【作者】 李敏；

【导师】 李岚；

【作者基本信息】 东北财经大学 ， 工程管理， 2016， 硕士

【摘要】 建筑业是我国的支柱型产业之一,对配合我国经济发展、改善人民生活质量有着非常重要的作用。但是,建筑行业在提高国家综合实力水平的同时也给各方各面带来了一些问题,特别是伴随着建筑项目的增加,产生的安全事故也逐渐增多。由于建筑行业有着项目周期长、体量大、各方参与者庞杂等一系列特点,其事故的产生相较于其他行业来说更加频繁,涉及的原因也非常多,而且建筑项目投入很大,影响也广,许多微小的错误可能会造成意想不到的重大损失。目前我国很多建筑项目参与者已经对安全管理有了一定的重视,例如施工企业逐渐开始看重安全意识培养,设立现场安全员来进行监督,并且对对施工过程中产生的安全隐患进行了及时的整改;国家、社会也有了一些管理机构来进行第三方监督;法律体系中也有了一些规章制度来对施工过程安全问题进行明确限定。但是,这些方法尚未形成全面系统成熟的安全管理体系,而且我国针对施工过程安全管理主要采用的方法是安全调查表和事后的原因追踪,不能够从开始时就对项目的安全生产进行监督管理,无法预防事故。因此本文从安全管理思想入手,将BP神经网络法运用于安全预警管理工作中,变定性为定量,建立了一套预警管理系统,以期能在建筑项目实施中实时掌握其安全状态,增强安全管理能力。本文首先介绍了安全管理和评价、预警理论现有的发展,阐述了相关的理论基础,包括安全经济学理论、事故致因理论和BP神经网络理论。然后对建筑工程的事故致因因素进行了分析归类,总结出一些安全生产问题的主要致因,将其作为安全预警指标体系的重要组成部分。同时,针对以往评价体系的不足之处,参考国内外评价体系的构建模式与方法,将影响建筑工程施工过程安全的诱因分成人、材料和机械、环境、管理四个类型,同时按照递阶层次搭建了21个评价元素的体系,并应用AHP法确定各指标权重。最后根据施工过程安全风险识别、评价、预警、反应的逻辑关系,建立了完整的建筑工程施工过程安全预警模型,并基于BP神经网络对预警模块的指标和结果建立了良好的对应关系,同时用实际工程案例证明了预警模型的适用性。本文的创新之处在于:(1)通过搜集相关资料及分析,同时结合专家意见及相关标准,确立了建筑施工安全预警评价指标体系,包括4个二级和21个三级指标,并给出了评分的相关依据;(2)建立了完整的安全预警系统图,分别从风险的识别、评价、预警和反应四个方面对系统的运作机制做了阐述;(3)根据文中建立的建筑施工安全预警体系,分析项目当前的情况并依据结果判断本项目的预警级别,给出相应的改进措施,同现场调查得到的各方结果进行对比论证后,进一步验证了预警模型的适用性。但是,由于安全预警评价指标体系的构建难以满足指标体系的相互独立完全穷尽的原则,本文还存在一定的不足,而且指标量化采用专家评分法也会由于各位测评人员理解的不一致出现不同程度的偏离。同时作者对BP神经网络的理解和使用还有很多欠缺,需要进一步研究和完善相关知识,相关的实证研究仅是在十三个项目安全数据的基础上进行的,其普遍适用性有待进一步探讨。本文也未对各个评价结果相对应的应对机制和应该采取的措施作出详细分析,在以后的研究工作中需要不断总结实际情况并对相应措施作出总结。更多还原

【Abstract】 The construction industry is one of the pillar industries in China,which plays a very important role in the development of China’s economy and the improvement of the quality people living.However,the construction industry not only improves the level of national comprehensive strength,but also bring some problems to the parties.Especially with the increase in construction projects,security incidents rating is gradually growing.Because the construction industry has a series of characteristics:a long project cycle,large size and complex of all participants,compared to other industries,it caused accident more frequently,and the reasons are more complex.Because the construction project needs lots of invest,it has a wide influence on the society,and many small errors may cause significant unexpected losses.At present,many of China’s construction safety management of the project participants already have a certain amount of attention,such as construction enterprises gradually began to value the security awareness training,the establishment of site safety officer to supervise,and the security risks in the construction process of timely rectification;the state and society also has some authority to the third party supervision;legal system also has some rules to limit the construction process of safety problems.However,these methods have not yet forormed a comprehensive system of mature safety management system,and our method for safety management of construction process is mainly used in tracking security questionnaire and afterwards,not from the beginning of production safety supervision and management of the project,no method of prevention of accidents.This paper starts from the safety management theory,the BP neural network method is applied to safety early warning management,quantitative qualitative change,establish a set of early warning management system,in order to real-time control the state of safety in construction project implementation,strengthen safety management capability.This paper firstly introduces the current development of the theory of safety management and evaluation and early warning,and expounds the related theoretical basis,including the safety economics theory,the accident causing theory and the BP neural network theory.Then it analyzes and classifies the accident causing factors of the construction project,and summarizes the main causes of the safety production problems,and it is the important part of the security early warning index system.At the same time,aiming at the deficiencies of the evaluation system,reference to domestic and foreign construction of evaluation model and method system,incentive process of construction safety,adult materials and machinery,environmental management,four types will influence,at the same time according to the hierarchical structures of the 21 elements of the evaluation system,and the application of AHP method to determine the index weight.Finally,according to the logical relationship between the construction safety risk identification,assessment,early warning,response,has established the construction process of complete construction safety warning model and indicators and results of the early warning module based on BP neural network has a good corresponding relation is established,at the same time with the actual engineering case demonstrates the applicability of the early warning model.The innovation of this paper lies in:(1)by collecting relevant information and analysis,combined with expert advice and relevant standards,to establish the safety early warning evaluation index system construction,including 4 two and 21 three grade indexes,and gives the basis of score;(2)establish a security warning system diagram complete,respectively from four aspects of risk identification,assessment,early warning and response mechanism of the system is discussed;(3)according to the construction safety early warning system established in this paper,the current situation and analysis of the project on the basis of the results of the project determine the warning level,corresponding improvement measures are presented,with field investigation the parties were compared after the demonstration,to verify the applicability of the early warning model.However,because of the establishment of safety early warning evaluation index system is difficult to meet the index system of independent MECE principle,this paper also has some shortcomings,and quantified by expert score method will be due to inconsistent evaluation of personnel understanding deviation degree.At the same time the author to understand and use the BP neural networks have some deficiencies,need further research and improve the relevant knowledge,the relevant empirical study was carried out in thirteen items based on data security,its general applicability for further discussion.In this paper,the corresponding mechanism and measures should be taken to make a detailed analysis,in the future research work need to constantly sum up the actual situation and the corresponding measures to make a summary of the corresponding measures.更多还原