针对城市水灾害系统的不确定性特征,提出了基于RBF神经网络和云模型理论的RBF-C风险等级评估方法。选取影响城市水灾害的4个基本评价因子,依据实测水文频率曲线确定相应于各风险等级的标准限值,并生成各评价因子下风险等级的综合云模型。用评价因子实测时间序列进行RBF神经网络建模,预测值代入综合云模型得到水灾害风险等级确定度分布。实例证明,RBF-C风险等级评估方法能够改进评价过程中风险归属不确定性问题,评估结果能较为准确地反映出城市水灾害的风险程度。%Aiming at the uncertainty characteristics of urban water disaster system,this paper proposed a risk assessment method based on RBF-ANN and cloud model (RBF-C). Selecting four basic evaluation factors of urban water disaster and according to the measured hydrological frequen-cy curve,this paper determined the risk limits corresponding to each level and generated comprehensive cloud models for risk levels of each evalua-tion factor. Using the evaluation factors of measured time series to establish RBF-ANN,the predictive values were got into the comprehensive cloud model to obtain the distributions of certainty degrees of water disaster risk levels. The study case shows that the RBF-C method can improve the un-certainty problem of the risk ownership in evaluation process and the evaluation results can reflect the risk degree of urban water disaster accurately.
展开▼