Regresar

Analysis of Calibration Transfer Strategies in Environmental Variable Monitoring Prototypes

Abstract:

Environmental measurement sensors often exhibit drift and require frequent calibration. Traditional methods can be expensive and time-consuming. This article explores the use of calibration transfer, a cost-effective alternative. Two identical sensor systems measuring temperature and CO concentration were built. Data from these systems were collected under various conditions and used for multivariate calibration with partial least squares regression (PLS). Subsequently, Piecewise Direct Standardization (PDS) and Spectral Space Transformation (SST) were employed for calibration transfer on one system. PDS achieved a 0.031% reduction in error compared to PLS, while SST showed a 0.533% increase. These results demonstrate the effectiveness of calibration transfer, particularly PDS, in maintaining sensor accuracy while reducing the calibration time and cost.