Bridge Damage Identification Based On Data Fusion Of Dynamic Fingerprint And D-S Evidence Theory

Traffic is the lifeblood of social and economic development,and bridge is an important part of traffic network.Since the reform and opening decades,with the development of national economy,our country’s transportation network has tended to perfect.For Bridges that have been in service for a long time,due to the long-term influence of environmental erosion,load action,aging of materials and other factors,the performance of the bridge structure deteriorates and the potential risk of damage is greater.Therefore,the research on bridge structural damage identification method is of great importance.By combining the deterministic damage identification method based on curvature mode theory and generalized flexibility theory with the data fusion uncertain damage identification method,this paper proposes a damage identification method based on dynamic fingerprint and improved D-S evidence theory.A continuous beam model was built with ANSYS finite element software.The effectiveness of the method was verified and the noise resistance was studied through numerical simulation analysis.The main work and research results of this paper are as follows:(1)Five indexes of superimposed curvature mode change rate SCMCR,frequencyweighted superimposed curvature mode change rate SCMQR,generalized compliance curvature matrix diagonal difference GFCCD,generalized compliance curvature 1-norm difference GFCCN1 and generalized compliance curvature 2-norm difference GFCCN2 were constructed by curvature mode theory and generalized compliance theory.To provide optimal indexes for data fusion,the damage recognition results and anti-noise performance of five indexes under different damage conditions were compared by numerical model of simply supported beam and continuous beam.The comparison results show that the GFCCD and SCMQR are superior to the other three indexes in damage identification,location,quantification and noise resistance.(2)The preferred index GFCCD and index SCMQR were used as independent information sources for D-S evidence theory data fusion,and the bridge structural damage identification method based on the improved D-S evidence theory was proposed.Through the numerical model of simply supported beam and three-span continuous beam,the identification effect of single damage index,traditional D-S evidence theory fusion index and improved D-S evidence theory fusion index on bridge damage and the anti-noise performance of the index are compared and analyzed.The comparison results show that the improved D-S evidence theory fusion index can effectively reduce the influence of undamaged units,enhance the damage probability of damaged units,improve the misjudgment and missing judgment of damage,and accurately locate the damage with better noise resistance.(3)Because the improved D-S evidence theory will be affected by information sources,there will be missing judgment in damage recognition of some working conditions,and a damage recognition method based on two-stage data fusion is proposed.Damage identification of numerical models of simply supported beams and three-span continuous beams is studied.The results show that: Compared with the single improvement of D-S evidence theory fusion,the two-stage data fusion index D has better accuracy for structural damage identification,reduces the missing judgment of structural damage,and has better noise resistance.Under the noise level of 15%,the structural damage can still be accurately located.