The research activities of the Leonhard Obermeyer Center cover a broad range of digital methods which are being developed for sensing, analyzing and shaping the built environment. They are driven by the scientific expertise of the individual chairs as well as by joint research projects. A particular emphasis is put on researching digital methods which are able to bridge the multiple scales involved with modeling and simulating the built environment – from the scale of a few millimeters to the scale of a whole country.
In detail, the research fields covered by the Center’s chairs include:
- Building Information Modeling, Infrastructure Modeling, Spatial-temporal Analysis, Pedestrian Dynamics, Construction Process Simulation (CMS)
- Geospatial Information Modeling, Geographic Information Systems, Spatial Databases and Data Infrastructures, Indoor Navigation (GI)
- Computer-Aided Architectural Design, Digital Design Methods, Case-Based Reasoning, Digital Fabrication (AI)
- Numerical methods, Finite Element Method, interactive and real time structural simulation, multi-scale and multi-physics problems (CIE)
- Photogrammetry, Remote Sensing, Image and Point Cloud Analysis, Computer Vision (PRS)
Urban Information Modelling + Smart Cities & Regions
In this student project a semantic 3D city model of New York City (NYC) has been created based on datasets provided in the NYC Open Data Portal. Different 3D feature types were derived from existing public 2D and 2.5D datasets using spatial and semantic transformations together with (some) photogrammetric analyses. The resulting 3D city model is represented in a homogenized and integrated way using the international standard CityGML of the Open Geospatial Consortium (OGC).
In order to test and ensure the safety of automated and connected driving, a virtual test field is required which enables the efficient simulation of a vast amount of traffic scenarios. The objective of this research project is the development of a multifunctional regional model. Therefore a digital twin of the city Ingolstadt is created and infrastructural elements (traffic lights, induction loops, …) as well as floating car data are connected to the city model.
CityGML is an international standard issued by the Open Geospatial Consortium (OGC) on the modeling, storage, and exchange of semantic 3D city models. Version 2.0.0 of the standard was adopted by the OGC in March 2012. The Chair of Geoinformatics is actively involved in the development of the upcoming version 3.0.0 and coordinates the ongoing creation of the specification documents and the data models.
The aim of the 2D+t LandModell research project is to develop a monitoring system for the detection and analysis of change processes in the agricultural landscape. An essential core idea of the project consists in the coupling of semantic data model and complex analysis methods. By defining relevant object classes and attributes, the semantic geodata model allows the representation of the agricultural landscape as a complex system of interacting and changing elements and thus provides a solid basis for the system understanding of the agricultural landscape as well as for in-depth analyses, for example, for the detection, documentation and description of spatio-temporal change processes.
The goal of this project is to provide a small-scale, climate-dependent forecast for feed-in power and consumption and, as a result, an assessment of the compatibility of the calculated loads and services with the components and lines installed in the distribution network. In a further step, the measures to be taken are then to be simulated and evaluated with an increased feed load or consumption. In the end, a forecasting and simulation system is to be developed, which allows network operators to optimise operation in distribution networks.
More than 50% of the world population is living in cities today. No matter if looking at environmental noise, air quality and particulate matter, energy usage and production, or traffic flows - in order to achieve or maintain a high quality of living in cities, municipalities and companies must take into account many different concerns. The concept "Smart District Data Infrastructure" (SDDI) provides planners with the necessary set of flexible tools. Entire cities or city districts are represented by 3D virtual city models which are linked with dynamic data, for example, on the traffic density or energy consumption. These models are used for monitoring and evaluation of the current situation, and especially for simulations of future developments and an early impact analysis.
The Energy Atlas Berlin is a tool supporting the strategic environment and energy planning in Berlin and London. It comprises the representation of the actual state of the environment, energy-related objects and parameters in the sense of an urban inventory. Furthermore, different options for measures such as the comparison of the estimated energy demands and the production and savings potential, e.g. by renewable energy sources and energetic retrofitting of buildings, can be both analyzed and visualized.
The last 20 years clearly show more frequency of extreme weather conditions with an increased summer heat waves in Central Europe. This is particularly evident in metropolitan cities, as the effect is intensified by contribution of Urban Heat Island Effect (UHI). These changing conditions will intensely influence planning both in the building sector and urban development scenarios. The Urban Microclimate modelling project is a software development based on TRNSYS simulation engine, where the model will be able to assess the transient behavior of outdoor spaces and microclimate effects.
