气液混输泵优化设计及其数值模拟

发布时间：2018-04-19 23:14

本文选题：混输泵 + 优化设计　；参考：《兰州理工大学》2012年硕士论文

【摘要】：多相混输技术是近年来发展起来的一种高效经济的油田开发方式。多相混输泵作为多相混输技术的核心设备，它是一种集常规液相泵和气体压缩机性能于一体的多相输送装置。叶轮作为混输泵的核心部件，直接影响着泵中流体的流动稳定性以及泵的效率。因此，对叶片进行优化设计，提高其工作性能对提高混输泵的效率有着重要的意义。本课题是基于CFD数值计算、神经网络及优化算法相结合的混输泵叶片优化设计。该方法首先对叶轮叶片进行了参数化设计，选取参数化后的12个控制点作为优化变量，确定优化变量的3个取值，通过改变优化变量的取值，从而达到改变叶型的目的。以优化变量为因素，3个取值为水平，使用L27_3_13正交实验设计表设计出27组实验方案。对每组方案进行建模，通过fluent软件预测每组方案的效率。27组方案做为神经网络的样本数据，其中20组做为网络训练样本，7组作为预测样本，优化变量作为网络输入，计算效率作为网络输出。训练好的神经网络分别作为遗传算法及粒子群算法的适应度值函数，利用遗传算法及粒子群算法的全局寻优能力，对叶片进行优化。对两种优化算法输出的最优解进行对比，结果显示遗传算法的输出解更优，最终以遗传算法的输出解作为混输泵叶片的最优模型。对优化后及优化前的混输泵叶轮分别在含气率为0.1、0.3、0.5、0.7、0.9五种工况下进行数值模拟，对比其优化效果。从优化结果来看，优化后的混输泵在各含气率下效率及扬程均有所提高，且在设计工况下优化效果最为明显，优化后的计算效率最大提高了1.91%。这也在一定程度上证明了本文的优化设计方法的可行性，该优化设计方法适用范围广、鲁棒性强，有较强的工程应用价值。本文还从叶栅理论、速度三角形及泵内多相流动理论分析入手，，初步探讨了混输泵重要结构参数与性能参数间的对应关系，推导出混输泵流量、扬程、效率的表达式，为进一步改善混输泵性能提供了理论依据。
[Abstract]:Multiphase mixing technology is an efficient and economical oil field development method developed in recent years. As the core equipment of multiphase mixing technology, multiphase pump is a kind of multiphase transportation device which integrates the performance of conventional liquid phase pump and gas compressor. Impeller, as the core component of the mixed pump, directly affects the flow stability and efficiency of the pump. Therefore, it is of great significance to improve the efficiency of the mixed pump by optimizing the blade design and improving its working performance. This subject is based on CFD numerical calculation, neural network and optimization algorithm combined with the mixed pump blade optimization design. In this method, the impeller blade was first parameterized, 12 control points were selected as optimization variables, 3 values of optimization variables were determined, and the change of blade shape was achieved by changing the value of optimization variables. Taking the optimization variable as the factor and the three values as the level, 27 groups of experimental schemes were designed by using the L27313 orthogonal experimental design table. The efficiency of each group is predicted by fluent software. 27 groups of schemes are used as neural network sample data, 20 groups as network training samples and 7 groups as prediction samples, and the optimization variables as network input. The computational efficiency is the output of the network. The trained neural network is used as the fitness function of genetic algorithm and particle swarm optimization algorithm respectively. The global optimization ability of genetic algorithm and particle swarm optimization algorithm is used to optimize the blade. The results show that the output solution of genetic algorithm is better than that of genetic algorithm. Finally, the output solution of genetic algorithm is taken as the optimal model of mixed pump blade. After optimization and before optimization, the impellers of mixed transportation pump were numerically simulated under five conditions of 0.1g / 0.3g / 0.50.50.7g / 0.9, respectively, and the optimization results were compared. From the result of optimization, the efficiency and lift of the pump are improved under each gas content, and the optimization effect is the most obvious under the design condition, and the calculation efficiency of the optimized pump is improved by 1.91%. It also proves to some extent the feasibility of the optimal design method in this paper. This method has a wide range of application, strong robustness, and has a strong engineering application value. Based on the cascade theory, velocity triangle and multiphase flow theory, the corresponding relationship between the important structural parameters and the performance parameters of the pump is preliminarily discussed, and the expressions of the flow rate, lift and efficiency of the mixed pump are deduced. It provides a theoretical basis for further improving the performance of the mixed pump.
【学位授予单位】：兰州理工大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TH311

【引证文献】