Practical implementation and evaluation of optimal carbon dioxide supply control

In state-of-the-art production of greenhouse crops, the air is enriched to higher than atmospheric carbon dioxide levels to increase photosynthesis and final yield. When there is ventilation, this means that some of the supplied carbon dioxide will be lost. Especially when the available carbon dioxide is not a by-product of heat production (e.g. pure liquid carbon dioxide), the enrichment has a significant cost. Commonly, controlling carbon dioxide level is a simple on/off control depending on the windows opening. Experienced growers try to increase the benefit-cost ratio by using more complicated strategies based on influences of climatic factors, e.g. sun irradiation and wind speed. The settings for the influences are made intuitively and it is hardly possible to get any feedback on the applied strategy; therefore they do not warrant optimal use of carbon dioxide. This study demonstrates as alternative a model-based optimization algorithm. This algorithm estimates the carbon dioxide concentration where the difference between benefits and costs of carbon dioxide dosing is maximized. The greenhouse ventilation rate, as the major determinant of the costs, is estimated from the energy and vapor balance of the greenhouse. The photosynthesis, as the major determinant of the benefits, is simulated as a function of solar radiation and carbon dioxide. The model is implemented in the process computer of a commercial tomato greenhouse. The method gives a reasonable estimation of optimum carbon dioxide concentration in varying conditions. The estimated increase in weekly margin compared to the strategy of the grower could be over 200 € ha^-1