A5031098775- Building Energy and Comfort Optimization
- Wind and Air Flow Studies
- BIM and Construction Integration
- Advanced Multi-Objective Optimization Algorithms
- Probabilistic and Robust Engineering Design
- Energy Efficiency and Management
- Time Series Analysis and Forecasting
- Simulation Techniques and Applications
- Health Systems, Economic Evaluations, Quality of Life
- Fibromyalgia and Chronic Fatigue Syndrome Research
- Advanced Clustering Algorithms Research
- Fuzzy Logic and Control Systems
- Impact of Light on Environment and Health
- Energy Load and Power Forecasting
- Music and Audio Processing
- 3D Modeling in Geospatial Applications
- Hygrothermal properties of building materials
- Smart Grid Energy Management
- Urban Heat Island Mitigation
University of Victoria
2019-2021
University of Lübeck
2021
University Hospital Schleswig-Holstein
2021
ETH Zurich
2020
Fast machine learning-based surrogate models are trained to emulate slow, high-fidelity engineering simulation accelerate design tasks. This introduces uncertainty as the is only an approximation of original model. Bayesian methods can quantify that uncertainty, and deep learning exist follow paradigm. These models, namely neural networks Gaussian process enable us give predictions together with estimate model's uncertainty. As a result we derive uncertainty-aware automatically identify...
The buildings sector is one of the largest contributors to CO 2 emissions, comprising up 33% global total (ürge-Vorsatz et al., 2007).Improved computational methods are needed help design more energy-efficient buildings.The Python library besos along with its associated web-based platform BESOS researchers and practitioners explore energy use in effectively.This achieved by providing an easy way integrating many disparate aspects building modelling, district optimization machine learning...
Statistical surrogate models, or meta-models, are used to emulate building simulation models. Their key advantage is the reduction of computational cost. This in particular matters if design analysis demands explore a large number different designs options as optimization uncertainty problems. To derive model, data set consisting in- and output generated. then train surrogate. process collecting may be time intensive designer has wait until model available. In this study we construct global...
Abstract Background Spironolactone (SPL) is a reversible mineralocorticoid receptor (MR) and androgen (AR) antagonist which attracts pharmacotherapeutic interest not only because of its beneficial effects in heart failure but also the pathogenetic roles MR AR activities neuropsychiatric diseases. Recently, rapid‐onset SPL have been documented case series women with fibromyalgia syndrome (FMS). To reaffirm this observation, we performed double‐blind placebo‐controlled randomized clinical...
Abstract Machine learning-based surrogate models are trained on building energy simulation input and output data. Their key advantage is their computational speed allowing them to produce performance estimates in fractions of a second. In this work we showcase the use deep convolutional neural network embedded into web application, users rapidly explore at high spatio-temporal resolution. Users can pick any climate an interactive map, customize design with thirteen decisive parameters, model...
This paper presents the besos Python library, which can parameterize EnergyPlus models and integrate these with ecosystem of tools including machine learning optimization libraries. library underpins BESOS (Building Energy Simulation, Optimization Surrogate-modelling) platform. The need for a flexible Python-based that integrates domains is outlined. A case study presented to demonstrate benefits this integrated approach, an overview given research works have leveraged benefits. We also...
Data-driven, black box machine learning models have received a lot of attention in the field building control. They been used successfully to predict behaviour given information like weather forecasts and real time sensor information. In these models, occupant is considered act exogenously on building. We consider users as active elements operation control loop. To make educated decisions they be informed about how will behave. Therefore, we propose prediction model which explains occupants...
Fast machine learning-based surrogate models are trained to emulate slow, high-fidelity engineering simulation accelerate design tasks. This introduces uncertainty as the is only an approximation of original model. Bayesian methods can quantify that uncertainty, and deep learning exist follow paradigm. These models, namely neural networks Gaussian process enable us give predictions together with estimate model's uncertainty. As a result we derive uncertainty-aware automatically suspect...
Parametric exploration and optimization of building geometry is a powerful tool for designing energy efficient buildings. However, in practice this process computationally expensive time-consuming. In research, we explore the use surrogate models, i.e. statistical approximations physics-based simulation to lower computational burden large-scale analysis. For purpose, developed novel dataset 38,000 residential models derived from real world floor plans (Wu et al. (2019)) train model emulate...