DESARROLLO DE UN MODELO AUTO-REGRESIVO CON ENTRADAS EXÓGENAS PARA LA PREDICCIÓN DE LA TEMPERATURA EN UN INVERNADERO DEL ESTADO DE OAXACA

Abstract

In the present work a dynamic linear model of auto-regression with exogenous variables (ARX) was developed which predicts the behavior of the air temperature inside a zenith greenhouse located in Santa Cruz Xoxocotlán, Oax., At an altitude of 1 530 masl and latitude 17 ° 01'35 "N, with a semi-hot arid climate, according to—classification. Measurements of air temperature, solar radiation, wind speed and relative humidity were taken outside of greenhouse and as the unique variable the air temperature was monitored inside. These data were taken with a DAVIS weather station located at height of 2 meters above of greenhouse zenith and the interior air temperature in the geometric center every minute in August of 2011. With the outside measurements, the model calculated the interior air temperature and it is compared with measured internal temperature till to achieve the optimal adjustment factor between both temperatures. The better structure was obtained generating 100 000 ARX models using a combination of process variables, later carry out them with Akaike’s final prediction error and residual analysis. The RLS allowed maintaining a right adjustment coefficient even varying the weather conditions in different seasons, through periodic recalibration of vector of parameters. As conclusion, the fifth-order or superior ARX models were useful to predict the temperature inside the greenhouse and the RLS, to maintain goodness of fit even with variation of characteristic weather conditions of every season.