中文摘要：

针对机采棉中残地膜与棉花分离容易受到残地膜质量、机采棉飞入速度及强静电场电压等因素的影响,提出一种机采棉清杂过程中棉花与残地膜运动规律的预测方法.以南疆机采棉新陆早26号为研究对象,利用棉花与残地膜之间的分离率来表征运动规律.利用残地膜质量级别（1、2、3、4级）、荷电后的机采棉飞入速度（9、10、11、12m/s）、强静电场电压（30、40、50、60kV）3个因素作为BP神经网络模型的输入量,再利用改进的遗传算法训练所设计的网络的权值和阈值,建立能表征机采棉运动规律的分离率的预测模型.试验结果表明：利用改进遗传算法结合BP神经网络得到的机采棉与残地膜分离率预测模型能够很好地反应新陆早26号机采棉与残地膜分离应力与主要控制因素之间的非线性关系,预测结果与实测结果之间的平均绝对百分比误差为0.82,测试的样本实测值与理想值之间的相关系数为0.91751,所得到的预测模型效果良好,可为机采棉清杂提供参考.

英文摘要：

For cotton and cotton to plastic film separation by plastic film quality, cotton picker machine speed, flying into the strong static electric field voltage influence factors such as machine mining, prediction method of cleaning process of cotton and plastic film movement presents a cotton picker machine.The southern machine picked cotton Xinluzao 26 as the research object, to characterize the rate of movement by the separation between cotton and plastic film.The use of plastic film quality level （level 1, level 2, level 3, level 4）, electric charge after the cotton picking machine fly into speed （9, 10, 11, 12m/s）, the strong static electric field voltage （30, 40, 50, 60kV） three factors as the input of BP neural network model.The weights and thresholds of the left transmission algorithm training network design improvement.A prediction model for the separation rate of the movement law of cotton picking is established.The experiment results show that the improved genetic algorithm combined with cotton and plastic film separation rate prediction model can reflect the nonlinear relationship between Xinluzao 26 cotton picker machine and plastic film separation stress and main controlling factors of the BP neural network by machine, the mean absolute percentage error between predicted and measured results for 0.82 the measured sample test, the correlation coefficient and the ideal value is 0.91751, the effect of prediction model are good, the results of this study can provide a reference for cleaning for cotton picker.