基于数据驱动的管道流致振动计算方法  

A Data-Driven Method for Calculating Flow Induced Vibration in Pipelines

作　　者：王斌斌 WANG Binbin(Institute of Nuclear Engineering Research and Design,China Nuclear Power Engineering Co.,Ltd.,Beijing 10084,China)

机构地区：[1]中国核电工程有限公司北京核工程研究设计院,北京100840

出　　处：《湖北电力》2024年第1期1-6,共6页Hubei Electric Power

基　　金：中国核电工程有限公司自主科研项目(项目编号:KY21097)。

摘　　要：高速流体在流经管路系统中的弯头、阀门、三通、孔板等节流组件时,流体冲击载荷容易诱发管道的振动,甚至导致管道及其附件的损坏。常规的管道流致振动分析需要开展流体在管内流场以及流体与管道流固耦合计算,耗费计算工时较长不具备通用性。通过基于数据驱动的神经网络方法分析,建立典型流体元件的流动特性数据库,采用神经网络分析,可以在较短时间内完成流致振动的计算。该方法可以在设计阶段对可能发生振动的管道进行分析和计算,并通过改变管道设计来避免或减轻管道的振动,该方法也可适用于其他行业管道系统的流致振动分析。When high-speed fluid flows through throttling components such as elbows,valves,tee joints and orifice plates in pipeline systems,fluid impact load is easy to induce vibration of the pipline,and can even cause damage to the pipeline and its accessories.Conventional pipeline flow-induced vibration analysis carries out the flow field of fluid in the pipe and the coupling calculation of fluid and pipeline flow-solid,which consumes long working hours and is not universal.This article conducts a data-driven neural network method analysis to establish a database of flow characteristics of typical fluid components.Through the neural network method,the calculation of flow induced vibration can be completed in a much shorter time.This method can be used to analyze and calculate the possible vibration conditions of pipelines during the design period and avoid or reduce pipeline vibration by changing the pipeline design.This method can also be applied to the flow induced vibration analysis of pipeline systems in other industries.

关 键 词：热力发电 核电 流致振动 数据驱动 神经网络 管道 

分 类 号：TL334[核科学技术—核技术及应用]