- Mean saving from daylight
- 28 % (LBNL meta-analysis, Williams et al., LEUKOS 2012)
- Combined strategies
- 38 % on average (LBNL meta-analysis)
- German reference building
- daylight-dependent control "dimmed, not switching off" (Annex 2 No. 3.2 GModG, from 1 Jan 2027)
- Energy calculation
- daylight dependency factor F_D (EN 15193-1)
- Illuminance when occupied
- minimum values under EN 12464-1:2021 and, in Germany, ASR A3.4
How does daylight harvesting work?
A light sensor, usually on the ceiling or built into the luminaire, measures the brightness in the room. The control compares the reading with a setpoint and adjusts the luminaires through their dimmable driver, for example via 0-10 V or DALI. Output falls as daylight increases and rises again as it fades.
The technical term is constant light control when the sum of daylight and electric light is held at a steady level. Constant light control has a second effect that works even without windows: new installations are deliberately designed brighter than needed when new, because of the maintenance factor. The control trims away this surplus and compensates for luminaire ageing over the years.
Many sensors in the SensMotion range measure presence and daylight in one device, such as IFS105SE and IFS108 for rooms and EFS104 and EFS106 for high bays. The exception is the EFS106W: its photocell only switches at a threshold (hold ON/OFF) and does not provide daylight-dependent dimming.
Dimming or switching: what is the difference?
A switching daylight control only knows on and off. When daylight exceeds a threshold, the light goes off; when it drops below, the light comes back on. It is inexpensive but can switch frequently under changing cloud, which occupants often find distracting.
A dimming control regulates continuously. Transitions are barely noticeable, and it also makes use of partial daylight: if the window covers half of the requirement, the luminaire supplies only the other half.
The reference building in the GModG (German Building Modernisation Act, Gebäudemodernisierungsgesetz, sometimes abbreviated GMG) uses the dimming variant. Annex 2 No. 3.2 specifies daylight-dependent control of the type "gedimmt, nicht ausschaltend" (dimmed, not switching off) under DIN/TS 18599-4:2025-10, section 5.5.4, including constant light control.
"Not switching off" means the reference dims down but never switches off completely. A real installation that dims further or switches off when daylight is sufficient or the space is empty can perform better in the calculation.
What is new in the German reference building from 2027?
Daylight-dependent control is not new to the reference building. It was already part of the previous Annex 2 of the GEG, as were presence detectors, constant light control and building automation class C. What changes on 1 January 2027 is mainly the light source: LEDs in LED luminaires instead of fluorescent lamps with electronic ballasts. The standards references (DIN/TS 18599:2025-10) and zone numbers are also new. The reference applies daylight-dependent control in zones of uses 1 to 4, 8, 12, 30, 33 and 38.
Because the reference now assumes LED plus controls, a real building with LED but no controls falls short of its benchmark in those zones. This affects new-build compliance, the energy certificate from 2027 and minimum energy performance. All zones are covered in reference building lighting from 2027.
Under German law, Section 56(6) GModG requires non-residential buildings with systems over 70 kW to have automatic lighting control by 31 Dec 2029, which should be appropriately zoned and include occupancy detection. The wording does not mention daylight explicitly. Where windows or roof lights exist, however, daylight zones are part of appropriate zoning. See Section 56(6) GModG explained.
How are daylight zones defined?
Daylight drops off quickly with distance from the window: the first row of luminaires benefits most, the second much less, rows further in hardly at all. Luminaires along the facade therefore get their own zone, and in open-plan offices possibly two staggered rows. In halls, luminaires under roof lights form their own zones.
With luminaire-integrated sensors, each luminaire measures and regulates based on its own reading, so the zoning is as fine as the luminaire grid. With separate sensors, one sensor controls a group, which must be chosen so that all its luminaires receive similar daylight. See also zoning and occupancy detection.
How is the system calibrated?
The light sensor does not measure the illuminance on the desk but the light reflected back to the ceiling from floor, furniture and work surfaces. During commissioning, illuminance is therefore measured on the working plane with a lux meter, and the sensor setpoint is adjusted until the required value is reached there. Ideally this is done after dark.
Under EN 12464-1:2021 and the German workplace rule ASR A3.4, minimum illuminance must be achieved while the space is occupied. Daylight control may therefore only dim to the point where daylight and electric light together reach that value. If flooring, furniture or racking change significantly, the calibration should be checked. Monitoring data per zone confirms this in daily use.
What savings are realistic?
The meta-analysis by Lawrence Berkeley National Laboratory (LBNL, Williams et al., LEUKOS 2012; 240 values from 88 studies) reports a mean of 28 % for daylight and 24 % for occupancy. Combinations averaged 38 %. Simulations frequently overestimate daylight, for instance because blinds or shading are missing from the model.
Several strategies are multiplied, not added: total = 1 − (1 − a) × (1 − b). Using the mean values this would give 1 − 0.72 × 0.76 ≈ 45 % on paper, while combinations averaged 38 % in the studies. In deep halls without roof lights, daylight contributes little and occupancy detection is the bigger lever. In halls and warehouses with low occupancy, project results from sensor-based control are often above 60 %, and up to 75 % depending on the project. Calculation and measurement are covered in calculating and verifying savings.
What mistakes are common?
- Sensor looking at the window: direct sunlight or bright sky in its field of view suggests too much brightness, leaving the room too dark.
- Zone too deep: one zone runs from the window to mid-room, so rear desks get too little light or nothing is saved.
- No calibration: factory settings rarely match the space.
- Control reacting too fast: without sufficient delay, the light responds to every passing cloud.
- Planning from simulation rather than reality: blinds, neighbouring buildings and dirty roof lights are missing from the model.
- Daylight instead of presence: presence detection is neglected in areas with little daylight, although it saves more there.
Frequently asked questions
Is daylight control mandatory under Section 56 GModG?
Section 56(6) does not mention daylight explicitly; it requires automatic lighting control that should be zoned and detect occupancy. In the German reference building from 2027, however, daylight-dependent control in specified zones is part of the benchmark.
Does daylight harvesting work in warehouses?
Yes, where roof lights or skylights exist. The luminaires beneath them form separate zones. In deep halls without daylight, occupancy detection saves far more.
Do I need new luminaires for daylight harvesting?
No, provided the existing LED luminaires have a dimmable 0-10 V or DALI driver. Sensors with integrated daylight measurement can be retrofitted. For choosing the sensor, see presence detector vs motion sensor.
What does "dimmed, not switching off" mean?
It describes how the German reference building is calculated: dimmed in daylight but never switched off. An installation that also switches off can score better.
How often should daylight control be recalibrated?
There is no fixed interval. Check the settings after major changes to furniture, flooring or racking.
Last reviewed: 18 November 2026. General information, not legal advice. The legal text is authoritative.