中国是世界上受台风灾害影响最为严重的国家之一。近年来，我国东南沿海地区台风灾害频发,对城市关键基础设施系统造成较为严重的破坏。关键基础设施是城市各项功能能够正常运转的重要保障，一旦受到破坏，不仅会造成巨大的经济损失，也会极大地影响了人民的正常生活和社会的正常运行。因此，掌握台风灾害下关键基础设施的破坏特征、恢复规律并开展风险评估工作，对于做好防范和减轻台风灾害风险、有效开展灾后恢复救援和建设韧性城市具有重要意义。

本文以海南省为研究区，基于历史台风灾害对海南省公路和电网造成的损失破坏和恢复数据，构建了台风灾害下海南省公路和电网的韧性模型，并基于台风灾害模拟事件集，评估了台风灾害下海南省公路和电网的灾害韧性以及面临的风险。主要研究结果有：

（1）基于历史台风灾害对海南省造成的公路阻断和恢复记录数据，得到每场台风中海南省公路网络真实的的韧性曲线，以公路网络效率性能指标，展示了灾后海南省公路网络性能随时间的变化情况，并通过历史经验函数拟合得到公路网络在台风灾害下的拟合韧性曲线。发现路网性能的恢复并不是连贯而是阶段性的，大部分中断公路会在较短短时间内经过抢修恢复通车，但也存在一些需要长时间修复的中断路段，从而导致路网性能长时间处于某一状态而不能及时恢复到正常状态。

（2）基于登陆海南省的台风造成的公路破坏数据和台风强度数据，构建公路脆弱性模型，量化台风强度和公路受影响程度之间的关系。模型的脆弱性指标包括公路破坏概率、公路损失比和公路破坏长度比。台风强度指标包括台风过程累积降水、台风过程最大风速和风雨联合效应。对比不同台风强度的脆弱性模型发现，降雨和风速联合作用下的脆弱性模型比单独的降雨或风速模型拟合效果更好，但是无论是降水模型还是风速模型都可以以令人满意的精度估计公路的破坏概率、破坏长度比和损失比。对比不同公路等级，发现高等级公路的脆弱性低于低等级公路。

（3） 基于2012-2019年间台风灾害对海南省电力系统造成的破坏数据，基于lognormal CDF函数构建了海南电力系统对台风灾害的脆弱性模型。模型提供了台风灾害致灾强度和海南省电力系统停电用户比例之间的定量化关系，分别建立了台风过程累计降雨强度-停电用户比例脆弱性模型、台风过程最大风速-停电用户比例脆弱性模型和风雨联合作用下-停电用户比例脆弱性模型。结果表明，风雨联合作用下的脆弱性模型拟合效果最优。

（4）在构建了脆弱性模型基础上，结合台风灾害模拟事件集，评估了海南省公路和电网面临的台风风险。结果显示，在台风灾害作用下，海南省公路的年均损坏长度为108.9公里，年期望直接经济损失约为1.324亿元，其中县道风险最大，为4540万元，占总经济损失的34.3%。单场台风事件有10%的概率公路物理破坏长度超过109千米有10%的概率公路直接经济损失超过1.8亿元。海南省平均每年累计约有32万户电网用户停电，约占全省电网用户总数的10.9%。平均每年需要投入1.1万人次的抢修人员用于恢复电力系统的供电，平均每年需要调用2646车次的抢修车辆用于抢修电网破坏。单场台风事件有10%的概率电网停电用户数量超过56万户。

China is one of the countries most severely affected by typhoons in the world. In recent years, typhoons occur frequently in the southeast coastal areas of China, causing serious damage to the urban critical infrastructure system. Key infrastructure is an important guarantee for the normal operation of various functions of a city. Once damaged, it will not only cause huge economic losses, but also greatly affect the normal life of people and the normal operation of society. Therefore, it is of great significance to master the damage characteristics, recovery rules and risk assessment of key infrastructure under typhoon disaster for prevention and mitigation of typhoon disaster risk, effective post-disaster recovery and rescue and construction of resilient cities.

Taking hainan province as study area, based on the historical typhoon damage to roads and power grid caused by the loss of hainan province and the restoration of data, build the roads and power grids of hainan province under the typhoon disaster vulnerability model and restorative model, and based on the typhoon disaster simulation event set, evaluate the roads and power grids of hainan province under the typhoon disaster resilience and risk. The main results are as follows:

(1) based on the history of typhoon disasters in hainan province road block and restoring the recorded data, to get each typhoon in hainan province highway network real toughness curves, performance indicators at road network efficiency, shows the post-disaster hainan province highway network performance with the change of time, and through history function fitting road network under the typhoon disaster resilience curve fitting. It is found that the recovery of road network performance is not consistent but phased, most interrupted roads will be restored to traffic in a short time through emergency repair, but there are some interrupted roads that need to be repaired for a long time, which leads to the road network performance in a certain state for a long time and cannot be restored to the normal state in time.

(2) Based on the highway damage data and typhoon intensity data caused by typhoons landing in Hainan Province, a highway vulnerability model was constructed to quantify the relationship between typhoon intensity and highway impact degree. The vulnerability indexes of the model include highway failure probability, highway loss ratio and highway failure length ratio. The index of typhoon intensity includes accumulated precipitation, maximum wind speed and combined effect of wind and rain. By comparing the vulnerability models of different typhoon intensities, it is found that the vulnerability model under the combined effect of rainfall and wind speed has better fitting effect than the model of rainfall or wind speed alone. However, both the precipitation model and wind speed model can estimate the failure probability, failure length ratio and loss ratio of highway with satisfactory accuracy. Comparing different highway grades, it is found that the vulnerability of high grade highway is lower than that of low grade highway.

(3) Based on the damage data of hainan power system caused by typhoon disasters during 2012-2019, the vulnerability model of Hainan power system to typhoon disasters was constructed based on LogNormal CDF function. Model provides a typhoon disaster to disaster intensity and the proportion of power system blackouts users of hainan province, the quantitative relationship between cumulative rainfall intensity respectively established the typhoon process - power user scale vulnerability model, the maximum wind speed of typhoon process - power user scale vulnerability model and under the joint action of wind and rain - power user scale model of the vulnerability. The results show that the fitting effect of vulnerability model is the best under the combined effect of wind and rain.

(4) On the basis of constructing the vulnerability model, combined with the typhoon disaster simulation event set, the typhoon risk of hainan highway and power grid was assessed. The results show that under the influence of typhoon disaster, the average annual damage length of highway in Hainan province is 108.9 km, and the annual expected direct economic loss is about 132.4 million yuan, in which the county road is the most at risk of 45.4 million yuan, accounting for 34.3% of the total economic loss. There is a 10% probability that a single typhoon event will cause physical damage to the highway over 109 km, and 10% probability that the direct economic loss of the highway will exceed 180 million yuan. On average, about 320,000 power grid users in Hainan experience power outages every year, accounting for 10.9% of the province's total power grid users. On average, 11,000 emergency personnel are invested to restore power supply to the power system each year, and 2,646 emergency vehicles are used to repair power grid damage each year. There is a 10% chance that a single typhoon event will result in power outages over 560,000 households.