基于Web的管廊可视化信息管理系统设计与实现

城市化进程促进了城市地下综合管廊的飞速发展，但目前管廊的巡检等运维管理工作大多以人工为主，且信息化水平低下，而且管廊内的管线复杂、人工运维管理效率低且危险隐患高。文章提出一种基于Web的管廊建筑结构与建筑信息轻量可视化及CO环境数据可通过浏览器在线训练与预测的系统。通过Revit二次开发技术将建筑信息模型轻量化解耦并通过WebGL技术将几何模型加载在浏览器前端页面，后台服务程序可以通过通信线程池将终端节点采集到的环境数据存入Mysql数据库，并与浏览器前端动态耦合。通过整合Tensorflow.js可以在浏览器前端实现基于LSTM算法的CO浓度预测模型在线训练与预测。经过测试，设计的系统工作可靠，满足设计需求。

Design and Implementation of Utility tunnel Visual Information Management System Based on Web

Urbanization promoted the rapid development of Utility tunnel. However, there are some problems in the management of Utility tunnel currently such as management mean depends on manual, the complicated pipelines in Utility tunnel, the low efficiency of manual management and the high risk factor. The article puts forward a web-based Utility tunnel visualization information management system based on web to realize the lightweight visualization of building structure and building information, online prediction and model training of CO environmental data in web browser. Management system uses the secondary development of Revit to decouple the BIM model and loads the geometric model in the web page by WebGL. The background services puts the environment data which is collected by the terminal nodes to mysql database through the communication thread pool, and dynamically couple with the browser web page. Integrating Tensorflow.js to implement online training and prediction of CO based on LSTM in the browser. According to the test, management system works reliably and meets the design requirements.