Research on temperature drift compensation method for piezoresistive pressure sensors based on multivariate differential algorithm

Purpose Due to the temperature sensitivity of intrinsic parameters such as the piezoresistive coefficient and resistivity, silicon-based piezoresistive pressure sensors exhibit significant temperature drift in output signals under high-temperature conditions. This leads to a sharp decline in measurement accuracy and severely constrains their application in high-precision measurement fields. To address this technical bottleneck, this paper aims to propose a low-temperature-coefficient passive resistor network compensation method based on a multivariate differential algorithm. Design/methodology/approach By establishing coupled equations for the sensor’s output characteristics, the authors analyze its nonlinear features and introduce a dimensionality reduction strategy. This approach simplifies the differential equations, enabling the solution of compensation resistor parameters using only the resistance values of the four bridge arms at threshold temperature and pressure. This significantly reduces data acquisition and processing complexity. To validate the method’s effectiveness, a comprehensive temperature-pressure calibration platform (25–235°C, 0–1 MPa) was constructed for pre-compensation and post-compensation sensor calibration. Findings Results demonstrate that after compensation, the zero-point temperature drift reaches 0.0781%FS/°C, and the sensitivity temperature drift reaches 0.0717%FS/°C. Compared with conventional methods, zero drift is reduced by 23% and sensitivity drift by 30%, substantially enhancing the sensor’s high-temperature measurement stability. Originality/value This study introduces a low-temperature-coefficient passive resistor network compensation method using a multivariate differential algorithm. The key originality lies in the dimensionality reduction strategy, which simplifies the differential equations and allows determination of compensation parameters with minimal data – specifically, the resistance values at threshold temperature and pressure. This innovation significantly reduces both data acquisition and computational complexity compared to traditional approaches.