Improvement Of Point Estimate Method For Statistical Moment And Its Application In Member Importance Evaluation Of Structural System

In the field of analysis of stochastic system and structural reliability, it is a frequentproblem for evaluating statistical moment of function. Point estimate method based onnumerical integration is the simplest and most efficient approach for statistic momentsevaluation. Though many point estimate algorithms have been put forward andmeanwhile proved to be accurate and effective by numerical cases, but it is doubtful toapply these methods to more common cases because of being lack of theoretical support.In order to apply point estimate method more rationally in practical engineering system,the computational performanc of several typicl point estimate methods are studied andcompared, and the most rational point estimate method is given out.By comparing the typical point estimate methods, it’s found that the nonlineartransformation between original variable and reference variable in Zhou&Nowak’spoint estimate method lead the precision of this method becoming lower. Hence,secondly, linear transformation is introduced into the point estimate method, and thecalculated points and weights for reference variable are determined in advance. Theproposed improved point estimate mothod by introduceing in linear transformationbetween original variable and reference variable for calculating moments is proved tohave a higer precision than Zhao&Ono’s point estimate mothod without loss of highefficiency.The importance assessment of structure’s members has directive significance forstructure design, detection and maintenance. The importanc assessment of structure’scomponents usually is implemented through evaluating of the component’s influence onstructure’s certain performance index. Except for the importance assessment ofstructure’s members based on reliability index, at present, all researchs on theimportance assessment of structure’s members are all for the analysis of deterministicstructure. However, in actual, the input parameters and output response of structure arestochastic, that is, structure’s certain performance index is a stochastic physical quantity.Hence, in this work, in order to consider the randomness of load，material properties andso on, statistical moments of structure’s performance index is proposed for theimportance assessment of structure’s members. The proposed structure’s memberimportanc index based on statistical moments and structure’s member importanc indexbased on reliability is used for member importance analysis of several example. Finally, the proposed method and the methed based on reliability index forimportance assessment of structure’s members are used for the importance assessmentof a electrical transmission tower’s main members. By analysis of the memberimportanc index based on structural ultimate bearing capacity of transmission tower, it’sfound that the members in two sides of the end of transmission tower’s crossarm and thethe vertical components of the bottom to middle of transmission tower are the mostimportant. The alternate load path need to set at the end of transmission tower’scrossarm in case of the corssarm turning yielded. The vertical components of the bottomto middle of transmission tower should be set at a high level of reliability.