Journal : Biosystems Engineering, 229, 1-17
Author : Uk-Hyeon Yeo, Sang-Yeon Lee, Se-Jun Park, Jun-Gyu Kim, Deuk-Young Jeong, Cristina S. Decano-Valentin, Rack-Woo Kim, In-Bok Lee*
Abstract
Monitoring inside pig houses for environmental control is usually conducted using either a single or a small number of sensors. It is therefore difficult to monitor the representative rearing environment and provide appropriate control strategies. It is necessary to determine proper monitoring locations to precisely monitor and then control the environment inside pig houses. Thus, this study aimed to identify the appropriate sensor location according to the number of installed sensors that can represent the entire environment inside a pig house and detect the locations where internal changes occurs that are greatly affected by the external environment. Optimal locations for each number of sensors were assessed by the application of an error-based method (EBM) to select monitoring locations that can represent the entire environment of the pig house and an information-entropybased method (IEBM) to select monitoring locations with significant environmental changes. When selecting air temperature monitoring locations with the EBM during the entire experimental period, the middle sensor installed along the corridor (P-10) indicated the air temperature similar to the average air temperature of all sensors installed inside the pig house. Root mean square error (RMSE) and mean absolute percentage error (MAPE) of the P-10 location showed errors of 0.9 °C and 2.9%, respectively. With the installation of three sensors, accurate air temperature monitoring was possible within MAPE of 1%. The sensor location selected with the IEBM was near the pig house entrance (P-16). The sensor location showed errors of 2.5 °C and 13% for RMSE and MAPE, respectively.
Keywords : Air temperature; Monitoring; Placement; Pig house; Sensor location
Download Link : https://doi.org/10.1016/j.biosystemseng.2023.03.010