A Novel Model Fusing ALA and Integrated Learning: Temperature Compensation for 700 kPa Pressure Scanners

Abstract

The pressure scanner, as a highly integrated multi-channel pressure acquisition device, plays a crucial role in the high-precision measurement process. However, since pressure scanners are extremely sensitive to ambient temperature, improper calibration may directly affect the accuracy of pressure data. Therefore, the use of accurate temperature compensation algorithms is particularly important for pressure scanners. This study proposes a novel hybrid temperature compensation algorithm that combines the advantages of the Extreme Gradient Boosting (XGBoost) model in integrated learning and the artificial lemming algorithm (ALA). The algorithm is validated using experimental data obtained from a self-built calibration system. A side-by-side comparison is made with other machine learning models such as Random Forest (RF), Support Vector Regression (SVM), and XGBoost. The results show that the ALA-XGBoost model performs well with R² of 0.99951, RMSE of 0.0226, and full-scale error of only 0.081 % FS. The results validate the feasibility and effectiveness of the integrated learning model for pressure scanners temperature compensation applications.