Modeling and identification of the hygro-thermal parameters of mechanically-ventilated broiler house using prediction error: Assessment in cold conditions

Abstract

Indoor air temperature and humidity moisture are of the foremost significance in climate control of broilers houses, and their impacts on poultry health and production depend on accurate control. The main objective of this work is to identify and assess a novel state-space model, to rapidly predict the hygro-thermal behavior of the livestock building. To achieve this analysis, various experimental measurements (e.g., ventilation rate, thermal heating, and air temperature and humidity) of two commercial poultry houses placed in the Mediterranean zone were monitored over cold conditions production cycle. The developed model was estimated and validated against a dataset of 25 days acquired under three different operation ventilation modes (min-ven, power and tunnel modes). Through simulation, the results showed that the predicted model and measured data were achieved a satisfactory accuracy with an averaged coefficients of determination R2 were 0.93 and 0.95, respectively, for the indoor air temperature and humidity models, and a root mean squared error (RMSE) of 0.3213 °C and 0.957 %. Additionally, the predictive model shows satisfying performances for the long horizon prediction with a final prediction error (FPE) equal to 0.084, which will prevent the intensely time-consuming process of getting precise physical parameters in regards the poultry house system.