GREENHOUSE PARAMETER ESTIMATION BY RECURSIVE LEAST SQUARES

A least squares method (LSM) with extensions for on-line calculations was used to estimate greenhouse thermal parameters. Because certain parameters of a greenhouse are time varying or influenced by operating conditions, a loss function with exponential weighting, a "forgetting factor," is introduced into the LSM to eliminate slowly the influence of old data.A linear thermal model of an unventilated greenhouse, the parameters of which were estimated by the LSM, was applied to study the estimating accuracy and the effects of the forgetting factor. This LSM has the advantages of estimating several parameters simultaneously, requiring little computer memory, and having short calculation time requirements. Therefore, it would be applicable to a control algorithm and a design of the heating/cooling system.