Penny Theodoropoulou – Optimized control strategy for Venetian blinds in an event space with fully glazed facades

Nomination Dutch Daylight Student Award 2024
Project type: Research

Optimized control strategy for Venetian blinds in an event space with fully glazed facades
The case study of the Co-Creation Center

Penny focused on improving shading control strategies for a conference center. Using simulations, she optimized indoor daylight levels and blind configurations to achieve a balance between daylight access, minimizing glare, and maintaining views. Her work contributes to creating more comfortable and productive environments in large venues.

• Topic: shading control strategy for conference centre
• Method: simulation of indoor daylight levels and optimisation of blind configuration
• Impact: optimal combination of daylight access, glare avoidance and view out

Nowadays, integrating smart system technologies in buildings is essential for achieving pressing sustainability goals, yet it remains challenging to design. The automatic operation of shading devices to regulate daylight entry is a common passive technology that aims at ensuring occupants’ visual comfort while limiting energy consumption for electric lighting. So far, automatic blinds controllers applied in practice are based on rule-based methods, which respond instantly to weather changes. However, their practical effectiveness is questioned, as they do not always respond to the actual visual needs of the occupants. On the other hand, Architectural Design Optimization (ADO) is getting popular among engineers that try to find optimal solutions, balancing conflicting aspects in their designs. Optimized control strategies are promising tools for optimal blinds operation, although their efficiency is still under research, as their high computational cost does not always allow for their real-life implementation.

The existing literature on control techniques for shading devices is mostly limited to simplified cases of side-lit rooms with fixed view directions. Nevertheless, the increasing use of fully glazed facades in modern designs complicates the assessment of indoor visual conditions, and hence the blinds’ operation. Daylight Glare Probability (DGP), a widely-used metric for reliable glare assessment, and other image-dependent metrics seem inappropriate in such cases, and thus the need for view-independent glare evaluation is raised. On the contrary, cylindrical illuminance (Ecyl), which represents the average of vertical illuminance in all directions around a viewpoint, offers potential for assessing glare risks independently of the occupant’s view direction.

This study explores the implementation of a glare-based control strategy for Venetian blinds in buildings with fully transparent facades. The case study focuses on the Co-Creation Center, a building located in the Green Village at TU Delft, which incorporates both passive and active building technologies to reduce its environmental footprint. The building hosts three types of events: presentations, meetings and workshops. The diverse occupancy patterns, coupled with the four fully glazed facades, make effective blinds control a challenging task to achieve using a traditional rule-based approach. To address this, an optimized control system is proposed, aimed at minimizing visual discomfort due to glare and reducing energy demand for electric lighting, depending on the event. The control strategy is developed within Grasshopper, a promising tool for parametric and optimization problems. Radial Basis Function Optimization (RBFOpt), a model-based optimization method, is utilized for the computation of the optimal blinds’ states.

Results show that the developed algorithm improved visual conditions in the Co-Creation Center by an average of 80% for all activity types, although it led to a 7% increase in the time steps where electric lighting was required compared to the current control system. The success of the control logic in this complex case suggests that it could be applicable to a broader range of buildings and scenarios. However, the time-intensive ray-tracing process performed by Grasshopper for each time step slowed down the optimization, making it unsuitable for real-time implementation. Despite this limitation, RBFOpt proved to be a promising optimization method. Finally, Ecyl showed a 92.5% overall agreement with DGP, indicating that in spaces with multiple windows and uncertain occupant view directions, a view-independent index can adequately predict glare risks when carefully correlated with a reliable view-dependent metric.