Multi-sensor Data Fusion And Safety Evaluation Of Extra-long-span Suspension Bridge

Suspension bridges have become one of the main bridge types for large-span bridges due to their strong spanning ability,beautiful appearance,and high material utilization.With the growth of operating years,under the coupling effect of the external service environment and the deterioration of internal materials,the performance of the bridge structure will inevitably deteriorate.In order to ensure the service safety of bridge structures,structural health monitoring systems have been widely used in the operation monitoring of long-span bridges.At present,bridge health monitoring has made great progress in smart sensors,system integration,wired and wireless transmission,massive data storage and analysis,but its applications mostly focus on the accumulation of historical data and the overall security under accidental events.Evaluation,etc.,the monitoring data accumulated by the existing monitoring system has not been fully interpreted.How to effectively deal with bridge monitoring data and excavate the information content of monitoring data is still a technical problem to be solved urgently for bridge health monitoring.Based on the investigation and combing of bridge health monitoring and safety assessment methods at home and abroad,this paper relies on Chongqing Egongyan Track Suspension Bridge,using theoretical analysis,numerical simulation and real bridge verification research methods to develop multi-sensor data for extra-large span suspension bridge Convergence and security assessment research,the main research contents and conclusions of this article are as follows:(1)Conducted research on data processing methods for health monitoring based on multi-sensor data fusion,and used Kalman combined filtering to fuse information from multiple sensors and compared it with traditional Kalman filtering.The comparison results show that:traditional Kalman filtering can reduce the interference of noise on monitoring data;Kalman combined filtering is better than traditional Kalman filtering to reduce the noise of monitoring data,and the accuracy of the obtained data is higher.(2)A full bridge state inversion method based on limited monitoring point deformation monitoring data for a super-long-span suspension bridge was developed,the influence of temperature load on the bridge deformation was analyzed.A multivariate linear regression model that changes with the month is proposed to remove the temperature load from the deformation monitoring values,and the deformation data from the temperature load is used to retrieve the state of the full bridge.The influence of the number of measurement points and the location of the measurement points on the inversion effect is analyzed,and a reasonable layout plan for the measurement points of Chongqing Egongyan track suspension bridge is obtained.The results show that the temperature load has a great influence on the deformation of the bridge,and the temperature response of the deformation data should be removed during the finite point inversion;the more the number of measurement points,the more accurate the stress state of the full bridge structure obtained,and the number of measurement points increases to a certain In terms of quantity,it is not obvious to improve the accuracy.The inversion accuracy of the measurement points evenly distributed is higher than the inversion accuracy where the measurement points are centrally arranged at the point where the internal force or deformation is the largest.Under the condition of limited measuring points,the reasonable layout of the measuring points of Chongqing Egongyan Track Suspension Bridge is:the number of measuring points is 13,and the location of the measuring points is the main span L/8,L/4,3L/8,L/2,5L/8,3L/4,7L/8;side spans L/4,L/2,3L/4.(3)A safety assessment method based on the deformation envelope of the super-long-span suspension bridge was carried out.Through the deformation monitoring data of Chongqing Egongyan Track Suspension Bridge,the safety assessment method was carried out by the deformation envelope method.The results show that the measured deformation values of the main span L/4 and L/2 measuring points are within the revised theoretically calculated deformation values within the selected time period,and the bridge is in safe operation.(4)The sensitivity of the train axle load and crowd load to the theoretically calculated deformation value is analyzed,and a correction method for the theoretically calculated deformation value based on the measured data is proposed.The revised theoretical calculation value is used to carry out the Chongqing Egongyan track suspension bridge security assessment.The results show that the change of train axle load and crowd load are more sensitive to the theoretical calculated deformation value.When the designed train axle load differs from the actual train axle load by 20k N,the relative error between the actual theoretical calculated deformation value and the design theoretical calculated deformation value is 10%The difference between the design crowd load and the actual crowd load is 0.5k N/m~2,and the relative error between the actual theoretically calculated deformation value and the design theoretically calculated deformation value is about 5%.Based on the measured axle load,mathematical statistical methods are used to correct the theoretically calculated deformation value.The corrected theoretically calculated deformation value is reduced by about 15%compared to the theoretically calculated deformation value before correction.