Research On Damage Diagnosis Technology Of Offshore Platform Based On Artifical Neural Network

In recent years, as more and more offshore platforms were constructed and threw in an usage, the monitor of structural health of ocean terraces becomes a popular topic. Because parts of structure of ocean terraces are under the sea level,finding out the structural damnification is not easy ; In the extremity bad sea condition, especially meeting huge storm, once some important parts of the structure are damnified, the damnification will expand rapidly, It may make the whole of the structure break down, So it is a problem to monitor the underwater parts over a long period of time. But with spreading of the Fiber optical sensing technology, it is easy to do.With collecting the data of damnification of the ocean structure by Fiber optical sensing technology and analyzing the damnification by neural network , I did a certain research to the harm diagnosis of the ocean terrace, main work contents are as follows:1) Analyzed the fiber optical sensing technology . Analyzed the characteristic, physics configuration and sending theory of the fiber sensor , and also analyzed the demodulation theory of signal of the FBG . Main method was showing theory and guiding expressions. At the same time , analyzed some points of the action of fiber sensor in the projects that must be noticed., include: the problems of sensor placement, laying and protecting of fiber sensors. Then I made a mathematic model of the FBGS,so that can know more about the fiber sensor.2) Made a model of the whole offshore platform by finite element method, and then did a Modal analysis. Applying a noumenal model of the offshore platform, in order to do the Modal analysis. By emulational model, did a Modal analysis to some parts of the offshore platform. Then collecting the model data of the damnified offshore platform.3) diagnosing the condition of the damnified offshore platform by BP network. I introduce the configuration of the neural network and also the theory of working, Especially for the BP network. Next, introducing the arithmetic and ameliorated arithmetic. At the end, working out the model frequency and the model shaking type by applying the finite element model of the damnified offshore platform. Finally, making up a BP network by the data got below, and then doing a damage detection to the offshore platform to find out the damnified parts of offshore platform.