Study On BFC Based POD Reduced-order Model And Its Application On Heated Oil Pipeline

Thermal simulations of heated oil pipeline in massive conditions are needed for its reliability-based flowing safety evaluation.However,the low computing speed of traditional numerical methods,such as Finite Volume Method(FVM),causes great inconvenience to the pipeline flowing safety evaluation.In addition,other problems of heated oil pipeline requuiring large scale of thermal calculations also demand on high computing efficiency.So,it is urgent to develop fast thermal calculation method for heated oil pipeline.POD(Proper orthogonal decomposition)reduced-order model is a powerful tool to achieve such goal.Nevertheless,POD reduced-order models recorded in literatures have some drawbacks.The main one is that they cannot be applied to thermal calculations of heated oil pipelines with different geometry parameters such as depth and diameter.This strongly limits the application of POD reduced-order model in heated oil pipeline's thermal calculation.Therefore,in this paper,POD reduced-order model and body-fitted coordinate(BFC)are combined to overcome the main drawback.First,the BFC based POD reduced-order models are established and systematically studied.Then,they are employed to obtain the fast thermal calculation of heated oil pipeline.The details are stated as following:First,this paper built the BFC based POD reduced-order model for heat conduction problem.And the influence of Jacobi spatial derivative on the application range of the established POD reduced-order model is invesigated.The harmonic scheme for the discretization of temperature basis function spatial derivative is deducted,which can increase the accuracy of the established POD reduced-order model.Then,BFC based POD reduced-order models for fluid flow and natural convection problem are developed.The influence of the velocity basis generation procedures on the application range of the established POD reduced-order model are theoretical analyzed.The ?calibrated method? is proposed to enhance the accuracy and stability of steady fluid flow POD reduced-order model.Application procedures of the established natural convection POD reduced-order model are also studied,on the condition that it is applied to fluid-solid domains(Natural convection and heat conduction coupling problem).Besides,by typical examples,it is shown that the established POD reduced-order models have good degree of accuracy and robustness,and the calculation efficiency of POD reduced-order model can be hundreds times faster,even more,than that of FVM.Based on the above study,the established heat conduction POD reduced-order model is employed to the fast thermal calculation of steady-state transportation,commissioning,batch transportation of cold and hot oil,and shutdown(?equivalent heat conductivity coefficient? method)of buried heated oil pipelines.The established natural convection POD reduced-order model is employed to obtain fast ?detailed thermal simulation?(in the case that flow is in laminar state during shutdown)for overhead-pipelines' shutdown.Through the given examples,it could be found that the established POD reduced-order models could achieve the fast thermal calculation of heated oil pipeline in different geometries and it could improve the computing efficiency of thermal calculation by 107~301 times.