A New Method Of Riser Design Based On Numerical Simulation Optimization Technique

Common methods of riser designs include modulus method, the ratio method and cubic equation method, etc. On the basis of the numerical simulation technology, a riser design combined with the simulation and optimization is more scientific and tends to become a mainstream design method in the future. In this paper, when the riser was emphasized to be the object of the study, and the numerical simulation software was used, coupled with the optimization algorithm, a new method of riser design was investigated, a scientific method of casting process design which is based on the solidification theory, castings forming theory, heat transfer, fluid mechanics and so on and so forth. The research is now still in its initial stage. The investigation will lay the foundation for the automation and integration of riser designing, thus reducing too much dependence on the experiences of technicians.In this paper, the application of the simulation optimization method to the riser design was studied. The optimization method using the simulation to verify the strategy of riser design was studied. The candidate schemes of the riser was one by one imported into the simulation model and then the effect of each strategy was compared by the enumeration method to complete the riser design. The method combining the simulation technology with the optimization technology has features of getting its optimization results simply, easily and rapidly. It is suitable for the research on the regularity problem. The method was used in the study of riser shape optimization.Aiming at the problems of the lower efficiency of multi-parameter optimization and its complex optimization procedure, a fundamental research was conducted on the riser shape optimization as well as the best feeding ability of the riser. In this paper, the effects of riser’s parameters on the feeding ability, such as the riser height, riser diameter and so on, were given in detail, and the problem of multi-parameter optimization for risers became the problem of single variable(diameter) optimization.In relation to the shape optimization to risers, the feeding capacity for risers with same module but different shape was studied. When equilateral cylindrical risers exactly can feed the castings, with the same module, the feeding capacity of spherical risers is lower than that of equilateral columns’ and cubes’. The results show that with the same modulus the influence of the shape on the riser volume makes the riser feeding ability quite different.Different shapes’ feeding capacities of risers were studied in the paper. When the safety height is zero, feeding the same casting with spherical, equilateral cylindrical and square respectively by means of the use of simulation results on the strategy verification gains the radio of volume for spherical, equilateral cylindrical and square risers 1:1.08:1.0825. The feeding capacity of spherical risers is the best, while square risers’ is the worst with slight difference from equilateral cylindrical risers’. The results show that in the case of the same tangent, different shapes have little effects on their feeding capacities. However, the production process of spherical feeders is so complicated that cylindrical ones are widely used.Aiming at common cylindrical risers, the effects of risers’ height on their feeding capacities were studied. When K≤0.7(K is the ratio of the height to diameter of risers), with the increase of the height of a riser, the upward moving speed of shrinkage gradually reduced. When K>1, with the larger distance from the riser top to the casting, the influence on the casting temperature field became smaller. So the increasing height of risers can hardly improve their feeding capacities.The effects of risers’ diameters on their feeding capacities were also studied. When K≤1, with the growth of risers’ diameters, the height of their porosity gradually rose, so the effect of increasing diameter is better than that of increasing the height. When 0.7<K≤1,either the height or the diameter of a riser can be increased to improve its feeding effect slightly.The best feeding shape of risers was studied, too. When the riser volume was the same, the feeding efficiency with different K values(the ratio of the height to its diameter of risers) was compared,. When 0.4≤K≤0.7, the riser has a good feeding efficiency. When K=0.7, the feeding efficiency of risers is the highest. Taking the K=0.7 as the best high diameter ratio, the problem of multi parameter optimization can turn into the single optimization problem with the diameter. Riser design process becomes simple, and the design efficiency is improved obviously.The paper covers the problem of single variable function optimization from several aspects, such as the basic idea of single variable function optimization, determining the search range, and the solution to the optimization problem, etc. In this paper the algorithm of the dichotomy was combined with the hill climbing, the anakonda algorithm was presented, and the mathematical model of the algorithm is given in detail. The advantages of this algorithm include a fast convergence rate of the optimization algorithm, a stable result and a high precision. With the increase of optimal times, its precision can be further improved, but the efficiency of optimization significantly reduces. Therefore, an appropriate choice of optimal precision can improve the efficiency of optimization and reduce optimal times. The contrast experiments on riser optimization showed that the optimal process reduced 4 times from the original 110 times when this method was applied.After analysis, the defect of the anakonda algorithm was found that the search range is the initial value plus two times of step. If the initial value is too small(the initial value is less than half of the target value), the result of the optimization cannot be obtained. Therefore, the optimal model of risers was modified and the algorithm double grew up was put forward. In addition, a mathematical model — the riser "growth" was established in which the double grew up was combined with the anakonda algorithm. The automated design of risers came true. This is a relatively new design method of risers. The static design of risers was changed into dynamic growth process during which the riser seed was first put on casting which needed feeding, then the natural growth of the riser in the process of optimization iteration was achieved, finally the riser would grow to a desirable size, thus the design of the riser was completed and examples were also given to verify.These studies laid a solid foundation for the integrated and intelligent design of casting process. It proved that the method of riser design has a strong technological advantage and wide application prospect, in which numerical simulation is combined with simulation optimization technique.