The Application Of Fuzzy Clustering And Comprehensive Evaluation In Deformation Analysis

Deformation monitoring is an important way of analyzing and evaluating the safety status of structures, validating the design parameter, feedback design and construction and the research on the rules of normal deformation monitoring and forecast the deformation. The work can be divided into two parts, one of which is the field monitoring work, the other one of which is data processing. The interior work of large engineering deformation monitoring is trival.lt has huge amount of data and involves too many contents. There are many ways of the stability analysis.Each of these commonly used methods for analyzing the stability has its advantages and disadvantages, as far as the disadvantages of these methods are concerned, the calculating way of the average gap method is trival, while the tolerance test to judge is general. But in practical work, only a single method for stability analysis is often used. The fuzzy clustering analysis can function as a simple calculation method of stability analysis of reference.If a variety of monitoring methods are used to monitor the same engineering structure, general analysis process is to analyze its different sub item of the project. Then use the index to judge the stability of each part or to forecast the trend of change through the analysis of the deformation. Finally make the maintenance or reinforcement scheme by experience to ensure safe operation. Comprehensive evaluation can consider the impacts of many factors by comprehensive analysis concerning different weights to each factor and providing an overall result finally.In this paper, we process the data of deformation monitoring mainly by three methods, the one of which is fuzzy cluster analysis, the other one of which is subordinate function analysis and the last one of which is comprehensive evaluation of the fuzzy mathematics. Do the clustering analysis test through one-dimensional and two-dimensional simulation data or measured data to judge the similarity of multiperiod deformation monitoring data. Then select suitable clustering model according to actual situation. Then use the appropriate model to analyze the similarity of three-dimensional measured data. Analyze the stability of the measuring point and compare with the result of commonly methods. According to the characteristics of element of the fuzzy similar matrix represents the similarity of two objects, I use fuzzy normal distribution function to establish fuzzy similar matrix for fuzzy cluster analysis and verify the classification results by simulation data. For overall stability analysis, I use different methods such as close degree, single comprehensive evaluation and analytic hierarchy process for analysis and judgment.