Research On Dynamic Characteristics And Damage Identification Of Reinforced Concrete Simply Supported Beam Bridge Under The Influence Of Temperature

Bridge structure,as a key component of modern transportation system,plays an important role in social and economic development.Its safe operating condition is consistently associated with economic sustainable development and security of people's lives and properties.However,material performance degradation and accumulated damage of bridge are inevitable in the service process due to the gradually increasing traffic load and harsh external environment,which even causes serious bridge collapse accidents.Therefore,it is of vital practical significance to guarantee bridge operation safety and improve service life through identifying bridge damage and evaluating service condition timely and effectively.With the development of technologies on sensor and modal analysis,vibration-based damage identification(VBDI)has been widely used in bridge health monitoring and damage diagnosis.However,environmental factors,especially temperature,cause a significant change in bridge modal parameters,which leads to the failure of VBDI method.The temperature variation range close to 70 ?in seasonal frost region will certainly causes a change of modal parameter and reduces the reliability of damage identification based on VBDI.Therefore,in order to improve the accuracy of VBDI and promote its application in practical engineering,it is necessary to analyze temperature effect on modal parameters,illustrate the influence mechanism,and propose some effective methods to eliminate temperature effect on modal parameters and identify existing damages.This study was founded by the National Natural Science Foundation Project “Research on modal parameter identification and damage detection of bridge structure considering temperature effect”.In this paper,environmental effect on dynamic characteristics of reinforced concrete simply supported beam bridge in seasonal frozen region has been systematically studied,and effective temperature effect elimination methods have been proposed.On this basis,bridge damage identification and detection under the effect of temperature have been achieved based on statistical process control,fuzzy clustering and fuzzy inference theories.Detailed researches are shown as follows:1.Environmental effect on modal parameters of reinforced concrete simply supported beam bridge and corresponding internal mechanism are studied based on long-term monitoring for a reinforced concrete simply supported beam.Influence degrees of environmental factors on modal parameters are quantitatively evaluated by use of grey relational degree analysis,so as to determine the essential environmental influence factors for modal parameters.Considering internal temperature variation,the influence rules of temperature on modal parameters are revealed through statistical analysis.Moreover,the influence mechanism of temperature on modal frequencies is explained according to temperature-dependent characteristic of concrete elastic modulus and structural internal stress change caused by temperature variation.2.Considering the influence of temperature variation and non-uniform distribution on modal frequencies,regression quantitative models between them are constructed to eliminate modal frequencies caused by temperature change.Linear principal component analysis is employed to reduce the dimension of temperature parameters and eliminate the correlation between them.Prediction performances of different quantitative models established by multiple linear regression,BP neural network,support vector regression,and support vector regression optimized by genetic algorithm(SVR-GA)are compared and examined using quantitative evaluation index including root mean square error and correlation coefficient.Based on the long-term measurements and predictions of SVR-GA,modal frequencies caused by temperature variation are successfully eliminated.Comparison analysis between variation degrees of modal frequencies before and after eliminating temperature effect verifies the effectiveness of SVR-GA quantitative model.3.Aiming at existing randomness and uncertainty in modal frequencies after eliminating temperature effect,this paper proposes a damage alarming and evaluation method based on Shewhart mean control chart and fuzzy C-means clustering improved by particle swarm optimization algorithm(FCM-PSO).Statistical analysis is conducted on modal frequencies under undamage condition to determine the limits of control chart.Abnormal modal frequencies beyond the control limit indicate damage occurrence,which realizes an early-alarming of damage.FCM-PSO is employed to perform clustering analysis for modal frequencies with typical damage conditions,and calculate the clustering centers corresponding to different damage grades.Damage degree evaluation and quantitative identification are achieved based on fuzzy memberships of modal frequencies under unknown condition to corresponding cluster centers.4.Considering that the internal temperature measurement is difficult or quantitative model is unavailable,this paper proposes a novel damage diagnosis and evaluation method which is independent of temperature measurement based on auto-associative neural network(AANN),multivariable Shewhart control chart and FCM-PSO.Taking the temperature as a potential affecting variable of modal frequency,AANN is trained by the modal frequencies under undamage condition to construct the potential mapping relationship between temperature and modal frequencies.Reconstruction of modal frequencies is performed using the completely trained AANN,and damage alarming is achieved through hypothesis testing on residuals between reconstructed and measured modal frequencies by use of multivariable Shewhart control chart.Reconstructed residuals under typical damage conditions are automatically classified based on FCM-PSO so as to calculate the cluster centers with respect to damage grades.Damage degree evaluation and quantitative identification are conducted based on fuzzy memberships of residuals in test condition to cluster centers.5.Due to the uncertainty in modal parameters after eliminating temperature effect increases the difficult in locating damage and identifying local damage extent,a damage identification method with strong robustness is proposed in this paper based on uniform load surface curvature difference(ULSCD)and fuzzy inference system(FIS).This method obtains ULSCD through uniform load surface curvature(ULSC)before and after structural damage,and selects ULSCDs at nodes near damage region as damage identification index.Normalized ULSCD,as the input parameter of FIS,is transformed into fuzzy linguistic variable using Gaussian membership function.Similarly,element damage status,as the output parameter of FIS,is also transformed into fuzzy linguistic variable using Bell-shaped membership function.Fuzzy rule base is constructed by corresponding element damage status and damage identification parameters.Damage location and local damage extent are identified through fuzzy inference by FIS.Damage identification results in numerical simulation and model experiment of simply supported beam indicate that ULSCD-FIS has accurate inference capability and strong anti-noise capability,which verifies the effectiveness of proposed method.