Conference Agenda

Overview and details of the sessions of this conference. Please select a date or location to show only sessions at that day or location. Please select a single session for detailed view (with abstracts and downloads if available).

Please note that all times are shown in the time zone of the conference. The current conference time is: 5th July 2022, 14:42:25 CEST

 
 
Session Overview
Session
Session W2.2: Ensuring high quality building simulations
Time:
Wednesday, 01/Sept/2021:
13:00 - 14:30

Session Chair: Laura Carnieletto, University of Padova
Session Chair: Frederik Maertens, boydens engineering
Location: Concert Hall - Artiestenfoyer
't Zand 34, Bruges

External Resource: Click here to join the livestream. Only registered participants have received the access code for the livestream.
Show help for 'Increase or decrease the abstract text size'
Presentations
13:00 - 13:18

Heat and moisture transport through a living wall system designated for greywater treatment

Hayder Alsaad, Conrad Voelker

Bauhaus-University Weimar, Germany

Aim and Approach

(max 200 words)

Façade greening systems designated for remediating greywater can help to relieve the water treatment centres while saving on irrigation water (Prodanovic et al. 2017). This study aims to numerically investigate the heat and moisture transport in such systems. As most heat and moisture simulation models cannot simulate the complex impact of vegetation on the simulated parameters, this study was conducted by coupling two simulation tools: ENVI-Met and Delphin. ENVI-Met is a high-resolution meteorological model that can simulate the interaction between urban geometry, vegetation, and the outdoor environment (Bruse and Fleer 1998). Delphin, on the other hand, is a simulation package for coupled heat and moisture transport in porous building materials (Grunewald 2000). In the present study, ENVI-Met was used to calculate the influence of the plants on air temperature, velocity, relative humidity, wind direction, and radiation (long wave and short wave) on the façade. Subsequently, the calculated parameters were then imposed on the façade in Delphin. Thus, ENVI-Met was used to determine the local climate conditions on the façade, which were used to conduct the hygrothermal simulations with Delphin. The hygrothermal simulations had a duration of four years to reach the equilibrium moisture content in the construction.

Scientific Innovation and Relevance

(max 200 words)

With the continuously increasing levels of pollution in cities and rising temperatures due to the urban heat islands, living walls have been growingly investigated because of their promising potential in improving the urban environment. In addition, these systems can have a significant impact on the performance of the walls on which they are mounted. The literature indicates that façade greening can improve the heating demand of the building (Tudiwer and Korjenic 2017). Yet, due to evaporation from the substrate and transpiration from the plants, the relative humidity on the façade can increase (Capener and Sikander 2015). Moreover, as the greening system investigated in this study is meant for greywater treatment, it involves continuous water flow in the substrate of up to 50-75 L/d. An increase in humidity can damage the building material and reduce the energy efficiency of the building by increasing the heat conductivity of the wall layers. As hygrothermal simulations of living walls designated for greywater treatment is not reported in the literature, this study aims to investigate the impact of relatively high exposure to moisture on façades.

Preliminary Results and Conclusions

(max 200 words)

To evaluate the impact of the living wall, two simulation models were created: a façade covered with a living wall and a reference facade with no greening. Both facades had a generic brick structure with a total thickness of 420 mm. The simulations showed that while the living wall was emitting water vapour, it did not increase the humidity content in the structure because of the ventilated air gap between the greening and the façade. In fact, the facade greening protected the wall from wind-driven rain and thus had a 16% less humidity content in the fourth simulation year in comparison to the reference case. The relative humidity of the interior surface of the wall was almost similar in both cases (58.3% with greening and 59.8% without greening). In the summer months (21 June – 21 September), the living wall cooled the façade's surface temperature due to shading, the thermal mass of the substrate, and the passive cooling of the plants. The maximum surface temperature behind the greening was 25°C compared to 40.6°C without greening. In the winter, the greening increased the minimum interior surface temperature by 1.2 K, which indicates an improvement in the thermal resistance of the construction.

