| The rapid development of China's economy urges the process of urbanization,but it also breeds "urban disease".As a good solution,the urban comprehensive management corridor emerges at the right moment.In recent years,most of our country's pipe corridors have entered or are about to enter the operation and maintenance stage.The application of building information model technology(BIM)can effectively improve the information level of operation and maintenance management of the pipe corridors,with the aim to deal with the complex distribution of pipeline.But the existing BIM applications are difficult to be applied to the operation and maintenance stage of the pipe corridors due to the software dependence and strong professionalism.This paper makes full use of the advantages of BIM model of two-dimensional and three-dimensional integration and multi-source data heterogeneity.WebGL technology is also applied to improve the current application situation of BIM Technology in the operation and maintenance stage of the corridor.On this basis,this paper integrates the development advantages of web technology with high degree of freedom and the deep learning technology of B/S architecture,database technology and other technologies to realize the management system of pipe gallery operation and maintenance.The management system is based on BIM model and equipped with the function of cross platform,visualization,intelligence and information.The main work of this paper is as follows:(1)To research the method of digital analog separation and lightweight of BIM model.This paper analyzes the structure of BIM model of pipe gallery,developsplug-ins to reconstruct BIM model on the premise of graphics geometric entity standard and omniclass classification principle.Plus,it applies vertex reuse and approximate curvature edge folding algorithm based on building classification to further reduce the model volume;(2)To research the lightweight rendering engine.By using WebGL technology and taking the digital analog separation product as the input object,this procedure is carried out in the following three steps.Firstly,the 3D scene of Revit and webg is differentiated and adjusted.Secondly,the object-oriented model is classified and encapsulated based on the model component units.The scene is managed by octree,and the scene camera controller,particle system simulation warning and digital analog binding are designed and studied.Finally,the optimized octree based view cone removal is proposed.The algorithm improves the real-time rendering ability of the engine to ensure the stability of the engine rendering.(3)To design and create a maintenance parameter prediction platform based on B/S.Firstly,it analyzes the similarities and differences between the client oriented deep learning mode and the traditional server mode deep learning platform.Secondly,it analyzes the architecture mode and framework selection and realizes the early warning platform.Finally,combined with the characteristics of operation and maintenance data,it selects the prediction model,and determines the network structure of one-dimensional convolution neural network and Gru stack through comparative experiments.Additionally it combines the early warning platform to predict the operation and maintenance parameters at different points in time.(4)The system design and analysis are described in combination with such principles as the software design principles,database design principles,interface design principles to show the realization effect of the key functions involved in the system.The non-functional indicators are used to test and analyze the system with cross platform,compatibility and responsiveness to verify the feasibility of the system. |
PDF Full Text Request