FPSO Motion Response Feature Extraction And Modal Identification Based On Monitoring Database

Floating production storage and offloading(FPSO)is a large offshore platform with the functions of offshore oil exploitation,processing,storage and offloading,which is used in many sea areas in China.In order to ensure the long-term safe operation of the platform,real-time monitoring and safety and reliability analysis of FPSO are required.The large amount of monitoring data and documentation generated by long-term monitoring poses challenges to data management and analysis.Given the problems above,based on the actual monitoring plan and monitoring data of an FPSO in the South China Sea,through the research on the monitoring data collection,management and maintenance of the FPSO platform,a monitoring database system that conforms to the characteristics of FPSO monitoring is designed,and based on the database,the monitoring data were analyzed,extract platform health information.FPSO six-degree-of-freedom motion response data has complex nonlinear relationship,which makes the linear analysis method inapplicable.Kernel principal component analysis(KPCA)can effectively extract the main features of nonlinear data by kernel function,it can be used to analyze FPSO motion response data.Aiming at the problem of kernel parameter selection in practical application of KPCA method,a multi-parameter integrated KPCA method is proposed,and the effectiveness of the method is verified through face recognition experiments.Relying on the roll and pitch data of the FPSO in the monitoring database,this method was used to extract its motion response characteristics,the relationship between the FPSO motion response energy change.The influence of the platform’s own operation is excluded through the T~2 statistic,and the relationship between the FPSO motion response energy change and the environmental load is established,which provides a reference for the safe and stable operation of the FPSO.As a time-varying systems with multiple degrees of freedom,the motion response of FPSO is non-stationary.When the structure is damaged,the instantaneous frequency of its response signal will change significantly.Therefore,the accurate identification of the instantaneous frequency is the premise of realizing FPSO damage diagnosis.A TVAR-HHT modal identification method is proposed based on HHT method and TVAR model to identify modal parameters of time-varying systems with multiple degrees of freedom.The modal identification results of the time-varying simulation signal and the simulation system show that the method can indeed extract the instantaneous frequency of the time-varying system more effectively than the STFT and HHT methods.Based on the roll data of the FPSO in the monitoring database,the TVAR-HHT method was used to effectively identify its instantaneous frequency,which provided a reference for further damage diagnosis.