Research On VIV Suppression Device Optimization Design Of Deepwater Pipe Structure

The pipe system is an indispensable part of marine oil equipment,it plays a vital role in marine oil and gas extraction,its operation has also been tested by the complex marine environment,the vortex-induced vibration(VIV)of the deep water pipe structure structure is the main reason for its fatigue failure.Although the research on flow field and VIV has lasted for hundreds of years,how to accurately and quickly solve the flow field has not been solved yet.The Reynolds stress is obtained by solving complex partial differential equations with the help of various alternative turbulence models,but the calculation is huge,and computations are difficult to converge,which hinders the development of computational fluid dynamics(CFD).In recent years,artificial neural networks have been developed and widely used to solve complex nonlinear dynamic problems.In this paper,BP neural network learning and training method was adopted to directly reconstruct the method for solving the flow field in cylindrical flow around the cylinder,and the surrogate model for solving the flow field was obtained,which greatly improved the solving efficiency.At the same time,combined with the surrogate model and genetic algorithm,the optimization design of the vortex-induced vibration suppression device is carried out,which provides a new idea for the suppression of vortex-induced vibration in engineering.The numerical simulation of two-dimensional cylindrical VIV under low Reynolds number and high Reynolds number is carried out,Beiler curve is used to establish the parameterized fairing model,and the VIV responses of Reynolds numbers Re=100,Re=6000and Re=12000 are studied respectively.Based on the data interaction interface between COMSOL and MATLAB,a column VIV numerical simulation database with additional fairings is established.An experimental method is proposed to study the hidden layer structure of neural network,and training function and normalization function are studied,and established the topology of BP neural network training learning.The prediction model was trained and verified using training data and test data in database.For different Reynolds numbers,lift acting on fairing is used as the optimization objective function,based on genetic algorithm and VIV response prediction model,the parameterized fairing model is optimized and establish a fairing section model with the best VIV suppression effect,CFD numerical simulation and experimental study on the theoretical optimal model is carry out.Finally,the optimal design scheme of VIV suppression device for engineering reference is proposed.A variable-form VIV suppression fairing structure design scheme and an adaptive control and active control control scheme for VIV control are proposed.The results show that:(1)based on the BP neural network learning training reconstruction flow field,the alternative model has higher prediction accuracy and solution efficiency for solving the vortex-induced vibration response of the additional fairing of risers.(2)The optimized fairing model obtained by the genetic algorithm optimization algorithm has a low VIV suppression efficiency improvement in the low Reynolds number flow field compared with the traditional drop-shaped fairing model,and has a high VIV suppression efficiency in the high Reynolds number flow field.The VIV suppression efficiency is obviously improved compared with the conventional drop-shaped fairing model.(3)Through experiments and simulation studies,it is feasible to optimize the design of VIV suppression device based on neural network training and intelligent optimization algorithm,which can provide new research ideas for VIV suppression in engineering.