Quantitative Analysis Of Human Factors In Building Workplace

The construction business’s growing is steady and rapid these years in China, promoting national economy’s continuous enhancement. Meanwhile, due to the instinct property of construction industry, such as long life-cycle of production, high intensity of work, complexity of construction techniques, the existing problems unceasingly highlights the production of construction engineering, various types of accidents emerge in an endless stream. Among all these accidents, that leading by human factor accounts for a substantive proportion.Depending on existing human reliability theory and human error causing theory, this thesis analyzes the typical characters of construction sets and operators. A human error causing model is constructed based on the analysis of Chinese construction working characteristics and behavior features of builders, analyzing the concrete flow of human error incident’s emergence in construction site by the relationship of human error’s inducing factors, engendering mechanism and representation.Directing against the two pivotal aspects of human reliability analysis, the identification of human error causing factors and human reliability quantitative analysis, this thesis proposes targeted human error inducing factors identification technique which is in accord with the characteristics of Chinese construction industry and causing factors’ interactive influence considered quantitative method of human reliability appraisal. At the same time, the human error data collecting and analyzing system based on the proposed human error analysis technique is also built to provide theoretical and technical support for the prevention and reduction of human errors in construction projects. Research on the human reliability quantitative appraising method for construction working site has definite theoretical and practical application value.Binary response model and dummy variable regression are introduced to quantitative evaluate the probability of the emergence of human error incidents. A logistic regression model is built to appraise human error probability in construction working site, Homser-Lemeshow test and Wald test are adopted to separately test the entire model’s significance and each variable’s significance among the model. Besides, a classified table is used to measure the predication accuracy of the built human error probability logistic model.Finally, taking advantage of Java Web techniques, the B/S architected construction working set human error data collecting and analyzing system is developed according to the structured human reliability quantitative appraisal method. This system separates model, view and controller using Struts2 framework, data persistent are made by object relationship mapping framework Hibernate, achieving aspect oriented programming and inversion of control via adopting Spring framework. By actualizing SSH architecture, development of each module in this system achieves "high cohesion, low coupling".