Search and Access Archived Conference Presentations

2019 ASHS Annual Conference

An Effective Algorithm for Controlling Greenhouse Light to a Target Daily Light Integral Using Dimmable Supplemental Lights.

Monday, July 22, 2019: 4:15 PM
Montecristo 3 (Tropicana Las Vegas)
Geoffrey Weaver, University of Georgia, Athens, GA
Marc W. van Iersel, University of Georgia, Athens, GA
Lighting recommendations for greenhouse crops are often made in terms of the daily light integral (DLI), the total amount of photosynthetically-active radiation received by a crop in a day. Methods for controlling greenhouse lighting to a target DLI using on/off control have been previously described. We developed an algorithm for DLI control which makes use of the dimmability of light-emitting diode (LED) lights, and accounts for the behavior of a nominal distribution of daily sunlight intensities. The objective was to provide exactly enough light to reach a target DLI within a specified photoperiod if supplemental light is needed, and provide no excess light if it is not needed. This algorithm was applied to typical meteorological year data for five U.S. cities (Athens, GA; Elmira, NY; Kalamazoo, MI; Seattle, WA; Yuma, AZ) with several photoperiods (16 to 24 hours) in MATLAB using a custom script. Simulations were conducted for each city and photoperiod based on reaching a target DLI of 17 mol m-2 d-1 with a maximum LED photosynthetic photon flux density (PPFD) of 200 µmol m-2 s-1. For Athens, GA the DLI from sunlight alone was less than 17 mol m-2 d-1 on 98 days. With a 16-hour photoperiod at this location, excess light was provided only on days for which sunlight alone exceeded the target DLI, and this occurred on 10 days (3.4 mol m-2 yr-1 total annual excess from the LEDs). The target DLI was not reached on 35 days, with an average deficit of 1.9 mol m-2 d-1, which was partially due to the limited PPFD output of the LEDs. With a 24-hour photoperiod (Athens, GA), excess light was provided on only one day, and the target DLI was not reached on two days (0.04 mol m-2 d-1 average deficit). Results were comparable at the other locations, with improved accuracy observed at longer photoperiods. The algorithm was least successful when applied to the Seattle, WA data with a 16-hour photoperiod; supplemental light was required on 199 days, excess light was provided on 23 days (10.4 mol m-2 yr-1 excess), and the target DLI was not reached on 137 days (average 3.2 mol m-2 d-1 deficit). To be applied in a greenhouse setting, this strategy only requires measurements of PPFD made at canopy level, dimmable LED lights, and a simple microcontroller. The algorithm controls greenhouse lighting to a target DLI with reasonable accuracy and can be readily implemented.