The award winning 3D City Database is a free 3D geo database to store, represent, and manage CityGML-compliant city models on top of a standard spatial relational database. The database model contains semantically rich, hierarchically structured, multi-scale urban objects facilitating complex GIS modeling and analysis tasks, far beyond visualization.
In the project, procedures are being developed to automatically generate digital twins of existing infrastructure structures (e.g. roads, bridges, hydraulic structures) as a basis for operation and maintenance. The sub-procedures to be developed include the recording and processing of point clouds (e.g. by means of laser scanning or photogrammetry), extraction of geometric and semantic information using digital image processing, use of knowledge databases for semantic model generation and the evaluation of technical drawings using machine learning. The main activities are the development of algorithms, data structures and technical procedures.
INTREPID aims to create a unique platform, seamlessly integrating Intelligence Amplification and eXtended Reality concepts, with unprecedented Smart Cybernetic Assistants and innovative deep indoor Networking and Positioning capabilities, to improve and accelerate the exploration and assessment of disaster zones. The project will validate its effectiveness, in iterative and complementary pilots, to support the rescue operations in areas that are complex or dangerous to explore.
Duration: 10.2020 - 09.2023
During construction, it is the responsibility of the site manager and other managers to adjust the coordination of the construction project to current events in order to ensure not only efficiency and quality, but also the safety of the workers. Until now, decisions in this regard have often been made on the basis of visual judgement and empirical values and are very rarely based on sound information. Due to the close interconnection of many influencing factors, making good decisions is such a challenge. The EU Horizon 2020-funded project BIM2TWIN therefore aims to develop a digital platform that serves as a sound basis for action for those responsible for the construction site and thus optimises the construction process.
Duration: 10.2020 - 09.2023
This project's aim is the automated generation of geometric and semantic models of car manufacturing plants, respresenting the actual state of the facility.
Duration: 11.2019 - 10.2022
During this project, a guideline is compiled for the systematic creation of a BIM model that records all relevant information and data of a building. This digital twin of the factory building is then further enhanced by adding production data. This way, a single source of truth is created that stores all important data. This data is then visualized using Augmented Reality (AR) to ensure interactivity.
Building Information Modelling
This sub-project (TP) of Early BIM 2 deals on the one hand with the documentation of qualitative decisions in the evaluation and selection of variants. On the other hand, variants are evaluated based on references and the use of (partial) aspects for further design development. The evaluation of potentials and performance in design branches of similar references is to be extended by imperfection (uncertainty, vagueness, imprecision).
Methods based on graph transformation systems are developed and their coupling with parametric design systems is investigated: The concept is that a graph represents a section of the building model at a certain level of detail and that the application of graph transformation rules manipulates this graph in such a way that it subsequently describes this section at a finer level of detail. In this way, it is possible to map detailing processes and the associated knowledge in a computer-interpretable manner and make it reusable. Complete automation is explicitly not intended; instead, the user is provided with a series of detailing rules from which he selects the most suitable one based on experience and context knowledge.
The BIMwood project develops solutions for the value chain of planning and building with wood against the background of digital transformation. It is based on Building Information Modeling (BIM) as a key technology in Architecture, Engineering and Construction (AEC) with profound effects on common working methods. The research project focuses on further developing methods, tools and actions in prefabricated timber construction to improve smooth planning and data management processes.
Duration: 08.2019 - 06.2022
RIMcomb's goal is, therefore, to develop a holistic IT approach to support the collaborative, multi-discipline planning processes of railway infrastructure equipment technology. The core of the approach pursued is the development of a vendor-neutral data model. This model allows the description of track systems and railway equipment with specific properties and object classes. Design processes in railway engineering require various representations from 2D representations of the track topology to highly detailed 3D representations for complex geometric situations. For this purpose, Building Information Modeling (BIM) methods are transferred and applied in the context of railway engineering. New deployment methods for model changes ("patching") are considered.
In response to the urgent demand of international infrastructure stakeholders for extending IFC for tunnels, the standard development project was initiated in Octover 2019 by Infra Room as a fast-track project with a duration of 2 years. Due to the limited time and resources available, it was essential that the project focused on "low hanging" fruits; i.e. selecting use cases to be supported that bring the most value to the future users of the standard. The IFC-Tunnel extension project followed the formal project execution guidelines of bSI that came into effect in 2015 (buildingSMART International 2015).
