周领, 胡安妮, 吴金远. 泵站水力瞬变的二阶Godunov格式模型构建[J]. 农业工程学报, 2022, 38(19): 42-50. DOI: 10.11975/j.issn.1002-6819.2022.19.005
    引用本文: 周领, 胡安妮, 吴金远. 泵站水力瞬变的二阶Godunov格式模型构建[J]. 农业工程学报, 2022, 38(19): 42-50. DOI: 10.11975/j.issn.1002-6819.2022.19.005
    Zhou Ling, Hu Anni, Wu Jinyuan. Construction of the second-order Godunov scheme model for hydraulic transients in pumping stations[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2022, 38(19): 42-50. DOI: 10.11975/j.issn.1002-6819.2022.19.005
    Citation: Zhou Ling, Hu Anni, Wu Jinyuan. Construction of the second-order Godunov scheme model for hydraulic transients in pumping stations[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2022, 38(19): 42-50. DOI: 10.11975/j.issn.1002-6819.2022.19.005

    泵站水力瞬变的二阶Godunov格式模型构建

    Construction of the second-order Godunov scheme model for hydraulic transients in pumping stations

    • 摘要: 为了提高复杂泵站系统水力瞬变数值模拟的高效性和稳定性,该研究基于泵站系统水力瞬变问题,建立有限体积法Godunov格式的数学模型,对简单管道系统和复杂泵站系统进行模拟研究。与常用的特征线法求解泵站水力模型方程不同,该模型引进有限体积法二阶Godunov格式对模型进行离散,用Riemann求解器对离散通量进行求解。使用MUSCL-Hancock方法进行界面数值重构,采用MINMOD斜率限制器避免虚假震荡。提出双虚拟单元边界处理方法,实现计算区域与边界同时达到二阶精度。将所建模型计算结果与精确解、经典算例数据进行对比,并针对库朗数取值和计算网格数进行敏感性分析。结果表明:所建模型模拟结果与精确解、经典算例数据吻合较好;与特征线法相比,二阶Godunov格式更加准确、稳定且高效。对于简单管道系统,特征线法计算耗时0.227 s,二阶Godunov格式计算耗时0.017 s。对于实际泵站系统,由于存在多特性的管道结构,二阶Godunov格式模拟时需稍微降低库朗数。而采用特征线法进行泵站水力过渡过程计算时,若不调整管道长度或者波速,管道中库朗数会小于1,在该文算例中,库朗数为0.72~0.76,模拟计算结果偏差很大。所以需要调整局部管道长度或波速,以达到库朗数为1的条件,这样处理因改变管道特性而引入计算误差。综上,二阶Godunov格式模拟方法可以更有效提高传统泵站系统水力瞬变模拟的高效性、稳定性以及准确性。

       

      Abstract: This study aims to implement a more efficient and stable numerical simulation of the hydraulic transient in a complex pumping station system. A finite volume method (FVM) Godunov scheme was established to simulate the simple pipeline and complex pumping station system. The FVM was then introduced to discretize the mathematical models, while the Riemann solver was selected to solve the discrete flux. The MUSCL-Hancock method was utilized to reconstruct the numerical data at the interface of control volumes. The higher numerical accuracy and stability were realized in the Godunov scheme, compared with the frequently-used method of characteristics. Meanwhile, the MINMOD slope limiter was used to avoid false oscillation. The boundary processing of the dual virtual unit was then presented for the second-order accuracy of both the computational region and the boundary, particularly for the simpler computation. The simulation of the improved model was in good agreement with the exact solution and the classical examples. The sensitivity analysis was also performed on the Courant and grid number. Furthermore, a more accurate, stable, and efficient performance was achieved in the second-order Godunov scheme, compared with the method of characteristics. More importantly, there was more outstanding attenuation with the decrease of the Courant number for a simple pipeline system. The computation time of the second-order Godunov scheme was 0.017 s at the same accuracy, compared with the method of characteristics (0.227 s). Consequently, a more stable and efficient performance was achieved in the second-order Godunov scheme. In the actual pumping system with the multiple-characteristics pipe structure, the second-order Godunov scheme required only a slight reduction in the Courant numbers, indicating the simple and convenient way for high numerical accuracy. Once the method of characteristics was used to calculate the hydraulic transition of the pumping station, the Courant number in the pipeline was less than 1 at the same length or wave velocity of the pipeline. By contrast, the Courant number was 0.72-0.76 in this case, indicating a very different simulation from the actual. Therefore, it is necessary to adjust the local pipeline length or wave velocity for the condition that the Courant number was 1. The tedious operation can lead to calculation errors, due to the change in pipeline characteristics. The accuracy can be improved but with less computational efficiency, if the wave velocity remained unchanged to increase the number of computational grids. In the method of characteristics, the number of grids can properly improve the accuracy of the calculation but with the doubled computation time, when the Courant number was less than 1. In the second-order Godunov scheme, there was little effect of grid number on the accuracy of the calculation but with the longer calculation time, whether the Courant number was equal to or less than 1. Therefore, a finer grid was preferred in the method of characteristics for the same accuracy requirements, when the Courant number was less than 1 in the transient process of the simulated pump system. Therefore, the second-order Godunov scheme can accurately simulate the process lines of rotational speed, discharge, and outlet pressure parameters during the hydraulic transient of the pump system. Anyway, the second-order Godunov scheme can be expected to effectively improve the efficiency, stability, and accuracy of hydraulic transient simulation of traditional pumping station systems.

       

    /

    返回文章
    返回