|
| GPU-accelerated third-order finite-volume scheme pipe transient flow solver model |
| Received:November 19, 2021 Revised:December 28, 2021 |
| View Full Text View/Add Comment Download reader |
| DOI：10.7511/jslx20211119002 |
| KeyWord:GPU acceleration finite volume method the transient flow ENO scheme long-distance water delivery |
| Author | Institution |
| 莫铁祥 |
西安理工大学 省部共建西北旱区生态水利国家重点实验室, 西安 |
| 李国栋 |
西安理工大学 省部共建西北旱区生态水利国家重点实验室, 西安 |
| 李海潮 |
西安理工大学 省部共建西北旱区生态水利国家重点实验室, 西安 |
|
| Hits: 551 |
| Download times: 406 |
| Abstract: |
| In order to efficiently and accurately solve the transient flow change process of a long-distance water conveyance system,the third-order ENO finite volume scheme is used to solve the one-dimensional pipeline unsteady flow equations.The interface flux is reconstructed based on the lax-Friedrichs flux splitting method.A set of efficient and numerical models for solving transient water hammer wave in the pipeline with high accuracy are established.GPU acceleration technology is introduced to realize the efficient calculation of large-scale water conveyance systems.The model is verified by the Method of Characteristics and the first-order and second-order Godunov finite volume schemes.The results show that the third-order ENO scheme can also maintain good performance of discontinuous capture at very low Courant without non-physical oscillation.At the same time,it is highly insensitive to Courant,making the model highly flexible in meshing and significantly improving the computing speed.The application of GPU acceleration technology shows that the model has an obvious acceleration effect when the number of grids is large,and the acceleration effect is significant with the increase in the number of grids.The model can provide theoretical support for efficient,accurate,and rapid simulation and prediction of the unsteady transient process of a long-distance water conveyance system. |