Research On The Intelligent Management Of Highway Construction Security Risk

Currently, the number of accidents, the loss of life and injury in the construction are the second greatest after mining. As a high-risk sub-sector, the safety situation of highway was extremely severe while it rapidly expanded during the "Eleventh Five-Year" period. And the severe losses have badly hurt the healthy of highway construction. Therefore, the Government issued a series of macro security notices in order to strengthen supervision, prevent accidents and improve emergency response capabilities, and the Government also required security management to achieve five changes. So the thesis focuses on some key technologies of highway construction safety risk management, and the following aspects are devoted to the main effects:(1) Intelligence, networking and informatization are the future research trends of highway construction safety management. This article focuses on the intelligent application in construction site to solve practical problems. Combined with practical features and security status of highway projects, this article details the theory superiority and application feasibility of some intelligent algorithms such as case-based reasoning, rule mining, support vector machine and variable fuzzy sets, etc.(2) Data preprocessing holds an important position in the artificial intelligence field, accounting for 60% of the entire process of intelligence. One of its main purposes is to provide data preparation for algorithm application, and the other is to prevent data sets unbalanced which might lead to errors. And data conversion, normalization and dimension reduction are the three important pretreatments. In the highway construction process, there are filled with many index types, such as subjective, objective, qualitative and quantitative indicators, on the other hand, different intelligent algorithms ask for different data types, so this article researches through actual cases study to prove the importance of data pretreatment in safety decision analysis.(3) Case-based reasoning is an analogical reasoning technology reference to similar project experience. Due to its own characteristics, the accident rate of highway project is higher than the general construction projects, plus construction information accumulated of China is quite scarce, and most of the problems are unstructured forms, so that security decisions making relied more on subjective experience. However, subject to the experts' background knowledge, decision-making always has some bias, which leads to insufficient early risk prediction, poor risk control in construction, emergency start slow, and inflexible plans, etc. To improve the effectiveness of security decision-making to these non-programmable problems, the best solution in theory is to retain. So based on this background and the theoretical groundwork, this article emphasizes that case-based reasoning is the best way to solve current security decision-making problems, and it is also the basis to achieve security management change. Combined with the data features and security requirements management of highway project, this article puts case representation and case retrieval the two key technologies of case-based reasoning into practice, and establishes a security intelligent management framework for risk prediction, early warning and emergency response of highway construction.(4) According to security risk management purposes, the database of case-based reasoning is set into five modules in this paper, in order to work on risk prediction, early warning and emergency response, etc. The whole workflow includes risk prediction based on project characterized, accident prediction based on project risk, risk causes query based on accidents, pre-control plans optimization based on risk levels and emergency response plans optimization based on accident levels. In the implementation process, this article uses case-based reasoning technology to store data, establish the appropriate retrieval mechanisms, and mine risks-accidents association rules in the finally; variable fuzzy weight model of subjective and objective to calculate indicators weight; variable fuzzy clustering model to cluster the existing project features, risk characteristics and accident characteristics; support vector machine algorithm to recognize risk level and accident level of the new project; quantitative and qualitative criterion model to monitor dynamic risk indicators in the construction process; variable fuzzy optimization model for multi-objective optimization of risk prevention plans and emergency response plans.(5) Combined with the five modules of case-base and security decision-making requirements, this article designs a C/S system, which takes VB as the interface platform, Access database as the foundation and Matlab as the mathematics programming. The system has solved the technical interaction problems among VB, Access and Matlab, and also achieved the algorithms function and results visualization, such as similarity search, pattern recognition based on support vector machines and variable fuzzy sets. Based on that, this system also could intuitively provide decision-making basis for security management.