Integration, Management And Feature Extraction Of Monitoring Data For Offshore Platform

Offshore platform is the important engineering equipment for oil and gas exploitation and transportation at sea. Because of the environment is complex and changeable, offshore platform has fatigue damage easily for longer-serving. Model tests can not simulate the real structure of the platform because of scale effect or other restrictions. In order to ensure the safe and stable operation of the platform, the offshore platform should be monitored comprehensively for a long periods of time so that the health condition can be understood. The existing mooring force monitoring system of Tiaozhanzhe FPS in South China Sea is independent. It can not be combined with other systems and the data of mooring force can not retain because of interrupt when typhoon come. In addition, for long term monitoring, a large number of monitoring data and documents have been produced, it causes inconveniences for management and analysis of data.Given the problems above, according to Tiaozhanzhe FPS, the mooring force system is integrated into monitoring system through surveying the feasibility and methods for uninterruptible power supply of force sensors on chain stopper. It makes possible for continuous measurement of mooring force even if typhoon evacuation. Given big data in platform monitoring, a monitoring database system which is applicable to offshore platform is designed and developed after studying management of monitoring data. It can manage monitoring data efficiently and scientifically and it can be used to help analyze the data.Extracting motion characteristics of platform by use of monitoring data and mastering its health condition is hot research field now. Modal parameters of structure and response statistics are common motion characteristics. Traditional modal analysis methods are analytic, which is not real-time. Given the problem of complex mode identification of a Multi-DOF (multi-degree-of-freedom) structure based on environmental loads, a method for extracting modal parameters based on a similarity search is proposed. The free vibration attenuation functions are used as the dictionary atoms. An MP(Matching Pursuit) and a GA (Genetic Algorithm) are introduced to search for the matching atoms that are most similar to the random decrement signal of the structure response in the dictionary; thus, the modal parameters of the free response functions can be identified. The method is applied to the analysis of the model experiment of the offshore platform FPSO single point mooring system and real monitoring data, obtaining good results.According to long-term monitoring of offshore, a method which based on KPCA to extract response feature of FPSO single point mooring system is proposed. This method realized the nonlinear mapping from raw data to feature space by using kernel functions to the original data space and calculated the energy feature of structure response in maximum direction. Analysis on long-term monitoring of offshore FPSO single point mooring system indicates:the variation characteristic of response energy basically depends on environmental loads; this method reduces the impact caused by offshore itself on response changes and can provide reference for offshore safe production.