Temperature Compensation for MEMS Mass Flow Sensors Based on Back Propagation Neural Network

Abstract

A temperature compensation method based on Back Propagation (BP) Neural Network is designed by taking advantage of the characteristics of neural network, whose performance is demonstrated in a mass flow sensor. The mass flow sensor is tested under different temperatures to obtain the sample data. The algorithm compensates the temperature effects by establishing a non-linear mapping relationship between temperature and mass flow rate. To settle the problem of accuracy degradation induced by ambient temperature variation, both BP neural network and polynomial fitting method are developed to compensate the drift of the mass flow sensor. The result shows that the method based on the BP Neural Network has high compensation accuracy and fast convergence speed, which can effectively compensate the influence of temperature on MEMS mass flow sensor and improve the sensor output accuracy.