A novel method for daylight harvesting optimization based on lighting simulation and data-driven optimal control

• To date, the best daylighting assessment technique is provided by climate-based simulation tools, which require remarkable efforts to create and calibrate realistic models. The data-driven approaches represent an interesting opportunity to support the physics-based modelling. This work proposes a novel method aimed at the optimization of energy use and luminous environment for a set of lighting control system solutions. The method processes experimental data of occupancy and lighting switch on/off events of an individual side-lit office in an academic building at high latitude via DIVA4Rhino; then, the climate-based simulation results provide the data necessary for the data-driven static optimal control that allow different control strategies of the lighting systems according to their lighting power density. The control allows optimal strategies giving priority to either energy saving or luminous environment improvement, depending on the energy efficiency of the lighting installation, while guaranteeing comfort base level. The results show that the method allows to achieve energy savings up to 18.6% by maintaining high visual comfort levels.

Subject headings

TEKNIK OCH TEKNOLOGIER  -- Samhällsbyggnadsteknik -- Arkitekturteknik (hsv//swe)

ENGINEERING AND TECHNOLOGY  -- Civil Engineering -- Architectural Engineering (hsv//eng)

TEKNIK OCH TEKNOLOGIER  -- Elektroteknik och elektronik -- Reglerteknik (hsv//swe)

ENGINEERING AND TECHNOLOGY  -- Electrical Engineering, Electronic Engineering, Information Engineering -- Control Engineering (hsv//eng)

TEKNIK OCH TEKNOLOGIER  -- Maskinteknik -- Energiteknik (hsv//swe)

ENGINEERING AND TECHNOLOGY  -- Mechanical Engineering -- Energy Engineering (hsv//eng)

Keyword

visual comfort

machine learning

Data-driven control

lighting control system

lighting

daylighting

Daylighting simulation

energy saving

energy efficiency

control strategies

Publication and Content Type

kon (subject category)

ref (subject category)