Main References

(max 200 words)

Bruse, Michael; Fleer, Heribert (1998): Simulating surface–plant–air interactions inside urban environments with a three dimensional numerical model. In Environmental Modelling & Software 13 (3-4), pp. 373–384. DOI: 10.1016/S1364-8152(98)00042-5.

Capener, Carl-Magnus; Sikander, Eva (2015): Green Building Envelopes – Moisture Safety in Ventilated Light-weight Building Envelopes. In Energy Procedia 78, pp. 3458–3464. DOI: 10.1016/j.egypro.2015.11.179.

Grunewald, John (2000): Documentation of the Numerical Simulation Program DIM3.1", Volume 2: User's Guide. Insitute of Building Climatology, Faculty of Architecture, Univesity of Technology Dresden.

Prodanovic, Veljko; Hatt, Belinda; McCarthy, David; Zhang, Kefeng; Deletic, Ana (2017): Green walls for greywater reuse. Understanding the role of media on pollutant removal. In Ecological Engineering 102, pp. 625–635. DOI: 10.1016/j.ecoleng.2017.02.045.

Tudiwer, David; Korjenic, Azra (2017): The effect of living wall systems on the thermal resistance of the façade. In Energy and Buildings 135, pp. 10–19. DOI: 10.1016/j.enbuild.2016.11.023.



13:18 - 13:36

A convolutional neural network for the hygrothermal assessment of timber frame walls

Astrid Tijskens, Staf Roels

KU Leuven, Belgium

Aim and Approach

(max 200 words)

Timber frame walls typically consist of a wind barrier at the cold exterior side and a vapour barrier at the warm interior side. In cold climates, the vapour barrier must have a higher vapour resistance than the wind barrier, to ensure vapour that entered the construction at the inside can dry out towards outside. However, there are no general guidelines available as to which combinations of wind and vapour barrier are safe in a specific context. Sometimes, a rule of thumb is used, which requires the ratio between the vapour resistances of vapour and wind barrier to be between 5 and 15 or even higher. This rule, however, does not take into account moisture buffering capacity of the structure nor specific climatic aspects, and hence does not guarantee an optimal solution. Because a hygrothermal simulation for every case would be too time-intensive, a metamodel is proposed in the current study, which allows quickly determining adequate combinations of wind and vapour barrier under given conditions. A convolutional neural network for time series is used to replace the hygrothermal simulations, thus allowing flexibility in the desired post-processing.

Scientific Innovation and Relevance

(max 200 words)

The use of neural networks for time series predictions is a fairly novel metamodelling strategy in the field of building physics. When evaluating the hygrothermal performance of a building component in a probabilistic framework, metamodelling strategies have in the past been applied to predict specific and single-valued performance indicators. This approach provides little flexibility and might not provide sufficient information for decision-making. Instead, a metamodel predicting hygrothermal time series, as calculated by the original hygrothermal model, provides more information and allows the user to post-process the output as desired. In [1-2], the authors proved the applicability of convolutional neural networks for hygrothermal calculations of massive brick walls. The current study explores the models to predict the hygrothermal response of timber frame walls.

Preliminary Results and Conclusions

(max 200 words)

First results show that it is possible to replace the time-consuming hygrothermal model with a much faster convolutional neural network, while maintaining high accuracy. Since material properties, such as (humidity dependent) vapour resistance and moisture buffering capacity, play a significant role in the hygrothermal response, the network requires additional input on this, compared to the network from [1-2], resulting in a slightly different network architecture.

Main References

(max 200 words)

[1] A. Tijskens, S. Roels, and H. Janssen, “Neural networks for metamodelling the hygrothermal behaviour of building components,” Building and Environment, vol. 162, no. June, p. 106282, 2019.

