STUDY ON PREDICTION MODEL OF SUBMARINE CABLE STIFFNESS BASED ON PSO-SVR ALGORITHM
Submarine cable stiffness is an important index to characterize the cross-sectional mechanical properties.Fine numerical simulation methods can consider nonlinear behaviors such as interlayer contact and friction to obtain the exact stiffness.However,in the preliminary design of the submarine cable cross-section,it is necessary to conduct a comparative analysis of the stiffness under different geometric parameters,which requires a significant amount of manpower and material resources to complete the modeling and calculation of multiple comparative models.This study establishes a finite element model of the submarine cable based on the Nysted offshore wind farm project,and uses the orthogonal test method to determine the main factors influencing the tensile stiffness,counterclockwise torsional stiffness and clockwise torsional stiffness of the submarine cable.Then,the main influencing factors are used as feature inputs of the Particle Swarm Optimization-Support Vector Regression(PSO-SVR)model to establish prediction models for tensile stiffness,counterclockwise torsional stiffness and clockwise torsional stiffness of the submarine cable respectively.Besides,the prediction performance of the PSO-SVR model is compared with the GRNN neural network model and the BP neural network model.The results show that the conductor diameter,steel wire diameter,steel wire pitch and the number of armor layers have more influence on the submarine cable stiffness,while the steel wire elastic modulus has less influence on it.The determination coefficient of the PSO-SVR model is higher than 0.95 and the error is relatively low,and the prediction performance is better than that of the GRNN neural network model and the BP neural network model.This prediction model can provide technical support for the preliminary design of submarine cable structures.
offshore wind powersubmarine cablesstiffnessparticle swarm optimizationsupport vector regressionorthogonal test
NETL
NSTL
万方数据