The TUM Open Infra Platform (short OIP) is a software for viewing alignment and digital elevation model data. The TUM Open Infra Platform supports several file formats for alignment and digital elevation data. It has export and import functions for IFC Alignment, LandXML and OKSTRA. Furthermore it allows to import ASCII-XYZ data and laser scan data in the LAS 1.1/1.2 format.
The aim of this research project is to holistically tackle the problem of disjointed 2D digital drawings and 3D Building Information Models by developing automatic methods for geometric consistency check between them, tracking changes made to drawings, positioning mechanism, and ultimately exchanging drawings and BIM models together in a vendor-neutral container-like manner.
The focuse of the project is on the development of methods for visual representation and exploration of simulation and analysis results for process-integrated presentation and comparison of different design variants. In order to gain insights from the datasets and support the decision making process of the designer, methods from the areas of visual analytics and visual representation are investigated, and concepts are developed based on a requirement analysis in order to interactively filter, link and visualise tendencies and multiple multi-dimensional data situation- and role-specifically from the simulations and analyses.
The design of a building is a complex process in which the solution is developed in an iterative manner in order to fulfill objectives and boundary conditions of multiple designs and engineering disciplines involved. If the Building Information Modeling (BIM) methodology is applied, the planning process starts with a coarse model, which is gradually developed into a more and more detailed model. These refinement steps are described as levels of development (LOD). Due to the increasing application and high potential of semantic BIM models for subsequent simulations and analyses, the implementation of a multi-LOD model seems essential for the future of digital planning and is thus the topic of this research project.
The BIMsite project is focused on the transition from construction planning to construction process. Until recently, building layouts and construction drawings have been designed and delivered in 2D, leading to problems and delays on the construction site due to insufficient depth of planning. The BIMsite project is dedicated to the question of how digital building models can be used for relevant questions in work preparation and during construction.On the one hand, it is to be examined which requirements building models must fulfil with regard to information content and level of detail so that a direct further use for the corresponding project phases is possible.
One of the most important issues during the planning of a construction project is to maintain the quality of the design planning constantly at a high level. Therefore this quality must be checked continuously in terms of accuracy and compliance to the applicable codes and guidelines throughout the duration of a project. The aim of this research project is, to develop an approach which enables the automation of this process with the help of digital methods. A cooperation project with Nemetschek Group and ALLPLAN GmbH
The research group aims to investigate and develop methods and techniques for the collaborative planning of infrastructural construction projects using 3D urban and building models. The 3D model shall be linked to a spatiotemporal database, to external geodata sources and GIS analysis methods. Using an augmented reality system planners on site will be able to localize the 3D planning model.
Human Computer Interaction
For pedestrian dynamics research, the chair of computational modeling and simulation created a new pedestrian simulator under the lead of Dr. Peter M. Kielar. The simulator links modern concepts of practical and theoretical computer science with our in-depth background in pedestrian dynamics. Thus, it is a perfect platform to implement new but also well-known behavior modeling approaches from the field of pedestrian dynamics.
The principles - as well as the concept of the Internet of Things (IoT) - form the potential for decentralized control by dynamically linking the individual MSR sensors and actuators in the building services, facade, sun protection, etc. and the user via direct networking (e. g. Internet). The result is a bottom-up regulation. This results in a more specific and efficient building control system, which is designed to increase user comfort and reduce the energy consumption of the individual building services systems. The aim is to examine at the conceptual level how and to what extent such a decentralized regulation is technically feasible and what potential for improvement it can bring.
Funded by the IGSSE (International Graduate School of Science and Engineering), “Game.UP” is an interdisciplinary research project assessing how gamification methods, as a form of visualisation, can be applied to the field of architecture, to address the challenges of participation and communication between driving bodies. Through a game-full confrontation and appraisal of planning information and technologies processes such as the dissemination of information; the education of planning proposals; and the education of non-traditional planning methods, technologies and structures can become easy to understand, stimulating, computer-based experiences. Furthermore, the effects of urban planning measures can be made directly accessible, visible and interactive through three dimensional digital models and real-time exploratory techniques.
In BEYOND, the applicants want to research how the planning of large construction projects in the infrastructure sector can be carried out more cost-effectively, faster and with a lower error rate while at the same time achieving higher quality by combining Building Information Modelling and people flow simulations. The implementation is carried out with the help of various building blocks: First, the parameterisation of digital building models will be investigated in order to subject these models to Deep Learning methods in the next step: This is intended to automatically investigate the interaction of geometry changes and people flows. If this is successful, a procedure will be developed that proposes optimised geometries with the help of these approaches.
