The Studies On Damage Detection Methods Of Offshore Platform Structures Based On EMD-HHT

Monitoring of structural integrity has become the focus of many engineering field.Sensor technology and the corresponding signal processing technology is necessary forstructural health monitoring. Intelligent signal processing builds the bridge between thesensor signal and the structural integrity detection. How to quickly and efficientlyextract the important characteristic information from the structure vibration responsesignal collected from sensors and judge the health state of structures is of greatapplication value. The damage detection method based on the HHT is a global damagedetection method. It has the advantages of adaptability and high time-frequencyresolution compared with the conventional damage detection methods which makes it avery broad application prospects.The paper summarizes the current main methods of time-frequency analysis, anddiscusses its present situation, research results and the problems comprehensively. Nextthe basic theory, developing process and application advantages in frequency analysis ofthe HHT is expounded. Combined with the preceding research results related to HHT, ithas a further research on the application of HHT in the field of damage detection.In view of the shortcoming of the empirical mode decomposition，whose precisionis subject to noise impact, Hilbert-Huang transformation combined with WaveletTransform for structural damage detection is presented. And a damage indicator is madebased on the low-order instantaneous frequency variation identified by HHT. Numericalsimulation study is conducted on a3-storey shear model. Damage detection of the modelunder different damage degree and different noise background is carried out by usingthis indicator. The result provides certain reference for the online monitoring of thestructural condition.Dynamic characteristic parameters of the structure may be influenced byenvironmental factors and operating conditions. And the changes of the dynamiccharacteristic parameters caused by the environmental factors and operating conditions may mask the small changes caused by structure damage. In view of this situation, thispaper proposes combining Hilbert-Huang transform with the Principal ComponentAnalysis for damage detection. Numerical simulation is made by using a cantileverbeam model. Principal Component Analysis is adopted to exclude the influence oftemperature on the modal frequencies obtained by Hilbert-Huang transform, and thenthe Novelty Detection technique is carried to analyze the prediction error matrix to judgewhether the injuries have occurred. The numerical research results show that this methodcan effectively eliminate the temperature influence on modal frequencies, and canaccurately detect the occurrence of damage.Jacket platform is a commonly used form in shallow water, so the damage detectiontechnology research of such type of platforms has important practical significance. Thefinite element models of a single column platform and a jacket platform are establishedin this paper. The HHT method is carried to study the damage detection of the modelunder different damage degree and external excitation. Besides, the influence of theexternal environment is simulated by adding noise. This study provides some usefulconclusions.