Research On Monitoring Data Management And Intelligent Damage Identification Of Offshore Platform

Because offshore platforms work in the complex and changeable environment,it isimpossible to analyze the real health status of the offshore platform only by short-term monitoring.Only through the long-term and close health monitoring of the offshore platform,the long-term monitoring data of the offshore platform can be obtained,and the long-term monitoring data can be calculated and analyzed to obtain the real structural health information of the offshore platform,so as to avoid the production and life loss of the offshore platform due to structural damage or fatigue failure and other reasons to the maximum extent.In order to obtain the real structural health information of the offshore platform,this paper studies the following contents:(1)Research on monitoring data management of the offshore platformThrough the long-term monitoring of deep-sea floating platform and FPSO soft York single point mooring system,the long-term monitoring data of each offshore platform is obtained,and an intelligent long-term monitoring data management platform of offshore platform is designed.The platform can be used for data storage,data analysis and data mining.The function of data storage part is to manage the data of all long-term monitoring offshore platforms in this project,and record the detailed information of monitoring location,sensor and channel in each offshore platform;the function of data analysis part is to carry out unified measurement calculation and signal processing analysis for any data segment.Through this part,we can calculate the statistics information of the selected data segment and process the signal change of this data segment,such as fast Fourier change,wavelet transform,Hilbert-Huang change,etc.in addition,we can also obtain the structural characteristic information of this data segment,such as auto-correlation function and random decrement(RD)signature.The last part of the data mining function is to obtain the structural characteristics of the offshore platform by segmenting a large number of long-term monitoring data,and then use intelligent methods to obtain the real structural health information of the offshore platform.(2)Research on intelligent damage identification of the offshore platformIn order to solve the serious problem of using high sampling rate sensor to collect data in long-term monitoring,it will consume a lot of computing resources and space resources.In this paper,the influence of the sampling rate of the sensor on the structural damage identification is studied.By using the second-order nonlinear system simulation and model experiment,it is found that the low-frequency data based on attitude can also reflect the structural damage,and the recognition effect is not bad.Finally,the correctness of the above method is verified by the real long-term monitoring data of the offshore platform.Based on the above research,a single classification online incremental learning method is applied to the study of intelligent online damage identification of offshore platform structures.The free response characteristics of the structure are extracted from the historical normal monitoring data of the offshore platform to establish a model.Through this model,all unknown working conditions of the offshore platform are continuously learned to achieve the effect of continuous online incremental learning.Finally,the model is used to verify whether there is any abnormality in the daily real-time monitoring data.If there is some abnormality,it means that the structural characteristics of the offshore platform have changed,that is,the structure may have recessive damage,and the early warning can be given in advance.