Research On Dynamic Statistical Analysis Model Of Dam Monitoring Data

At present,extensive achievements have been made in the theoretical research and practical application of data analysis models and methods for dam safety monitoring.Statistical model is the most widely used,the most extensive,and the most developed data analysis model for dam safety monitoring because of its advantages of clear expression and strong interpretability.However,as the dam structure is driven by the external environment and constantly evolves internally,the dynamic characteristics of the dam safety monitoring data appear.If the static statistical analysis model of the traditional dam safety monitoring data is continued,the functional relationship between the dam deformation effect size and the environmental variables can not be described well.Therefore,in view of the defects and limitations in the statistical model and analysis theory of the existing dam safety monitoring data,this paper focuses on the analysis of the deformation monitoring data of concrete gravity dam,comprehensively applies statistical methods and theoretical research in other disciplines to the analysis of dam safety monitoring data,and is devoted to the study of the dynamic statistical analysis model of dam safety monitoring data.By understanding and studying the operation law of dam with dynamic thought,it is beneficial to evaluate and analyze the timevarying system of dam more reasonably,so as to meet the needs of some engineering applications.The research content of this paper is as follows:(1)In order to ensure the effectiveness,accuracy and applicability of the construction of dynamic statistical analysis model of dam safety monitoring data and the selection of analysis methods,the dynamic features in the dam safety monitoring data caused by the dynamic changes of the external environment or the dynamic changes of the dam structure are first identified.Through comprehensive comparative analysis and research of statistical methods related to mutation point analysis and gradient trend analysis in time series,in order to solve the problem of mutation feature identification in dam safety monitoring data,this paper identifies the mutation feature of dam safety monitoring data based on Pettitt nonparametric test method,and overbleaching treatment is introduced to the original monitoring sequence.The correlation problem in the monitoring sequence was effectively eliminated under the condition that the original mutation and trend characteristics of the original monitoring sequence were kept and the accuracy of the test results of Pettitt mutation point test was guaranteed.Through the example analysis,the Pettitt mutation test accurately estimated the number of mutation points,the time of mutation and the significance level of mutation in the monitoring sequence of horizontal displacement deformation of the measuring point,and also verified the applicability and reliability of the Pettitt mutation test.In order to solve the problem of gradual gradient feature identification in dam safety monitoring data,this paper proposes a hybrid test method based on the combination of linear function test method and MITA non-parametric test method to identify gradual gradient feature of dam safety monitoring data.Through case analysis,the paper mainly studies the existence analysis,type analysis and significance degree analysis of the gradient trend in the dam safety monitoring data.(2)The dam system changes from one stable state to another due to the dynamic changes of the external environment(such as the impact of earthquake and other factors),which indirectly leads to the existence of individual abrupt points in the dam safety monitoring data.In order to study the running state of the dam system in different time periods before and after the influence of external environmental dynamic factors such as earthquake,whether the development law and structural mechanism are the same,this paper proposes an adaptive piecewise statistical model of the dam safety monitoring data.The basic modeling process is based on the establishment of a single statistical model of dam horizontal displacement deformation and water level,temperature and cross-terms of water level and temperature,and the segmentation of the residual sequence of a single statistical model is processed by using BG segmentation method,and the segmentation results are used as the basis for the establishment of an adaptive segmentation statistical model.The data in the same section can be guaranteed to have similar statistical properties and the monitoring data in different sections can be guaranteed to have different structural characteristics.Moreover,the dynamic change characteristics and development rules of dam deformation can be simulated and reflected in stages.(3)Due to the influence of dynamic changes of dam structure(such as aging of dam material parameters),gradual characteristics appear in dam safety monitoring data.In this case,when the statistical model of dam deformation monitoring data is established,the aging term is usually considered to reflect the aging process of the dam system.However,the specific form of the aging term is not clear,and it is often expressed as a linear combination of one or more mathematical expressions of time,and it is considered that the influence of environmental factors such as water level and temperature on the dam deformation effect size is fixed.Such models ignore the variation of the strength of external loads(water level and temperature load)on the dam deformation effect quantity at different times.Based on this,this paper reflects the aging process of the dam structure from the dynamic change of the strength of the external load acting directly on the dam system,establishes the time-varying parameter state space model of the dam safety monitoring data,and uses the Kalman filter algorithm to realize the accurate estimation of the time-varying parameter model of the dam safety monitoring data.Through the example analysis,the model is verified to have good applicability and rationality for the analysis of statistical problems with gradual trend in the dam safety monitoring data.Moreover,the time-varying parameters of each variable in the model and the continuous change process of the displacement component generated by external loads are studied,and the dynamic influence law of environmental variables on the dam system is described.