基于Multi-agent的隧道变形监测系统

Tunnel Deformation Monitoring System Based on Multi-agent Technique

研究目的:为消除传统隧道监测维护中低频率、长周期、高费用等缺点,适应隧道即时监测护养的新兴要求,本文通过对上海地铁变形预测研究,提出一种基于多智能体的隧道变形在线监测系统。研究结论:(1)在监测系统中提出一种基于分层组件结构设计的通用Agent模型,该模型包括客户感知层、智能业务层与服务效应层,并以粗细两种化度归纳描述智能体内部结构;(2)依据Agent通用模型及隧道变形监测自身特点搭建具有6层复合结构特点的隧道变形监测系统,提出基于熟人协作沟通机制,实现系统的松耦合和低负载构造;(3)利用上海地铁变形实测数据验证预测模块Agent的工作效能,实验表明该系统具有较强的自治性、合作性,并具有一定的实用价值,对隧道监测系统的搭建有一定的指导意义。

Research purposes： In order to eliminate low frequency and high cost factors in traditional tunnel monitoring and maintenance, a new tunnel deformation online monitoring system based on multi - agent is proposed. It can adapt to the new requirements for real - time monitoring maintenance. Research conclusions：（ 1 ） Based on the hierarchical component structure design idea, this paper presents a general agent model. The model consisted of three layers, which were customer perception layer, intelligent service layer and service effect layer, respectively. The internal structure of the agent was descripted separately in coarse - grained and fine -grained. （2） According to the characteristics of tunnel deformation monitoring and the general agent, a tunnel deformation monitoring system has been set up using acquaintance cooperation communication mechanism, with six - laye （3） The experimental results showed that the system had strong autonomy, cooperation and a certain practical value, and may give references to the construction of the tunnel monitoring system.