Design and validation of a low-cost data-driven poultry heat and moisture production monitoring system

Poultry are the primary source of heat and moisture in confined livestock housing systems. Accurate measurement of poultry heat and moisture production (HMP) is critical for intelligent and sustainable livestock farming, including effective environmental control, energy analysis, and reliable facility simulations. Currently, the indirect calorimetry method is widely applied, but it has estimation uncertainties and high construction costs. Meanwhile, the direct calorimetry method, despite its higher accuracy, is limited by complex equipment design and its sensitivity to environmental variations. This study developed an innovative data-driven poultry HMP monitoring system designed to improve measurement accuracy while reducing operational complexity and costs. The monitoring system comprises a poultry rearing chamber and a data acquisition subsystem. To accurately monitor the HMP inside the chamber, a dynamic heat and moisture prediction (DHMP) model was developed, and its parameters were identified by integrating experimental data with the Non-dominated Sorting Genetic Algorithm II (NSGA-II). Experimental data were collected under various heating and humidification power settings and ambient temperature conditions to train and validate the DHMP model. The results demonstrate that the DHMP model has good adaptability to ambient temperature variations across different heating power conditions. In validation datasets, the mean absolute percentage errors for heating and humidification power predictions were 3.30 % and 3.71 %, respectively, with corresponding root mean square error values of 0.961 W and 0.389 g·h⁻¹. Field experiments further confirmed that the HMP values predicted by the system closely match those reported in the literature, supporting the reliability of the system. Based on cost statistics, the total manufacturing cost was reduced by approximately 50 %-80 % compared with existing calorimetry methods. The developed data-driven HMP monitoring system effectively overcomes the limitations of traditional calorimetry methods in terms of complexity and high costs, providing an innovative and practical approach for accurately monitoring poultry physiological parameters to support precision environmental control and production management.