论文部分内容阅读
为解决非常规突发洪水事件风险分析中所存在的多因素高维数据与小样本数据信息不完备问题,将非常规突发洪水事件系统进行粒度分解,在知识元层建立基于投影寻踪及信息扩散的非常规突发洪水事件演化风险分析模型.利用知识元理论,将事件间的演化风险转变为事件所处情境内承灾载体状态属性变化引发输出属性变化的风险.根据承灾载体知识元内部状态属性阀值建立后续事件演化规则,利用投影寻踪方法与信息扩散理论,将监测样本所包含风险信息扩散到输出属性的风险指标论域控制点,计算事件演化概率分布,确定非常规突发洪水事件演化风险.以桃曲坡水库为例,在上游发生暴雨的情景下,计算水库水位上升向漫坝事件演化风险.实例分析结果表明,该演化风险分析模型能够根据风险等级标准和少量监测数据,对非常规突发事件风险概率分布进行动态定量分析与评估,帮助应急决策者实时掌握事件发展状况,及时做出应急响应,减少次生事件所造成的损失.
In order to solve the problem of incomplete multi-factor high-dimension data and small sample data information in the risk analysis of unconventional flood events, the unconventional flood event system is decomposed by granularity. Based on projection pursuit and Information diffusion of unconventional flood event evolutionary risk analysis model.Using the knowledge of meta-theory, the evolutionary risk between events into situational events within the context of changes in the state attribute of the risk of triggering output attribute changes.According to the carrier knowledge The inner-state attribute threshold is used to establish the follow-up event evolution rule. By using the projection pursuit method and the information diffusion theory, the risk information contained in the monitoring sample is diffused to the risk index domain control point of the output attribute, the probability distribution of the event evolution is calculated, Taking Taoqupo Reservoir as an example, under the scenario of heavy rainstorm in the upper reaches, the evolution risk of reservoir dam rising to diffuse dam event is calculated.Example analysis results show that the evolutionary risk analysis model can be based on the risk grade standard and A small amount of monitoring data to dynamically measure the probability distribution of risk of unconventional emergencies Analysis and evaluation, emergency help policymakers grasp the real-time event development, timely emergency response, reduce the losses caused by secondary events.