Co-simulation and validation of the performance of a highly flexible parametric model of an external shading system

Ellika Taveres-Cachat,Francesco Goia

doi:10.1016/j.buildenv.2020.107111

Ellika Taveres-Cachat, Francesco Goia

Open Access

https://doi.org/10.1016/j.buildenv.2020.107111

Copy DOI

Abstract

The article presents a validation study of a modelling approach implemented in a numerical script for external louvred shading systems based on an experimental analysis in a full-scale test facility. The model developed to abstract the system was entirely parametric and used co-simulation to predict the indoor air temperature and illuminance levels in two points of the test cell.The calibration of the model of the test facility was carried out using a combination of two methods: automated calibration based on multi-objective optimization with a genetic algorithm and manual calibration. In total, six different configurations of the external shading system with varying complexity were investigated to validate the script. Its performance was assessed using three metrics: the root mean square error, the coefficient of variation of the root mean square error, and the normalized mean bias error.The results showed that the thermal environment was simulated with consistent accuracy for all the cases investigated, predicting air temperatures with an error well within the tolerance of building performance simulation tools and the experimental uncertainty. The daylighting model satisfactorily captured the different dynamics of illuminance peaks and dips, replicating the variations between different configurations, but with a lower degree of accuracy than for the thermal simulations.

Highlights

The results showed that the thermal environment was simulated with consistent accuracy for all the cases investigated, predicting air temperatures with an error well within the tolerance of building performance simulation tools and the experimental uncertainty
Co-simulation is a method used in building performance simulation which allows coupling different models that describe parts of the building, each of which is run in a different simulation tool in a way that they can exchange simulation data during run-time [3]
The CV Root Mean Square Error (RMSE) was +2% and the Normalized Mean Bias Error (NMBE) was 2%, which indicates that the distance between the measured and simulated data points was small, and the level of accuracy of the model is well within the acceptable error of building performance simulation tools

Summary

Introduction

Parametric software allow the exploration of a larger solution space in early design stages when changes are still relatively easy to implement and less costly. These tools rely on an explicit dynamic linkage between geometric definitions of the buildings elements, system parameters, and whole-building performance [1,2]. Because these tools can simulta neously be used as interfaces to different building simulation engines, they can help increase the interoperability of simulation tools by sup porting co-simulation frameworks. The development of new simulation approaches is useful as it can help investigate advanced control strategies or complex geometries [5,6,7,8] as well as support the development of design approaches such as free form facades and shading elements [9]

Methods

Results

Discussion

Conclusion