Study On Temperature Effect On Tibetan Heritage Building Based On Long-term Monitoring Data

Different types of damages and deterioration of structural performance may existed in a timber heritage architecture due to material degradation,environmental effect,human aceivities and sudden-onset disasters.The structural state can be monitored in real time for any needed preventive protection through structural health monitoring with existing technology.Temperature effect is an important factor of long-term monitoring data and in the structural condition assessment of heritage timber structure.Therefore,it is of great benefit to analyze the temperature effect on long-term monitoring data of the structure.This thesis reports on a series work on the temperature effect on the long-term monitoring of Tibetan heritage structures with statistical analysis,theoretical derivation and experiment based on a typical Tibetan heritage building.The main work and important findings are listed below:(1)The influence of ambient temperature and humidity on long-term monitored structural strain is analyzed.The principal component regression model of strain and with respect to the ambient temperature and humidity is established in every month.Sensitivity analysis on the influence of temperature and humidity on strain is conducted with Morris method and the sensitive factors of temperature and humidity are compared.Results show that the humidity effect in Lasa during July and August in a year can not be neglected,while it could be neglected in the rest of the year and the temperature effect only may need to be considered.A method combining the singular spectrum analysis and polynomial regression(SSA-PR)is proposed for modeling and removal of the temperature effect based on the ten months monitoring data from Sep.1,2013 to Jun.30,2014.Results show that the correlation between temperature and strain could be well reflected by the proposed method,and it is accurate and effective for eliminating the temperature effect from the measured strain.A standardized Novetly Index is proposed based on the residual strain from SSA-PR model for condition assessment using X-bar control chart and CUSUM control chart.(2)According to.the construction features of typical Tibetan heritage timber beam,a rotational spring and a nonlinear tension and compression spring are adoptede for modeling the boundary conditons of the beam.Then the simplified mechanical model on the temperature effect on the timber beam is derived and a theoretical model of the strain in the grain direction at the bottom at mid-span of beam with temperature effect is proposed.A timber beam of a Tibetan heritage building from a monitoring system is employed as an example to analyze the strain variations in one year.The calculated strain variations match well with the measurement.Finally,analysis of the structural parameters affecting the strain variations is conducted based on the above work.(3)The non-load test and gradation loading test under temperature effect are carried out on an unrestrained timber beam and a dovetail joint timber beam in Lasa.According to the test data,the elasticity modulus parallel to grain of the wood,the tension and compression stiffness and rotation stiffness of the dovetail joints are identified.Different parameters are determined as variables relate to the temperature variations.According to the construction details of the dovetail joint,the tension and compression stiffness of the beam in the temperature increasing stage and temperature decreasing stage are different.The influences of load and boundary conditions on the strain parallel to grain are analyzed.Results show that the extra strain is induced by temperature and load effect of the dovetail joint timber beam,which is increasing with the temperature and the load.The measured and calculated values of the extra strain matched well with each other,which indicated the theoretical basis on the extra strain is correct.(4)The SSA method is applied for short-term and long-term forecasting of the deformation of a Tibetan heritage building.Parameters L and r are selected according to the minimum MSE for the optimal forecasting scheme.The forecasting results from SSA and ARIMA are compared,and the former is found to have smaller MSE indicating a better performance compared to ARIMA.