New Methods of Extracting and Eliminating Background Noise in the Impact-Based Grain Flow Sensor

摘要

In order to explore the influence of background noise on the measurement accuracy of the impact-based grain mass flow sensor and its elimination methods, the study designed a dual-plate differential impact-based grain mass flow sensor test bench. Background vibration noises under different field conditions were simulated in the bench through frequency adjustment. Output signals of the measurement plate and reference plate of the dual-plate differential impact-based grain mass flow sensor were acquired synchronously under 20 Hz, 35 Hz, and 50 Hz mains frequency of the driving motor respectively. The effects of the main and median filtering method on extracting random noises were compared. In addition, the effects of the frequency domain differential and time domain differential signal processing methods on rejecting background noises were also compared. For the time domain differential, differences between the filtered results of the corresponding signals acquired synchronously from the measurement plate and reference plate were calculated. For the frequency domain differential, DFTs were carried out to the filtered signals firstly, differences between the transformation results of the corresponding signals acquired synchronously from the measurement plate and reference plate were calculated next, and IDFT was executed to the differential result finally. The results showed that the mean filtering has a better effect on random noise filtering in the output signals of the sensor, and the frequency domain differential method can better eliminate the background vibration noise of the impact-based grain mass flow sensor.