Workshops

Heating, ventilation, and air conditioning (HVAC) systems of large commercial or office buildings are significantly complicated. The traditional design processes of HVAC systems are usually tedious and repetitive. And the engineers spend a lot of time in drawing the layouts of ducts and pipes, because they should change the design repeatedly in response to changes in upstream architecture design. There have been many studies on automated architectural, engineering, and construction design processes that provide efficient and convenient technical methods for HVAC system design. On this basis, adding artificial intelligence technology can make the traditional design process more streamlined and reduce engineers’ repetitive work. This workshop aims to introduce a framework for HVAC system automated design and discuss the challenging points in research.

The widespread adoption of building automation and IoT technologies have enabled the collection of real-time measurements of large-scale building energy systems. It is promising to extract valuable knowledge and incorporate advanced metering, automated controls and demand side management in the interaction between buildings and users/grids. In practice, the fully exploitation of building operational data can be challenging due to the intrinsic complexity in building operations, interaction with the energy grid and low-quality of measurement data. As a consequence, robust data-driven solutions are urgently needed to fully unleash the potentials of building data to deploy both energy information and automated system optimization services during operations. This workshop aims to present an overview of data-analytics applications in the energy and buildings field, covering typical applications of predictive data analytics for energy forecasts and fault diagnosis, descriptive data analytics for operating pattern and rule identifications, and novel machine learning solutions for heterogeneous data utilization and control optimizations at different scales.

SIM-VICUS is a cutting-edge building and district simulation software, designed for dynamic evaluation of energy efficiency within buildings and districts. It is uniquely equipped to effectively manage complex calculations of large buildings and networks while efficiently handling vast data volumes. With SIM-VICUS, users can directly create and modify the geometry of buildings and district heating networks using its intuitive 3D graphical interface. In addition, it supports the integration of BIM models and GIS data via IFC and geoJSON respectively. All model parameters are easily editable, drawing from extensive databases. Moreover, these parameters are presented in an accessible manner through interactive false color views. SIM-VICUS's state-of-the-art variable time-step solver drastically reduces simulation times for detailed models and is compatible with the Functional Mock-up Interface (FMI). SIM-VICUS is open-source, championing transparency and continuous innovation. Further enhancing its accessibility, the software is available for free download, inviting users globally to leverage its advanced capabilities.

The BOPTEST framework is used to evaluate the performance of HVAC control algorithms by enabling co-simulation. The framework provides a standard simulation environment and is organized into specific building test cases. The models are implemented in Modelica packaged as FMU for use in standardized simulation environment. An API for this environment allows users to advance the simulation of the detailed building model with input from their control algorithms, and report key performance indicators. The BOPTEST-Gym interface enables the implementation of Reinforcement Learning algorithms in BOPTEST emulators out of the box. The focus of the workshop is to provide a hands-on introduction to the BOPTEST framework and its Gym interface, which aims to enable benchmarking and facilitate the development of advanced control strategies for buildings. Key activities:

• an overview of the BOPTEST project goals and approach
• a guided, hands-on experience with the software
• an opportunity to use the software for their own research and development
• an opportunity to provide feedback for project development

The virtual CityLearn Workshop will introduce the software tool CityLearn to model and analyze advanced control approaches in grid-interactive smart communities, e.g., demand response and load shaping in buildings. CityLearn is an open source OpenAI Gym environment targeted at the easy implementation and benchmarking of advanced control algorithms, i.e., model predictive control or deep reinforcement learning. Main applications to-date consist of controlling the charging and discharging of active storage systems i.e. battery and thermal storage tanks and heat pump power in the buildings. CityLearn has been introduced in 2019 and is being actively developed in the building energy community.
In this workshop, we will provide a walk-through tutorial on how to set up the environment using input data from residential building energy models in public End-Use Load Profiles (EULP) for the U.S. Building Stock database. Participants will be able to follow along using the provided Jupyter notebook scripts. We will also provide notebooks on how to use a simple rule-based control architecture, advanced soft-actor-critic methods, and the MARLISA multi-agent reinforcement learning control architecture. Finally, participants can optimize hyperparameters of algorithms and compare their findings against each other.
The target audience of the workshop includes academic, private and commercial researchers that are interested in the topics of single and aggregated building energy systems load management through the use of advanced smart control systems. We also aim to attract artificial intelligence researchers whom are drawn to solving complex control problems or are interested in learning about a new reinforcement learning environment that deviates from the typical toy problems to a more real-world problem with urgency like climate change mitigation and decarbonization of building end-uses.

The talk delves into the transformative potential of Building Information Modeling (BIM) when integrated with Building Performance Simulation (BPS) techniques. This synergy marks a significant advancement in the architecture, engineering, and construction industries by fostering informed decision-making, enhanced collaboration, and sustainable building design.
The presentation opens with an overview of the challenges faced in traditional construction practices, where disjointed workflows and communication gaps often lead to suboptimal building performance. The convergence of BIM and BPS emerges as a solution to bridge these gaps, allowing stakeholders to visualize, simulate, and optimize building performance throughout its lifecycle.
The talk concludes by highlighting the paradigm shift that BIM-Centric Building Performance Simulation brings to the construction industry. By combining advanced modelling, simulation, and collaboration tools, professionals can create sustainable, efficient, and optimized buildings that meet the challenges of the 21st century. Embracing this approach promises a future where buildings are designed, constructed, and operated with utmost precision, resulting in improved performance, reduced environmental impact, and enhanced occupant experiences.

Cities consume more than two-thirds of the world’s primary energy and account for more than 70% of global greenhouse gas emissions. Urban building energy and environment modeling is an excellent method to explore opportunities and strategies to reduce building energy consumption and carbon emissions and improve indoor and outdoor living environments. Communities are not only the primary space to create true sustainability but also the starting point for the use of healthy and low-carbon ideas to address climate change and energy efficiency and promote a carbon-neutral society. This workshop will discuss the applications and challenges of applying urban-scale advanced modeling techniques (energy modeling, CFD analysis, etc.) to evaluate the indoor and outdoor environment, analyze energy conservation measures, support urban planning, mitigate urban heat island effects, etc.

Building energy simulation (BES) plays an increasingly important role in evaluating and optimizing building energy performance as cities strive to be energy efficient and sustainable. By leveraging data science and machine learning in BES, stakeholders can gain valuable insights that aid decision-making and optimize energy performance. Computational reproducibility is also essential in open science to guarantee all results are transparent and can be independently verified and thus be built upon by others, reducing duplicated efforts and improving the quality of BES research. This seminar aims to explore the integration of data science and machine learning in BES from individual buildings to the urban scale. The seminar will focus on the following key aspects:

• Tools and techniques for the application of data science and machine learning to BES
• Addressing challenges and considerations when applying data science and machine learning to large-scale urban building energy modeling (UBEM)
• Best practices for enhancing reproducibility in BES research