Abstract:
Objectives Complex construction organization strategies such as parallel same direction and reverse boring for long railroad tunnel projects lead to sparse spatiotemporal distribution and large spatiotemporal overlap in the digital twin three-dimensional model. The existing spatiotemporal indexes for the continuous uniform spatiotemporal distribution characteristics can not satisfy the needs of efficient retrieval of time, mileage, and semantics in multi-dimensions, and they also restrict the efficiency of the digital twin application in real-time virtual and real mutual feeds.
Methods This paper proposes a multi-dimensional efficient spatiotemporal indexing method applicable to the digital twin model of railroad tunnels. First, a 3D tile data structure of railroad tunnels associated with time, mileage, and semantic multi-dimensions is designed. Second, an oriented bounding box corrective mechanism constrained by the global optimal solution is established, and then a multi-granular spatiotemporal index of railroad tunnel lines-tiles-components is realized. Finally, the effectiveness of the method is verified using typical long railroad tunnel twin model data.
Results The results show that the average value of the proposed method in the mixed retrieval time of time, mileage and semantic multi-dimensions is 135.44 ms, which improves the retrieval efficiency of the 3D spatial single dimension and the time-mileage mixed dimension comparing with the existing typical spatial and temporal index.
Conclusions The proposed method can effectively support the efficient spatial and temporal retrieval of the digital twin 3D model of railroad tunnels, and the application of high-performance visualization and analysis.
下载:
