基于大坝安全监测的复合不确定模型优化

Complex Uncertain Model Optimization Based on Dam Safety Monitoring

由于大坝安全监测对含有不确定性信息效应量的预测精度要求越来越高,常规模型已经满足不了监控需求,为此,对常规的复合不确定模型进行了改进:考虑到监测样本信息的不足,分配函数用考虑信息增量的二次型代替线型;考虑到隶属度随自变量的变化有一定规律,模糊推理的矩阵运算用贴近度推理替代Max-Min推理,以期提高模型精度;采用综合有效度对建模的结果进行评价。实例分析表明,优化后的不确定模型有效提高了拟合和预测精度,为利用不确定信息对大坝安全进行分析提供了更为精确的手段。

Because the demanding on prediction accuracy of information volume effect with uncertainties is continuously increased in dam safety monitoring,it needs to optimize the conventional composite uncertainty model.The following improvements are carried out：the quadratic distribution function is used to replace the linear function after considering the lack of monitoring sample data,and the close degree reasoning is used to replace the Max-Min reasoning of fuzzy relation matrix operation after taking into account the changes of degree of membership with the independent variable being certain regularity.Finally the comprehensive availability index is used to evaluate the results of modeling.The application example shows that,the optimized uncertain model can effectively improve the precision of fitting and prediction and provides a quantitative basis for the analysis of dam safety by using uncertain information.