Regional airflow velocity control to meet the diverse requirements of unevenly distributed maize threshing mixtures

Abstract

Cleaning is a critical operation in maize production. Uneven lateral distribution of maize threshing mixture on the cleaning sieve surface is a common issue, which can lead to material overload or underload in certain regions, severely impacting the quality of the cleaning operation. This paper designs a combined cleaning fan, consisting of three fans that can independently control their speed, achieving precise airflow control over three regions of the cleaning sieve surface. A monitoring device is designed to detect the load distribution of maize threshing mixture in real-time, providing signals for the control of the combined fan. Material load, vibration frequency of the cleaning sieve, and sieve opening are used as the input layer, while fan speed serves as the output layer. A Gradient Boosting Regression (GBDT) fan speed prediction model is established to provide a data model for the precise control of the combined fan. A control system for the combined fan is developed, adjusting the fan speed in real-time based on model calculations. Cleaning performance tests showed that the combined fan control system precisely adjusted fan speed according to changes in material load, enabling differentiated operations for unevenly distributed threshing mixture, thereby achieving lower loss rates and impurity levels. Compared to traditional fans, the combined fan demonstrated superior operational performance. The combined cleaning fan improved airflow mismatch caused by uneven threshing mixture distribution, significantly enhancing maize cleaning performance. It also provides a valuable reference for the cleaning operations of other crops and the intelligent development of agricultural machinery.