Duration: 03.2020 - 03.2023
The aim of this research is to provide the architect with IT support in the early design stages. In essence, the project extends two of the architect’s primary tools – the sketch and the use of case study references – through the use of information technology, the data storage capacity of the computer and the ability to rapidly transport information in data networks.
Using a computer in order to realize creative design tasks is still cumbersome an inefficient. One of the most challenging problems is the inadequate Human-Computer Interaction of most of the current computer systems. Within an interdisciplinary project a collaborative design platform has been developed on a scale of 1:1.
USP aims to generate interactive tools to support the development of inner-city planning strategies. By monitoring key building codes and providing visualizations and simulation results in real-time they serve as an informed basis for debate and argumentation in the political decision-making and planning process and in turn support the development of inner-city planning strategies that are well-suited to their urban context.
Pointcloud Processing + Remote Sensing
The MultiGo project is funded by the Bavarian Research Foundation and deals with the non-contact surface analysis of vehicle bodies. The aim is to identify dents, dings and waves in a fully automated scanning process. In car body construction, a hybrid sensor system is to measure the geometry and detect surface irregularities - in a single operation. In this way, quality can be checked quickly and efficiently in a small space.
A key issue is the automatic monitoring of the progress of the construction site to detect deviations and to forecast delays. This project concentrates on matching 3D point clouds recorded in a changing environment to an as-planned 3D model for change detection, object extraction, estimation of occluded building parts with integration of construction process knowledge or on-site changes of the planned construction process.
In this subproject of the joint project For3D of the Bavarian Research Foundation, new methods for change analysis in point clouds are to be developed. Point clouds from different sensors or evaluation methods will be merged and compared with images taken at different times.
The project serves two principal goals: (i) to extract and interpret non-visible features of buildings from thermal infrared (TIR) data based on image analysis and (ii) to update and enrich the 3D building models with the information derived from non-visible features.
This project deals with the possibilities and challenges of utilizing airborne decimeter resolution InSAR data for the analysis of densely built-up urban areas. In this context, multi-aspect as well as multi-baseline approaches are investigated to derive surface models as well as to extract building models. The specific properties of SAR imagery have to be taken into account: shadowing, foreshortening, layover areas and scatterers show a completely different behavior dependent on the viewing direction of the sensor. This is a big challenge in combining images from different viewing directions, but allows the reduction of areas with shadows and layover.
In this project, a quantitative ultrasonic imaging approach based on elastic full waveform inversion (EFWI) is developed for accurate reconstructions of physical properties in isotropic and anisotropic structures. The forward model is computed in the frequency/time domain by solving a full-wave equation in a 2D/3D elastic model, accounting for mode conversions and multiple scattering. The inversion is based on local optimization of a waveform misfit function between modeled and measured data.
The numerical framework developed at the chair combines multi-level hp-refinement with a phase-field model for brittle fracture to allow for a locally refined mesh which dynamically adapts to the crack path. In a second step, the Finite Cell Method (FCM) was successfully integrated into the model to enable the efficient simulation of complex geometries.
This work addresses the aforementioned shortcomings by employing a decentral approach to domain organisation. The essential idea is to limit the domain view of each participating unit to their direct neighbours. Transfer of data and updates of topology are only realised between them, hence global updates are not necessary. Since there is an upper bound to the amount of neighbours each subdomain can have, regardless of total domain size, this approach promises to scale even when computing on the largest clusters.
Osteoporosis compromises bone strength, increasing the risk of vertebral fractures with severe health consequences. The development of an accurate and reliable patient-specific vertebral model would be of major clinical relevance, for both the prognosis of fractures and the investigation of implant systems. The finite cell method (FCM) is a high-order embedded domain approach which can handle complex geometry without the need for mesh generation. In this work, we apply the FCM to simulate the biomechanical behaviour of human vertebrae. The flexible nature of FCM regarding complex geometry and different types of geometric representation allows us to develop a robust simulation pipeline for patient-specific analysis.
High performance computing is an increasingly relevant field in computational mechanics for the solution of partial differential equations with many unknowns. This project seeks to develop efficient algorithms and data structures for the solution of complex real-life engineering problems by utilizing the finite cell method, hp-refinement, iterative solution strategies and the computational power of state-of-the-art parallel computing architectures.