[2] A. Tijskens, H. Janssen, and S. Roels, “Optimising Convolutional Neural Networks to Predict the Hygrothermal Performance of Building Components,” Energies, vol. 12, no. 20, p. 3966, 2019.



13:36 - 13:54

Besos: a python library that links energyplus with energy hub, optimization and machine learning tools.

Theodor Victor Christiaanse1,2, Paul Westermann1,2, Will Beckett1,2, Gaelle Faure1,2, Ralph Evins1,2

1Energy in Cities group, Department of Civil Engineering, University of Victoria, BritishColumbia, Canada; 2Institute for Integrated Energy Systems, University of Victoria, British Columbia, Canada

Aim and Approach

(max 200 words)

The goal of the BESOS library is to create an easy way for academics start building modelling experiments that involve linking EnergyPlus with optimization, energy system design and machine learning techniques. The Building and Energy Simulation, Optimization and Surrogate-modelling (besos) library provides a Python-based software bridge between EnergyPlus and various other environments. Furthermore, software abstractions are specifically setup such that building energy modellers can quickly create numerous model variations using parametric definitions. These include optimizing the building design and energy systems, and integration of the modeling results with machine learning techniques. Accompanying the Python library, a web platform (BESOS) [1] is freely accessible for academics to use the software, learning through our extensive library of examples and developing new software extensions on top of the existing software paradigms.

Scientific Innovation and Relevance

(max 200 words)

The optimal exploration of building design and operation requires the use of many software tools. Among them, EnergyPlus (E+) is a commonly used physics-based building energy simulation tool. The use of machine learning (ML) is finding adoption among engineers in the field and may have a huge impact on the speed and breath of modelling experiments. ML techniques can perform different tasks that could inform building design and operation such as; (i) capture the dynamics of a physics-based models in a fast and accurate surrogate model [2] reducing the cost of expensive exploration, (ii) large building datasets of E+ or real timeseries data can be analysis on a building-by-building level using ML techniques [3], (iii) retrofit measures may be identified by understanding the timeseries sensor data through black box techniques [4], (iv) forecasting future energy performance of the building stock using ML for forecasting [5].

The potential of these techniques ML techniques is evident. The integration between these new Python-based libraries and physics-based modelling tools used for energy modelling was limited. Our software library gives modellers the ability to create building models that work in E+ and manipulate the inputs and outputs of the E+ model within a Python environment.

Preliminary Results and Conclusions

(max 200 words)

The Python environment allows for the integration with these new novel machine learning libraries such as TensorFlow and Scikit-learn. Furthermore, we have also included links to powerful optimization solvers and energy system design tool Energy Hub. These three tools provide numerous options for energy modelers to combine multivariant data sets and E+ models to create experiments that stack many different techniques into a single interface.

We will discuss the challenges and successes we have had building the software library besos and BESOS platform. A demo of the software capabilities will be shown and demonstration how multi-model ecologies can be built are presented. A list of recent projects and papers that are would not be possible without the platform [2-5]. Finally, we will share how we continue to improve the underlying software for the next possible versions by partnering with computer scientists and making innovative technologies available to the building modelling domain.

Main References

(max 200 words)

[1] Faure G, Christiaanse T, Evins R, Baasch GM. BESOS: a Collaborative Building and Energy Simulation Platform. In Proceedings of the 6th ACM International Conference on Systems for Energy-Efficient Buildings, Cities, and Transportation 2019 Nov 13 (pp. 350-351).

[2] Westermann P, Evins R. Surrogate modelling for sustainable building design–A review. Energy and Buildings. 2019 Sep 1;198:170-86.

[3] Baasch G, Wicikowski A, Faure G, Evins R. Comparing gray box methods to derive building properties from smart thermostat data. In Proceedings of the 6th ACM international conference on systems for energy-efficient buildings, cities, and transportation 2019 Nov 13 (pp. 223-232).

[4] Westermann P, Deb C, Schlueter A, Evins R. Unsupervised learning of energy signatures to identify the heating system and building type using smart meter data. Applied Energy. 2020 Apr 15;264:114715.

