摘要

近几年随着地铁工程的大规模建设，安全事故也频繁发生，施工前的风险识别显得尤为重要。目前地铁工程风险识别的难点包括：专家数量少、专业技术人员少、主观性强、大量隐性知识未被集成利用等，为了克服风险识别的难点并提高风险识别结果的准确性，本文运用基于数据驱动的风险识别方法建立了地铁工程风险识别知识库系统，以期为地铁工程风险识别提供一种准确、高效的识别途径。

本文针对地铁工程的四个主要特点，将地铁车站工程的基本参数划分成了6类33个因素，又根据对国内外文献的总结确定了基坑开挖过程中常见的5种主要风险事件，并建立了风险事件与基本参数之间的关系。依据确定的风险事件与基本参数，文章引入了风险识别知识库的概念，介绍了本文风险识别知识的概念、风险识别知识的表示方法及风险识别知识的获取方法，结合文章引用大量数据的特点提出将支持向量机作为获取风险事件与基本参数之间隐性知识的推理机制。在风险识别知识库的基础上文章建立了风险识别知识库系统，并给出了系统结构设计、数据库形式、风险识别过程设计。最后，以武汉地铁7号线湖北大学站为例介绍了风险识别的过程，并得出了风险识别的结果。

本文将知识库、数据驱动的概念引入到地铁车站的风险识别中，总结了影响基坑开挖安全性的33个参数，归纳了地铁施工常见的5种风险事件，建立了地铁工程风险识别知识库系统。本文在地铁工程风险识别方面具有创新性。

关键词：地铁车站明挖法风险识别知识库数据驱动

Abstract

In recent years, alongwith large-scale subway construction projects, safety accidents occurfrequently, risk identification before construction become veryimportant.Currently,the risk identification of subway projects have somedifficulties which include: a small number of experts, fewer professionalpersonnel, subjectivity, a lot of tacit knowledge has not been integratedutilization and so on. In order to overcome these difficulties and improve theaccuracy of results of risk identification,this paper establish a subwayproject risk identification knowledge base system which based on data-driven toprovide an accurate and efficient way for Metro risk identification.

In this paper, the riskfactors of subway station are divided into 6 categories and 33 factorsaccording to the four main features, and this paper identified five major riskevents which based on the literature and established a relationship betweenrisk events and risk factors.Then, the paper introduces the concept of riskidentification of knowledge base, knowledge representation and knowledgeobtain, combined with the characteristics of citing a large number of dataproposing the SVM as a mechanism for obtaining tacit knowledge between riskevents and risk factors.On the basis of risk identification knowledge base, thepapere stablish a Subway Station Risk Identification knowledge Base System,andgive the design of system structure and risk identification process.Finally,put Hubei University Station as an example to describe the risk identificationprocess and obtain the results of risk identification.

The paper introduce theconcept of knowledge base and data-driven into the risk identification ofsubway station, summarizing 33 parameters that affect the safety of theexcavation and 5 common risk events, establishing a Subway Station RiskIdentification Knowledge Base Systems. This article is innovative in the riskidentification of subway station.

Keywords: Subway station   Open-cut method   Risk identification  Knowledge base  Date driven