Vibration analysis and structure optimization of grain cleaning screen based on VMD

Cleaning is one of the most critical steps in the harvesting process. The cleaning sieve is a vital component in a combine harvester. However, the unbalanced vibration can often occur during machine operation, due to the processing and assembly faults with the complex external excitation. The wear of the key components can be accelerated to reduce the cleaning efficiency, even resulting in an increase in the loss rate and impurity rate during the harvesting. This study aims to focus on the unbalanced vibration that caused by the unbalanced rotor system during the operation of the grain cleaning screen. Taking the cleaning system as the research object, vibration testing and analysis were conducted with the crank slider serving as the driving mechanism, the louver lighting sieve as the upper screen, and the woven screen as the lower screen. Structure parameters were then optimized to reduce the unbalanced vibration. The air-screen combination grain cleaning system was used as the research object to tackle the unbalanced vibration that induced by the unbalanced rotor system during the operation of the grain cleaning screen. Firstly, the transmission mode was determined using working principle. Secondly, the DH5902 dynamic signal acquisition instrument was used to conduct the vibration testing on the system. The vibration signal was collected at the bearing seat of the driving mechanism rotor system. Thirdly, the time- and frequency-domain analysis was carried out to calculate the root mean square (RMS) and power spectral density. The vibration intensity was measured on the major frequency components. An aggregative indicator was constructed using variational mode decomposition (VMD), fuzzy entropy, and kurtosis for the feature extraction and analysis of the signal. The sensitive component of the original signal was calculated for the envelope spectrum using the feature component with the minimum comprehensive index. As such, the unbalanced vibration was extracted from the system. Finally, the optimization function was constructed to optimize the driving mechanism using the penalty function method under the Matlab computing platform, where the mass and installation position of the counterweight block were taken as the variables. The results show that the VMD signal decomposition feature extraction and the comprehensive index formed by kurtosis and fuzzy entropy were quickly retrieved the feature components of the signal using the ideal mode component u and the envelope spectrum. The smallest comprehensive index K was obtained in the values of 0.706 and 0.241 for the intrinsic mode components u₁-u₇ after the signal VMD at the bearing housing of the cleaning screen body, respectively. The envelope spectrum was also obtained for the initial signal sensitive component u₁. The uneven vibration was primarily distributed around 30.13 Hz, with the vibration intensities ranging between 0.24 and 0.29 dB, when the motor speed was 900 r/min. The weight block that optimized by the penalty function technique was weighed 3650 g as eccentric by 136.7 mm from the disk center under a vibration strength of 0.11 to 0.12 dB at 30.13 Hz, then was reduced vibration energy by 58.62%. The finding can be used to guide the vibration reduction analysis and structural optimization design of a grain combine harvester and new types of working parts.