【摘要】 近些年来塔式起重机安全事故频发,造成了较大的人员伤亡和财产损失。为降低塔式起重机安全事故发生的概率、辨识关键的诱发因素,根据塔式起重机安全管理相关的法律、安全技术规范和标准,采用系统分析方法对事故致因进行层次分解,并以194份事故调查报告为数据样本,构建了一种基于随机森林算法的塔式起重机安全事故预测模型,并以某塔式起重机较大坍塌事故为例对该模型进行了验证。结果表明:利用随机森林算法对塔式起重机安全事故等级和类型进行预测的准确率分别达到了0.9和0.8;导致塔式起重机安全事故的关键因素为专项施工方案不完备（P₂）、人员安全意识淡薄（H₁）、安全技术交底不充分（P₃）、安全生产检查不充分（P₄）和工人无证上岗（H₆）,实例验证结果显示该模型对塔式起重机安全事故等级和类型预测的准确度较高。更多还原

【Abstract】 Tower-crane safety accidents occur frequently in recent years, resulting in heavy casualties and property losses.In order to reduce the probability of tower-crane safety accidents and identify critical causes, according to relevant laws, regulations and standards, this paper adopts the system analysis method to decompose the accident causes hierarchically, and then takes 194 accident investigation reports as data samples to build a tower-crane safety accident prediction model based on random forest algorithm.Subsequently, an example of a major tower crane collapse is given to verify the model.The results show that the prediction accuracy of random forest algorithm for accident level and accident type is 0.9 and 0.8 respectively, and the critical causes of tower-crane safety accidents are incomplete special construction plan（P₂）,low safety awareness（H₁）,inadequate safety disclosure（P₃）,inadequate safety inspection（P₄）,and unqualified workers（H₆）.The case study shows that the model has a high accuracy in predicting the levels and types of tower-crane safety accidents.更多还原