Representing a Digital Twin City Model Using Open Source Tools and Integrating It With Dynamic Sensor Data

Introduction

The results achieved in 2016 thanks to the European 7th Framework Programme (EU FP7) District Information Modelling and management for Energy Reduction (DIMMER) project (Osello, 2016) showed that was possible and effectively useful the integration of building data at city scale with energy-related data from sensors at building and district level: with a wise choice of representative use cases the results could be extended in the near districts or to similar pilot cities to compute and simulate energy consumptions / savings and effects on emissions, so that appropriate decision and strategies could be adopted at administrative level by competent authorities.

The relation between sensors and buildings in the case of the Turin pilot, and overall electricity consumption in case of the Manchester University Campus was defined at level of the Building Information Modeling (BIM) model: every representative model had one or multiple readings of energy or electricity consumption that could be compared with the volume of the construction and its materials.

The objective of the chapter is to show that the adoption of City Geography Markup Language (CityGML) as descriptive format of the district can improve the capabilities of relating the buildings with energy production, distribution and consumption networks: the new Dynamizers concept in the forthcoming definition of CityGML version3 can also make possible to simplify the computations at district level by reducing them at the scale of every single building. The previously stated relation between sensors and buildings can be explicited in a more detailed CityGML format of the building in which geometry, materials and sensors placement are included in the same model.

Methodology

The task of representing cities in a 3D environment is challenging. Indeed, the game industry achieved excellent results by adopting dedicated and costly rendering engines (Andrade, 2015). Public authorities have access and own a large variety of datasets such as cadastral databases, aerial and satellite imagery, local and small to medium-scale laser and drone scans. While the former are usually displayed in a lot of bidimensional Geographic Information System (GIS) applications and frameworks, the latter generate complex points clouds and tridimesional structures that requires dedicated and specialized viewers and editors. All these data differ from extension to form and resolution, and one of the complex tasks is to integrate them in a simple and usable way for the end-user. The tridimensional representation of the city and of the environment allows for effective thematization and classification of various phenomenona in dense urban environments.

Here below a list of data sources starting from the simplest to the more detailed, not necessarily exhaustive, is provided. They must be harmonized and displayed at different scales defined by the user according to the concept of progressive Level of Detail (LOD) (Biljecki, 2016). The OGC CityGML’s classification (Gröger, 2006) defines the level of detail of data considered from the simplest, bidimensional LOD0 to the complete LOD4.It is worth considering that the city, as CityGML specification suggests, is made not only of buildings but also from roads, street signs, trees, rivers and all sorts of anthropic and natural elements. Thus, the same principles of twin models apply not only to the city but to the open country and to landscapes, too.

After an evaluation of the visualization tools available, the steps needed to convert and adapt the various kind of datasets to the streaming and rendering of possibly huge geospatial content are considered.

In the results section, then, examples of conversion of BIM models into 3D Tiles format are presented, and a final example section shows different representation of sensors’ definitions as CityGML Dinamizers.