Font Size: a A A

Fatigue Damage Analysis And Power Spectrum Prediction Method Based On Wide-band Stress

Posted on:2018-01-10Degree:MasterType:Thesis
Country:ChinaCandidate:Y Z ShenFull Text:PDF
GTID:2322330518454682Subject:Engineering
Abstract/Summary:PDF Full Text Request
The threat of fatigue destruction in ships and Marine engineering structures is very large.The accuracy of fatigue analysis will directly determine the economic and security of ship and marine engineering using.Therefore,domestic and foreign scholars and practitioners have attached great importance to this.The classification societies of the big manufacture countries have their own specification for structural fatigue assessments,however,due to the fact that the calculation of the load is narrow in the form of the norms,it should be in the form of wide-band in fact.So the results and the actual situation are often different.This paper provides the new method of structural fatigue analysis of a ship,the main contents of this paper are as follows:The first part is the basic theory of structural fatigue and analytical methods including Miner linear cumulative damage theory and fatigue damage calculation of spectral analysis method.The second part is the method of calculating the fatigue damage based on the equivalent stress method under the wide-band random load,taking the error function as the primary consideration,and then get a new equivalent stress range approximation model.On this basis,establish a finite element model of a bulk carrier,then calculate the results of fatigue damage of the model and the existing equivalent stress model according to the structural fatigue spectrum analysis method,then the conclusions from these are compared.The third part is deal with the double peak wide-band stress power spectral density,using neural network to predict the probability density function of stress range.Including frequency domain-time domain conversion of stress probability spectrum,stress-time history of rainflow statistics,fitting the probability density of the stress statistical results,the data obtained are used as samples to train the neural network.And then,a neural network with probability density function for predicting stress range is established.
Keywords/Search Tags:Finite element model, Fatigue damage, Equivalent stress range, Artificial neural network
PDF Full Text Request
Related items