论文部分内容阅读
以内蒙古茅荆坝隧道为工程背景，在广泛获取国内外地下工程施工专家知识的基础上，利用现代计算机技术对开挖过程力学行为进行了研究。 首先，开发研制了非线性有限元程序ＵＳＥＰ２Ｄ，可视化地模拟了开挖前后的力学行为、不同开挖步序对隧洞稳定性的影响，同时也可对智能辅助决策及系统决策方案进行验证和准定量分析。 其次，采用面向对象的编程技术，提出并开发了施工智能辅助决策系统。该系统含６个子系统和１个公共子系统，即施工方法决策子系统、支护设计子系统、量测预报子系统、工期控制子系统、钻爆设计子系统、施工事故处理系统和围岩分类子系统。同时对施工智能辅助决策系统的知识获取问题进行了探讨，总结并利用现有的专家知识，建立了隧道工程围岩分类、施工方法、支护设计等３个知识库。 另外，探讨了遗传神经网络方法在开挖过程位移演化规律预测预报中的应用，用遗传算法优化神经网络结构，首次将遗传算法与神经网络结合起来用于预测、预报施工过程中围岩位移演化规律，使优化后的结果为全局最优，对下一步施工时围岩位移进行了预测，并与实际工程进行了对比分析。 最后，工程实例中将决策系统和ＵＳＥＰ２Ｄ程序应用于茅荆坝隧道工程的辅助决策和优化。
Based on the Maijingba tunnel in Inner Mongolia, based on extensive knowledge of underground engineering construction experts both at home and abroad, the mechanical behaviors of the excavation process are studied by using modern computer technology. First of all, the nonlinear finite element program USEP2D has been developed to visually simulate the mechanical behavior before and after excavation, the influence of different excavation steps on the tunnel stability, as well as to verify and approve the intelligent decision support and system decision-making plan Quantitative analysis. Secondly, the object-oriented programming technology is used to propose and develop the construction intelligent decision support system. The system consists of six subsystems and one public subsystem, namely construction method decision subsystem, support design subsystem, measurement and forecast subsystem, construction period control subsystem, drilling and blast design subsystem, construction accident processing system and surrounding rock Classification subsystem. At the same time, the problem of knowledge acquisition of construction intelligent decision support system was discussed. Three knowledge bases of tunnel engineering surrounding rock classification, construction method and support design were established by using existing expert knowledge. In addition, the application of genetic neural network method in predicting and predicting displacement evolution of excavation process is discussed. The genetic algorithm is used to optimize the structure of neural network. For the first time, genetic algorithm and neural network are combined to predict and predict the displacement of surrounding rock during construction The results show that the optimized result is the global optimum, and the displacement of surrounding rock during the next construction is predicted, and compared with the actual project. Finally, the decision-making system and USEP2D program are applied in the engineering decision-making and optimization of Maojingba tunnel project.