Earlier Warning Control Model And Simulation Research Of Safety Risk For Large Scale Engineering Based On Control Theory

The large scale engineering is an integrated complex giant system. Due to complexity of the internal structures and external interfaces, and uncertainty factors of the construction process, it often lead to significant disaster accidents, which will bring massive loss to the enterprise and even serious harm to the national public safety. Currently, researchers at home and abroad are mainly aimed at understanding of safety risk, identification of safety risk factors and assessment of safety risk, by using education, procedural and engineering method to mitigate safety risks. In order to improve the large scale engineering safety risk management level and promote systematic safety risk management, It is needed that establish an scientific, quantitative, automatic, and intelligent safety risk early warning and control platform. So this paper mainly studies from the following aspects based on typical characters of large scale engineering:Dynamic character is one of the typical characters of the large-scale engineering. In practice, it is difficult to give a quantitative warning to these dynamic early warning safety risk indicators. In this paper, upper-down control line in quality control charts is referenced and three division model based on optimal division model is established. Firstly, by dynamic programming method, give a classification to sample data, which is the observation values of the early warning safety risk indicators; then calculate the quantification warning threshold of safety risk by numerical iteration steps; finally draw the safety risk warning control chart based on the green warning threshold and yellow warning threshold. The result shows that the optimal division model can alarm to abnormal data effectively which does not need assume the statistics distribution of observation data.In order to avoid frequent mobility problem of constructionengineers, this paper establishes a balanced staff entropy model in the view of the construction planning. In the model, taking into account the long duration and uncertainty factors of the large-scale engineer, updates the existing model by normalization and adding buffering time, then optimize the overall construction plan by particle swam optimization. Finally, based on the above output, learning from the idea of rolling optimization of model predictive control, go on balance the staff during the fluctuation period, by using gradient descent method of the maximum theory. By examples, the model shows it makes the overall staff configuration achieve the optimal balance, and also verify the construction plan given by this model is robustness.Cross-operation is another characters of the large-scale engineering, in order to avoid the safety risks resulted from it, this paper uses the relative entropy to quantify the safety risk of cross-operation. Firstly, according to the permutations of devices, calculate the probability of cross-operation accident; and then quantify the safety risk of cross-operation referring to the relative entropy model which come from thermodynamics; finally solve the model by the particle swarm optimization algorithm. The cross-operation safety risk is relative to the non cross-operation safety risk, it is clearly showed by examples, and the goal is achieved to eliminate the safety risk resulted from cross-operation by establish the optimization model based on cross-operation relative entropy model.The current large-scale engineering will set a quantitative safety goal, such as "No deaths, no serious injuries, and minor injuries less than 0.6 ‰". To achieve this quantitative objective, Based on the existing research on the basis of the qualitative theory, this paper takes Yanfang Line of Beijing Subway construction project as an example to give systematic analysis and verify all the models provided by this paper. It include verify three division model based on optimal division model, which need indentify dynamic safety risk warning indicators in the all safety risk warning indicators. It also include verify safety risk entropy quantification model, which need establish optimization model based construction plan and solved it by optimal control algorithm. In general, in order to improve the level of safety risk management, use control theory methods to solve large-scale engineersafety risk warning and control model, which give aactive exploration and research about safety risks of quantitative earlier warning and quantitative control.