Research On Bridge Damage Dynamic Data Based On Data Fusion And Decision Tree

In the actual project,the bridge structure is affected by internal and external factors in the long-term service stage and inevitably appears damage.At the same time,with the increase of the service time of the structure,the damage degree is deepening,threatening people’s life and property safety.Therefore,it is an important measure to identify the structural damage and check the missing in time to ensure the normal use of the structure.However,the damage identification method based on dynamic fingerprint has been constantly improved by many scholars,but there are still some shortcomings,such as the lack of modal information measurement,the influence of environmental noise and the dependence on accurate finite element model,which makes it difficult to implement structural damage identification.In this context,based on the modal strain energy theory in the deterministic method of structural damage identification,combined with the data fusion theory and decision tree algorithm in information processing technology,this paper carries out the research on damage identification methods,and proposes a two-stage damage detection method based on Bayes data fusion theory and CART decision tree algorithm.The finite element model was established for simulation,the modal information of the structure was obtained by using the additional mass method,and the structural damage was located and quantitatively analyzed by calculation,and the noise resistance of the damage identification method proposed in this paper was verified and analyzed.The main work and research results of this paper are as follows:(1)Based on the theory of modal strain energy,three damage identification indexes were constructed:modal strain energy difference(MSED),modal strain energy ratio change rate(MSERC)and modal strain energy damage location(MSEDLI).The finite element model of simply supported beam and continuous beam was established with ANSYS,the first three modal information of the structure was extracted,the mean value of the first three modal strain energy indexes was calculated with MATLAB software,and the damage location ability of the modal strain energy indexes after adding mass was compared and analyzed under different working conditions.Moreover,various noise levels were added to analyze the antinoise performance of each index.The results show that the sensitivity of local damage is improved by using the method of structural additional mass,and the modal strain energy index constructed with the structural modal information obtained by the method can effectively locate the structural damage and has a certain anti-noise property.(2)Based on BAYES data fusion technology and combined with MSED,MSERC and MSEDLI three damage indicators,(Internal Fusion of Fndicators IFI)and(Fusion Between Indicators,FBI)two new damage recognition techniques.The finite element models of simply supported beams and continuous beams were established for simulation,and the modal information after the additional mass of the structure was extracted.Three kinds of damage indexes were calculated respectively,and the identification results of each index were used as the information source,and the data fusion method was used to fuse them within and between indexes,and the best synergistic results were obtained for damage location.The results show that Bayes data fusion technology can significantly improve the reliability and accuracy of damage identification,and has good noise resistance,and the identification effect of IFI index is better than that of FBI index.(3)The numerical fitting method is used to quantify the structural damage degree,and it is verified that the polynomial fitting method can effectively identify the structural damage degree,but it has a large amount of calculation and cannot meet the requirements of rapid damage identification.Therefore,the advantages of CART decision tree algorithm in data mining were utilized to classify and identify the damage degree,and the noise resistance of this method was analyzed.Finally,the simple supported beam and continuous beam were used to verify the analysis.The results show that CART algorithm can effectively identify the damage degree of bridge structure,and the two examples have high accuracy and strong robustness under different noise levels.