The reconstruction of three-dimensional objects from images is a widely applied approach in many engineering fields. The main goal of these techniques is to transform the pixel data that are available in the images into a–potentially application dependent–three-dimensional representation. Apart from standard applications, there has been an increasing interest towards methods that provide three-dimensional geometric representations which are suitable for numerical analysis by means of the Finite Element Method (FEM). The research conducted at the chair for Computation in Engineering addresses the possibilities of coupling photo-based shape measurement techniques to FEM and its derivatives, e.g. the Finite Cell Method.
The chair for Computation in Engineering (CiE) has been involved in research in the field of mesh generation since the early 1990ies with developing a surface mesh generation algorithm for discretizing arbitrary freeform surfaces with triangular and quadrilateral elements (DoMesh). In more recent stages of research the surface mesh generator has been enhanced to a mesh generation framework discretizing arbitrary BRep-volumes with curved high-order tetrahedral elements and thin walled shell-like structures with curved high-order hexahedral elements (TUM.GeoFrame).
This project is integrated in a collaborative research funded by DFG and conducted together the Prof. Wohlmuth (TUM, Mathematics), Prof. Schanda (FH Rosenheim) and Dr. Rabold (ift Rosenheim). The objective of this sub-project is to perform numerical simulations using the p-version of the Finite Element Method (p-FEM) to compute and estimate the vibroacoustic behavior of timber constructions. The efficient integration into modern planning workflows motivates the coupling of simulation models with parametric whole-building information models (BIM).e
Within this project, an interactive simulation tool for risk assessment and real time prediction of floods in complex environmental settings, like urban environments, shall be developed. Floods and impact of floods on infrastructure will be predicted within a multiscale framework reaching from the river down to the scale of the built infrastructure, such as railway, subway, tunnels, waste water channels, buildings, and building infrastructure.
In this project, we concentrate on constructions and built infrastructure with special interest in providing information at a varying and dynamically changing depth. We develop a system that provides the possibility of receiving detailed information (even on very fine-grain levels) about constructions in real time still preserving interactive information steering despite a huge data advent.
The basic idea of the research project is to further develop the modularization principle by using faceted structural elements made of carbon reinforced ultra-high performance concrete for concrete bridge construction. The approach, comparable to the geometrical division of load-bearing structures into finite elements, is characterized by the systematic decomposition of the overall structure into modules that are easy to manufacture. The boundary conditions to be observed here result on the one hand from structural-mechanical properties, and on the other hand from the requirements of manufacture and the subsequent assembly. The complex optimization problem can only be solved by further development of computer-aided methods and the design of suitable modules, their joining technology and production.
The project will contribute to closing the gap between digital design and additive manufacturing (AM). The conceived Design Decision Support System (DDSS) will enable the identification of components suitable for AM technology, founded on a formal AM knowledge base. Methods will be developed for generating Fabrication Information Models (FIM) from Building Information Models (BIM). These methods will be based on graph theory and algorithmic geometry. By incorporating manufacturing information from robots and sensors, methods for creating digital twin representations reflecting as-built geometry and properties are developed. The approaches are based on the concept of multi-LOD BIM ensuring the consistency between multiple levels of detail.
Digital models for additive manufacturing (AM) involve many different geometric scales. These scales start from micrometres for metal- or concrete-based processes. AM product scales, which are in the centre of this project, characterise parts as well as entire const- ructions and range up to tens of metres. This project has closely connected central goals: The first one is to develop a consistent description for the relevant geometric models. The second focusses on efficient simulation methods for AM products in construction, based on multi-scale geometric models. Finally, special emphasis is placed on validation of the developed techniques.
Recent developments in additive manufacturing have provided a unique possibility to create complex structures with porosity on micro and meso-structural levels. Such designs can outperform conventional materials in specific industrial applications, e.g. turbine blades with a transpiration cooling. However, the high flexibility of the input parameters for the 3D printers, such as the laser diameter, hatch distance etc., challenges a reliable estimation of the mechanical behaviour of the final parts. Moreover, a high variation of porosity in all material directions limits the application of analytical bounds based on the void fraction ratio.
Additive Manufacturing, also called 3D printing, is a general term for the production of artifacts by successively adding material in a layerwise fashion. There are numerous processes such as fused deposition modelling, stereolithography and laser power bed fusion. We focus on modelling the process of laser power bed fusion. In this process, a highly focused laser selectively melts powder. Thereby the local state of the material changes from powder to liquid to solid. Once one layer of powder has been treated selectively, a new layer of powder is added. The process is repeated until the final artifact emerges.