的过程中实现冒口的自然生长,最后“长”到所需要的尺寸,完成冒口的设计。根据冒口“生长”模型给出实例验证。这些研究为铸造工艺设计的集成化和智能化打下了很好的基础,说明了将数值模拟与仿真优化技术相结合进行冒口设计的新方法,具有强大的技术优势和非常广阔的应用前景。
[Abstract]:The common methods of riser design are modulus method, proportion method, cubic equation method, etc. Based on numerical simulation technology, combined with simulation optimization method is a more scientific riser design method, and is the mainstream design method of the future riser. This method is based on metal solidification theory, casting formation theory, heat transfer theory, fluid mechanics and other disciplines, scientific casting process design method, research is at the initial stage. The research results can lay a certain foundation for the new method research of riser automation and integration, get rid of riser design to warp. The application of simulation optimization method in riser design is studied. The optimization method for strategy validation is applied to riser design. The candidate riser strategies are input into the simulation model one by one, and the effect of each strategy is compared by enumeration method to complete riser design. This method is applied to the study of riser shape optimization. Aiming at the problems of low efficiency and complex optimization process of multi-parameter optimization in riser shape optimization design, the basic research on riser shape optimization is carried out. The influence of riser height and diameter on feeding effect is given in detail. The multi-parameter optimization problem of riser is transformed into a single variable function optimization problem with diameter as reference. For riser shape optimization, the feeding capacity of risers with the same modulus and different shapes is studied. The feeding capacity of spherical risers with the same modulus is lower than that of equilateral cylinders and cubes when the feeding of castings is just completed. By simulating the optimal design method for strategy validation, the spherical and equilateral cylinders are obtained. The volume ratio of the cube is 1:1.08:1.0825. The spherical riser has the best feeding capacity and the cube has the worst feeding capacity, but the difference between the spherical riser and the equilateral cylinder is not great. In the case of the same common-cut sphere, the different shapes have little effect on the feeding capacity of the riser. But the spherical riser is troublesome in the manufacturing process, so the cylindrical riser is often used in production. When the riser height is less than 0.7, the upward movement speed of the shrinkage hole decreases gradually with the increase of the riser height. When the distance between the riser top and the castings increases, the influence of the riser top on the temperature field of the castings decreases, so increasing the riser height can hardly improve the feeding capacity of the riser. In the circumference, with the increase of the riser diameter, the shrinkage height increases gradually, and the effect of increasing the riser diameter ratio is better than that of increasing the riser height. When the riser height-diameter ratio is 0.4 < K < 0.7, the riser has a good feeding efficiency. When K = 0.7, the feeding effect of the riser is the best. The optimization problem is transformed into a one-dimensional optimization problem for finding the minimum diameter of riser, and the optimization design process becomes simple and the design efficiency is improved obviously. Combining dichotomy with hill-climbing algorithm, a "snake-eating algorithm" is proposed, and the mathematical model optimization algorithm of the algorithm is given in detail, which has the characteristics of fast convergence, high precision and stable results. In the optimization process, the appropriate optimization precision should be selected to improve the optimization efficiency and reduce the number of optimization times. By using this method, the comparative experiment of riser optimization design is carried out, and the optimization process is reduced from 110 times to 4 times. After analysis, the shortcomings of the algorithm are found: the search range is the initial value plus two times the step size. If the initial value is too small, and the initial value is less than 1/2 of the target value, the optimization results can not be obtained. Therefore, the optimization model of the riser is modified and the upside doubling length is proposed. A new riser design method is proposed, which changes riser design from static design to dynamic growth process. In castings where risers are needed, a riser "seed" is placed and the iterative process is optimized. The natural growth of riser is realized, and finally the riser is "long" to the required size, and the riser design is completed. An example is given to verify the riser growth model. The method has strong technical advantages and very broad application prospects.
【学位授予单位】：沈阳工业大学
【学位级别】：博士
【学位授予年份】：2015
【分类号】：TG244.4

【相似文献】