Abstract
Measuring with Coriolis flowmeter under the condition of gas-liquid two-phase flow, the vibration signal frequency, phase and amplitude of the measuring pipe fluctuate randomly. The traditional phase difference extraction algorithm is difficult to ensure the measurement accuracy of mass flow. In order to improve the extraction accuracy of mass flow, this paper presents a dynamic error correction method of mass flow based on LSTM (Long Short-Term Memory). The water flow test platform and data acquisition system are built, and the mass flow corrected by this method is compared with the mass flow corrected by ARMA (Autoregressive moving average) model and the mass flow obtained by flow meter. The experimental results show that the mean error of mass flow corrected by this method is 0.0357, and the variance is 0.0453, which is far less than the static error correction of 0.0988 and 0.1068, and far less than the mass flow meter of 0.3816 and 0.4045.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have