[5] Westermann P, Braun J, Murphy E, Grieco J, Evins R. Insight Into Predictive Models: On The Joint Use Of Clustering And Classification By Association (CBA) On Building Time Series. In Rome, Italy; [cited 2020 Jul 16]. p. 1564–71. Available from: http://www.ibpsa.org/proceedings/BS2019/BS2019_211236.pdf



13:54 - 14:12

Data-driven black box model of building dynamics

Sophie Bernard1,2, Valery Ann Jacobs1, Bert Belmans1,3, Arjen Mentens1, Filip Descamps1,4, John Lataire1

1Vrije Universiteit Brussel, Belgium; 2Université libre de Bruxelles, Belgium; 3Universiteit Antwerpen, Belgium; 4Daidalos Peutz, Belgium

Aim and Approach

(max 200 words)

In this research a modelling tool is presented to derive surrogate models of thermal energy transfers in buildings, to support the development and testing of smart control algorithms.

A data-driven approach was used to identify a model able to predict the indoor temperature in a case-study building when an electric heater was turned on. The data about the system was generated by EnergyPlus simulations, which resolve the heat balance equations to simulate the thermal response of a building. The model structure that was selected is a second-order ARMAX transfer function whose parameters were identified with a Least Squares optimization criterion. The model inputs were limited to the heater’s power, the global horizontal solar radiation and the outdoor dry-bulb temperature.

Scientific Innovation and Relevance

(max 200 words)

Buildings account for one-third of the global energy consumption in the world, which is more than the industry sector or the transport sector. For environmental and economic reasons, it is thus important to reduce their energy consumption, while preserving a comfortable environment for their occupants. In this context, computer-aided control techniques for building comfort systems can be a valuable asset. Advanced control techniques, such as model predictive control, require supporting system models for development and testing. However, detailed physical models that carefully take system dynamics into account are too computationally intensive for practical applications. This is why surrogate models of low complexity are highly relevant.

The innovative element is that a frequency domain modelling approach is used. This allows to conveniently select the frequency band of interest. Namely, the dynamics are mostly important at low frequencies. Discarding the high frequencies implicitly removes a significant amount of noise.

Also, the use of a data-driven approach implicitly takes into account influences which, in a first principles approach, might have been neglected. That is, the model is validated on the data rather than on physical insight.

Preliminary Results and Conclusions

(max 200 words)

Simulation experiments have been conducted where random binary sequences (a sequence of step inputs) have been applied as the heater's input, and historical weather conditions have been used for the solar radiation and the outside air temperature.

A transfer function model was obtained, describing the relation between the inputs (heater and weather conditions) and the resulting ambient temperature in the room. A data set corresponding to the month of July was used for the identification, and the resulting model was then validated on a data set of the month of December.

It was demonstrated that the model was able to predict fairly accurately the indoor temperature when the building was subject to winter or summer weather conditions. Further improvements and refinements will be carried out, including taking into account the difference in time constants of the heater and the weather conditions.

Main References

(max 200 words)

P. Abrahams et al., Method for Building Model Calibration to Assess Overheating Risk in a Passive House in Summer, Proceedings of the 16th IBPSA conference, Rome, Italy, Sept. 2-4, 2019, DOI 10.26868/25222708.2019.210768

S. Mostafavi et al., Model Development for Robust Optimal Control of Building HVAC, Proceedings of the 16th IBPSA conference, Rome, Italy, Sept. 2-4, 2019, DOI 10.26868/25222708.2019.211331

S. Royer, Energy and Buildings, Volume 78, pg. 231-237, A procedure for modeling buildings and their thermal zones using co-simulation and system identification, 2014. DOI 10.1016/j.enbuild.2014.04.013

