Vibration Sensing And Analysis Of Precast Concrete Pavement

Pavement vibration,which is the dynamic response of pavement structure under vehicle loads,is closely related to the loading conditions as well as the structural conditions of pavement.Understanding the influence rules of loading and structural conditions on pavement vibration’s response and modal characteristics,sensing the dynamic change in temporal-spatial-spectral domains of the pavement’s vibration field with accuracy,and analyzing the coupled effects of structural and loading conditions on pavement vibration reasonably.They are helpful for traffic monitoring and structural health monitoring and also an effective way to smarten pavement.Precast cement concrete pavement can be assembled with high precision and replaced/repaired conveniently.It could be a reliable carrier of smart sensors and an important structure for future smart pavement.Thus this research work is focused on the vibration of precast concrete pavement,including the vibration analysis models,the response and modal characteristics,the distributed vibration sensing method,the data processing methods of multidimensional transform,and the analysis methods of structure-related/load-related vibration features.A series of tests and engineering applications also validate this research work.Conclusions are as follows:(1)The flexible function layer of precast pavement was regarded as a part of the transverse isotropic plate.The joint conditions were expressed based on the elastic boundary.The theoretical model of the precast pavement was established using beam functions,while the three-dimensional numerical model was developed by the finite element method.Considering the simulation of joints and the influence of thickness,those two models were compared to clarify the appropriate application situation of each other.Model optimization was also carried out to develop complementary models of pavement vibration analysis,as well as find out the key parameters affecting pavement vibration.(2)Through theoretical analysis and numerical simulation,the influence mechanism of load and structural parameters on vibration mode and dynamic response are investigated: joint stiffness affects mode shape and natural frequency by boundary conditions;density evenly affects the natural frequencies by inertial force;modulus affects the natural frequency(higher-order is more significant)by bending internal force;the support reaction force coefficient affects the natural frequency(lower order is more significant)by the support force;the magnitude of the load affects the response amplitude by the excitation strength;the load position affects the coordinate of vibration energy’s center;the load speed affects the movement direction and speed of the coordinate of vibration energy’s center.(3)A sensing unit of optical fiber ring structure adapted to the multi-dimensional characteristics of pavement vibration was developed,and the spatial resolution was increased to 0.30 m.Boundary-arranged,cross-arranged,and full-covered array layouts of sensing units were designed for different scenarios,and the recommended range of layout parameters was determined(e.g.distance of gap≤0.60 m,angle of layout ≤45°).The vibration sensing system’s architecture was proposed to establish the conversion relationship between the optical fiber axial coordinate and the real position coordinate.The control mechanism of the sensing system’s performance was also investigated based on the parameters of the measurement device.(4)Based on wavelet transform and energy ratio of short-term to long-term,data preprocessing methods such as noise reduction and time interval extraction were developed,which can improve the signal-to-noise ratio and retain effective information(deviation in frequency<0.01 Hz).The analysis methods of short-time energy in temporal-spatial domains and power spectral density in spectral-spatial domains were proposed,as well as the multidimensional transform for those extracted vibration characteristics between optical fiber axial coordinate and the real position coordinate.The accuracy is 0.30 m and the sensitivity is 0.35 m for the identification of vibration energy’s center.Compared with the accelerator’s measurements,the modal assurance criterion is above 0.90 and the accuracy in frequency is below 1 Hz.(5)Indicators of the load-related/structure-related vibration features were proposed according to the influence rules and mechanism of loads and structural parameters.Through multi-order modal identification,binary linear fitting,and parameter estimation,an algorithm was developed for decoupling the influence of loads and structural conditions.and their analysis methods were proposed based on the multidimensional transform of vibration characteristics.There were sixty-one scenarios conducted during the full-scale test.Results indicate that the vibration analysis results are in good agreement with the measured results: The analytical error of load position is less than 4.40 cm,the analytical error of speed is less than 0.86 km/h,and the Rsquare between load magnitude and analytical results is 0.90～0.96,and the correlation coefficient between analyzed structural parameters and deflection-based backcalculated results is 0.85～0.92.(6)The vibration sensing and analysis methods were applied in two cases,located in Chenhai Highway and Tongji University respectively.Cross-arranged and fullcovered layouts of sensing units were utilized in these two cases for vibration sensing and analysis.Results indicate under the actual engineering environment and real vehicle’s excitation,vibration systems(both cross-arranged and full-covered layout)and vibration analysis methods can provide stable performance: the sensing system can measure the dissipation characteristics of short-time energy in temporal-spatial domains,as well as the transfer characteristics of power spectral density in spectral-spatial domains.The analyzed structure-related features can quantitatively reflect the changes in the pavement modulus and support force after 10 months;the load-related features can quantitatively reflect the vehicle speed and position changes(the average error of speed is 3.05 km/h,and the average error of position is 14.64 cm)and also reflect different driving behaviors.