Construction of Multi-granularity Spatial and Temporal Entities for Groundwater in Polluting Enterprises Based on Digital Twin Technology

Abstract. The adequate supervision of groundwater pollution in industrial enterprises primarily focuses on post-pollution treatment, and there is an urgent need to develop methods that go beyond merely describing pollution conditions and thoroughly understand the characteristics of pollution entities. In response to the current challenges of groundwater pollution visualization and entity modeling, based on digital twin technology and multi-granularity spatiotemporal object models, this study innovatively proposes an entity modeling framework that integrates the spatial, temporal, and behavioral attributes of the enterprise-groundwater system, clarifying the definition and division basis of multi-granularity levels to support decision-making requirements at different scales. Taking the groundwater pollution of a specific chemical plant in Huizhou as an example, a digital twin of groundwater pollution with dynamic data assimilation capabilities for this site was constructed based on borehole data, historical monitoring records, and the deployed sensor network. The results demonstrate that the digital twin system enables the visualization and simulation of pollution dynamics by integrating near-real-time data. The multi-granularity spatio-temporal object model successfully captured the spatial heterogeneity, temporal evolution characteristics, and polluter-receptor interaction behavior of pollution. This research not only provides a more refined and dynamic cognitive tool for groundwater pollution but also lays scalable technical support for integrating multi-granularity spatiotemporal object models with digital twin technology for environmental supervision.