基于区间分析理论的变形监测数据处理方法

A Method for Deformation Monitoring Data Processing Based on Interval Analysis Theory

基于点变量方式的变形监测数据分析与处理,由于变形观测数据的近似或扰动,与真实形变存在一定误差,在一定的观测手段和方法下,这类误差导致目标函数的变化并不大,但其所求的理论分析模型参数却会存在巨大差异。针对点变量表达方式不能充分反映结构变形信息的复杂性,提出了以区间分析理论来进行变形监测数据分析与处理。基于区间分析理论,建立起基于区间变量和区间分析理论的变形监测数据处理方法。提出了区间参数变量的4种取值方法:概率统计取值法、仪器精度法、最大最小值法及综合误差法;基于区间超宽度扩展理论,提出了减小区间超宽度的3种措施:改写监测数据处理的表达式、缩小监测参数的区间宽度及选择合理的数据处理模型。基于时间序列模型及灰色系统模型,以区间参数代替点参数,采用区间分析定义的运算法则,对工程变形监测数据序列进行了分析。工程算例表明:一般点参数预测模型的预测值,由于变形体变形规律的复杂性,即使采用与变形规律拟合程度很好的数学模型,也很难与实测变形值符合,但对于预测的区间值,实测变形值落在预测值的区间内,概率是比较大的。区间预测值能较点预测值进一步提高理论预测模型的预测准确性,也能更好地反映工程实体变形的复杂特点。

Due to the approximation or disturbance,the deformation monitoring data analysis and processing based on the point variable method has certain errors with the real deformation.Under certain observation methods,such errors lead to little change in the objective function,but there are huge differences in the theoretical analysis model parameters that are sought.In view of the expression of point variables cannot fully reflect the complexity of structural deformation information,the analysis and processing of deformation monitoring data by interval analysis theory is proposed.Based on the interval analysis theory,a deformation monitoring data processing method based on interval variable and interval analysis theory is established.Four methods for determining the interval parameter variables are proposed:probability and statistical value method,instrument precision method,maximum and minimum method,and comprehensive error method. Based on the interval hyper-width expansion theory,3 measures to reduce the interval hyper-width are proposed:rewriting the expression of monitoring data processing,narrowing the interval width of monitoring parameters,and selecting a reasonable data processing model.Based on the time series model and the ~ay system model,the interval parameters are used to replace the point parameters,and the engineering deformation monitoring data sequence is analyzed by the algorithm defined by the interval analysis.The engineering example shows that for the prediction value obtained by the general point parameter prediction model,due to the complexity of the deformation rule of the deformation body,even if using a mathematical model with a good degree of fitting with the deformation rule,it is difficult to match the measured deformation value,but for the predicted interval value,the measured deformation value falls within the interval of the predicted value,and the probability is relatively large.'The interval prediction value can further improve the prediction accuracy of the theoretical prediction model compared with the point prediction value,and can better reflect the complex characteristics of the engineering entity deformation.