S. Royer et al., IFAC Proceedings Volumes, Volume 47, Issue 3, Pages 10850-10855, Black-box modeling of buildings thermal behavior using system identification, 2014. DOI 10.3182/20140824-6-ZA-1003.01519

R. Pintelon and J. Schoukens. System Identification: A Frequency Domain Approach. John Wiley, 2nd edition, 2012.



14:12 - 14:30

Summer passive strategies assessment based on calibrated building model using on site measurement data

Obaidullah Yaqubi1,2,3, Auline Rodler1,2, Sihem Guernouti1,2,3, Marjorie Musy1,2,3

1Equipe de recherche BPE, Cerma Ouest, Nantes, France; 2Institut de Recherche en Sciences et Techniques de la Ville (IRSTV) , Nantes, France; 3CNRS UMR 6183, GeM, Université de Nantes, France

Aim and Approach

(max 200 words)

With the changes in worldwide climate conditions, extreme summer heat events will become more frequent and severe rendering buildings uncomfortable. This paper in this context, presents the application of co-simulation on practical design issues of mixed mode ventilated buildings. It is based on a 2-months field study measurement of outdoor and indoor air temperatures and window operation of an existing residential building during the hottest season of the year in Nantes (France) in 2018.

The aim of this study is first to use measured indoor temperature to calibrate a building co-simulation model and second to evaluate how openness ratio of windows and operation of window shutters affect the indoor thermal comfort during summer.

Scientific Innovation and Relevance

(max 200 words)

Two modelling tools, Contam and Trnsys, were coupled to simultaneously simulate airflow and temperature dynamics of the whole building. Since each of the five storeys had similar thermo-physical features, it was decided to consider only the last storey for the purpose of the present study because it is the most sensitive to outdoor conditions. Each piece in the apartment i.e. bedroom, living room, and bathroom was treated as a separate zone. The time step of the simulation has been set to 15 min.The agreement between measured and simulated indoor temperature values at every hour was evaluated with the Coefficient of Variation of the Root Mean Square Error (CV(RMSE)) and Mean Absolute error (MAE).

For assessing the thermal comfort, international adaptive comfort standard methodologies such as EN 16798 and ASHRAE 55 as well as PMV were used to measure and to compare indoor comfort in the apartments.

Preliminary Results and Conclusions

(max 200 words)

After making necessary adjustments to the model, the co-simulation model was calibrated to the lowest (CV(RMSE)) values, between 3 and 5%, on indoor temperatures in the living rooms of 3 apartments and stairwell. Analysing indoor temperature of calibrated simulation building with adaptive thermal comfort indices showed that apartments and stairwell in the last floor of the building experienced higher temperatures than the maximum allowable operative temperature for category I and II of EN 16798 and ASHRAE 55 category of acceptability of 80 and 90% up to 5% of the studied period. Passive strategies such as adjusting the openness ratio of windows at night and day (natural ventilation) and closing window shutters to 0.9 i.e. ratio of no-transparent area during the day proved to reduce overheating risks for category I and II of EN 16798 and ASHRAE 55 acceptability category of 80% but not enough for acceptability category of 90%. The latter may need to use mechanical means to reduce overheating risks.

Main References

(max 200 words)

Dols, W.S., and Polidoro, B.J. (2015). CONTAM User Guide and Program Documentation Version 3.2 (Na-tional Institute of Standards and Technology)

ASHRAE-55-2017, «Thermal Environmental Conditions for human occupancy,» Atlanta, 2017.

Bienvenido-Huertas, D., Sánchez-García, D., Rubio-Bellido, C., and Oliveira, M.J. (2020). Influence of adap-tive energy saving techniques on office buildings located in cities of the Iberian Peninsula. Sustain. Cities Soc. 53, 101944



 
Contact and Legal Notice · Contact Address:
Privacy Statement · Conference: Building Simulation 2021
Conference Software - ConfTool Pro 2.6.143+TC
© 2001–2022 by Dr. H. Weinreich, Hamburg, Germany