Research Of The Collaborative And Life-Cycle Maintenance Based On BIM

In recent years,with the continuous application of BIM in construction projects,the data sharing and management in construction projects have undergone tremendous changes,and BIM has become a key approach to the informatization of construction industry.BIM should be integrated into the entire life process of a construction project,from architectural design to construction and post-construction operation and maintenance.In the life cycle of a project,project collaborators often obtain highintensity cooperation from relevant BIM models based on their needs in different fields.However,as model data continues to expand,traditional version-management systems have found it hard to support BIM Application status.Therefore,there is an urgent need for an efficient BIM versioning mechanism to support collaborative work throughout the building life cycle.At the same time,the development of BIM lacks the full life cycle of the construction project,including the stages of design,construction,operation and maintenance,a unified platform or series of tools,and how to provide full life cycle support for construction projects has also become an urgent problem to be solved in the development of BIM.First of all,in order to support high frequency data interaction in the collaborative work,a distributed version control mechanism is proposed,supports the partial-model collaborative design with the following three designs.An operation tracking mechanism is proposed to record the user's design applied to the model.It realizes the version control by providing functions such as the operation concatenation,operation reversal,etc.In order to support distributed collaborative design,a partial-model layer,in which different partial model can be revised independently.In order to keep the consistence between central model and partial models,a version synchronization mechanism is proposed that achieve the models' global coordinative synchronization.The system has been implemented on the BIMServer(an open source platform).Compared with Git,the system has obvious advantages in many major metrics: partialmodel support,network transmission,resource consumption and version synchronization time.Secondly,in order to support the model unification in different stages of the construction life cycle,a BIM-based state-of-the-building detection mechanism for building facilities is proposed to realize the state identification of building facilities based on convolutional neural network and realize the operation and maintenance phase of BIM in the construction life cycle Support for BIM technology and deep learning combined.By YOLO algorithm through increase the max pooling,increase the learning rate and change the batch size of the three improvements,to reduce the model training and recognition time,reduce the model hardware resources,and implemented on Tensorflow.Judging the state of building facilities by using the images of building facilities can help to accurately locate the problems during the operation and maintenance phase and reduce the resources needed in the traditional operation and maintenance process,which helps to improve the use of BIM models in the construction life cycle effectiveness.Finally,to achieve resource isolation and rapid deployment of project modules,a container-based project deployment was proposed.Packing the project code with running environment in the container contributed to bind the container with the project module,and the isolation of the physical resources between services.Orchestrate the environment by using the Rancher container to realize functions such as service rolling upgrade and online expansion in the cluster.Container-based project deployment solutions help to refine the service granularity,efficient use of cluster physical resources and improve the system's robustness.