HU Shun-ren1, CHEN Wei-min1. Dynamic threshold interval prediction for optoelectronic imaging deflection system[J]. 2009, 28(8): 59-62. DOI: 10.13873/j.1000-97872009.08.025.
光电成像挠度系统中数据动态阈值区间预测
摘要
对桥梁挠度数据的影响因子进行了深入分析
揭示了传统基于固定阈值方法判断数据信号正确性的方法是不适用的
从而在回归理论分析基础上
建立回归模型来描述挠度数据局部趋势曲线
以此预测挠度数据阈值区间范围
形成动态阈值包络。仿真实验证明:回归模型与实测曲线吻合较好
2个实验的均方误差分别为0.684 7和0.485 4
实测值都在预测值的动态阈值范围内(当输入温度值为15.61
实测值为288.89
样本大小为130时
预测的挠度阈值范围为289.048 8±0.470 8;样本大小为32时
预测的挠度阈值范围为289.019 3±0.137 6)。
Abstract
The bridge deflection scope has changed greatly
and the traditional threshold method has became unfitted.The regression model is built to describe the partial deflection curve
then to predict the next deflection threshold interval.Simulation results verify the effectiveness of the designed method and theoretical discussions.The mean-square error(MSE)of two experiment are 0.6847and 0.4854
and the true deflection value are in prediction threshold interval when the input temperature value is 15.61
the true deflection is 288.89
the stylebook data number is 130
the prediction threshold interval is 289.0488±0.4708
and the stylebook data number is 32
the prediction threshold interval is 289.0193±0.1376.