基于社会网络的地铁施工班组行为安全研究

熊超华

摘要：中国各个城市地铁建设速度加快，对于地铁施工工人的需求进一步上升，大量未经专业技能培训的工人进入施工一线，使得由于工人的不安全行为而引起的施工安全生产事故时有发生。工人的不安全行为控制已经成为地铁建设过程中一个重要的研究课题。本文以行为安全（Behavior-Based Safety，BBS）为基础，从社会网络视角着手对班组不安全行为进行研究，通过多维度行为识别技术，结合人类动力学方法、胜任力理论，观察、分析、矫正和评估班组不安全行为。具体研究内容入下:（1）地铁施工班组行为安全表现测量方法研究。首先，从地铁施工相关标准、技术方案以及事故案例分析提炼出地铁施工过程中存在的不安全行为，构建包含各工种、各施工阶段的地铁施工工人不安全行为知识库；其次，采取多维度不安全行为识别方法和技术，采集施工现场班组作业人员不安全行为数据；最后，为定量描述施工班组行为安全表现，综合考虑班组规模和不安全行为发生频率因素，提出施工班组行为安全表现测量指标。（2）地铁施工班组行为安全表现规律研究。首先，利用人类动力学的研究方法对班组行为安全表现进行时间维度统计分析，发现班组不安全状态时间间隔分布具有重尾特征，服从幂律分布。然后，从社会交互角度对班组行为规律进行解析，根据构建出的工人群体社会网络结构的特征分析，发现工人群体社会网络具有小世界和非标度特征。（3）地铁施工班组不安全行为网络干预研究。首先，根据构建出的班组社会网络，研究各个工种班组社会网络结构特征，发现各班组均存在网络中心节点以及凝聚子群。然后，根据班组社会网络节点度和节点相似性对班组成员重要性进行排序，识别出班组社会网络行为干预的关键人员。最后，在不安全行为频发的模板工班组和架子工班组进行了社会网络行为干预研究，通过随机对照试验发现，针对关键人员的行为干预方式相比较传统的行为干预模式矫正率更高，并且网络行为干预后的安全表现更好。与此同时，通过关联规则分析还揭示了两种类型的关键人员。（4）地铁施工班组关键人员安全胜任力研究。评估关键人员安全胜任力能够更有针对性进行社会网络行为干预。基于胜任力理论，结合探索性因素分析，验证性因素分析和结构方程模型，建立了地铁施工班组关键人员的安全胜任力模型。模型构建结果表明，存在四个维度因素对关键人员的行为安全有显著影响，即安全相关的操作资格、工作态度、自我调节和适应性以及个人性格和倾向。本研究将社会网络分析方法应用于班组行为安全研究中，从群体视角探讨地铁施工现场不安全行为控制，实现地铁工人群体行为数据的全样本采集，初步探索班组不安全行为规律，提出网络行为干预方法，构建关键人员的安全胜任力模型。本文研究内容有助于以群体角度深入了解班组行为模式，增强行为干预效果，提升组织安全管理效率。

关键词：地铁；施工班组；社会网络；行为规律；胜任力模型

ABSTRACT

With the rapid development of metro construction in various cities in China, the demand for metro construction workers is further increasing. A large number of workers without professional skills training enter the construction field, which makes the construction safety accidents caused by workers' unsafe behaviors happen from time to time. Workers' unsafe behavior control has become an important research topic in the process of metro construction. Based on Behavior Based Safety (BBS), this paper studies unsafe behaviors of construction teams from the perspective of group with social network method. It observes, analyzes, corrects and evaluates unsafe behaviors of construction teams through multi-dimensional behavior recognition technology, human dynamics method and competency theory. The specific research contents are as follows:

(1) Research on measurement methods of behavioral safety performance of metro construction teams. Firstly, a knowledge base of unsafe behaviors of metro construction workers including various types of work and construction stages is constructed, which extracted based on the analysis of metro construction related standards, technical schemes and accident cases. Secondly, multi-dimensional behavior identification methods and techniques are adopted to collect the behavior data of construction team members. Finally,put forward behavior safety performance measurement index of construction team to quantitatively describe the safety performance of the construction team, which comprehensive considerate the size of the team and the frequency of unsafe behavior.

(2) Research on behavior safety performance law of metro construction team. Firstly,the human dynamics research method is used to analyze the behavioral safety performance of the team. It is found that the time interval distribution of the insecure state of the team has a heavy tail characteristic and obeys the power law distribution. Then, from the perspective of social interaction, the behavior patterns of the team are analyzed. According to the characteristics analysis of the social network structure of the construction teams of each type of work, it is found that the social networks of each team have small world and non-scale characteristics, which further explains the behavioral performance of the team.

(3) Research on social network intervention of metro construction team behavior.Firstly, the social network structure characteristics of each construction team are studied. It is found that there are network center nodes and cohesion subgroups in each team. Then,according to the importance of the social network node degree of the team and the similarity of the nodes, the importance of the team members is sorted, and the key personnel of the social network behavior intervention of the team are identified. Finally,the behavioral social network intervention was conducted in the template work group and the shelf work group with frequent unsafe behaviors. Through randomized controlled trials,it was found that the network behavior intervention for key personnel is more effective than the traditional intervention model. At the same time, two types of key personnel were revealed through correlation analysis.

(4) Research on the construction of Safety Competency Model of key personnel in metro construction team. Assessing key personnel safety competencies enables more targeted behavioral social network interventions for key personnel. Based on the competency theory, combined with exploratory factor analysis, confirmatory factor analysis and structural equation model, the safety competency model of key personnel in the metro construction team was established. The results show that there are four dimensional factors that have a significant impact on the behavioral safety performance of key personnel, namely safety-related operational qualifications, work attitudes,self-regulation and adaptability, as well as personal traits and tendencies.  Social network method is used to explore the behavior safety of construction teams from the perspective of groups in this study. Firstly, the full sample collection of group behavior data of metro workers is realized. Secondly, the unsafe behavior law of construction team is preliminarily explored. Thirdly, the network behavior intervention method for key personnel is proposed. Finally, safety competency model of key personnel is built. The research content of this paper is helpful to understand the construction team behavior mode from the perspective of group, enhance the effect of behavior intervention,and improve the efficiency of organization safety management.

Key words: Metro; Construction team; Social network; Behavior law; Competency model