DEVELOPMENT OF STRATEGIES FOR TEMPERATURE AND CO2 CONTROL IN THE GREENHOUSE PRODUCTION OF CUCUMBERS AND TOMATOES BASED ON MODELBUILDING AND OPTIMIZATION

ISHS Acta Horticulturae 260: International Symposium on Growth and Yield Control in Vegetable Production

DEVELOPMENT OF STRATEGIES FOR TEMPERATURE AND CO₂ CONTROL IN THE GREENHOUSE PRODUCTION OF CUCUMBERS AND TOMATOES BASED ON MODELBUILDING AND OPTIMIZATION

Authors:

K. Reinisch, E. Arnold, A. Markert, H. Puta

Abstract:
The ideal of every process control consists in doing this optimally according to predetermined objectives. In greenhouse production publications the adjective optimal has been used some times. These strategies are based on the knowledge of experts and on experiences from simulation runs (scenario techniques). They are not optimal in a stronger sense. In the case of the complicated dynamic processes in question and dynamic criteria, it is hardly possible to get optimal solutions in this way.

Optimal strategies will be got by applying optimization, especially optimal control. For this we need:

A quantitative model describing the outputs (the yield of cucumbers/tomatoes) depending on the controllable (temperature and CO₂-concentration) and the uncontrollable inputs (radiation) in analytic or algorithmic form.
A criterion (several criteria) evaluating quantitatively the fulfilment of the objective(s).
A prediction of the environmental inputs (global radiation) as far as available. The optimal strategy for a dynamic problem depends for every t on the environmental inputs in the whole optimization interval (growth period).

When starting 10 years before the interdisciplinary cooperation with the Inst. of Vegetable Prod., we proposed to apply methods of optimal control (probably for the first time) to greenhouse climate control, beginning with cucumber and including tomato production later on. Simulation runs and several tests in the greenhouse confirm that considerable increases in (especially early) yield can be reached. In the following, improvements in modelling and optimization will be reported which have been obtained

Full Text (PDF format, 355593 bytes)
Citation

Download Adobe Acrobat Reader (free software to read PDF files)