Analysis of optimum drop size distribution of greenhouse air-assisted mist sprayer by using response surface method and genetic algorithm
Zhong Liping1,2, Qi Lijun1,3, Ji Ronghua4, Yuan Xue5, Mei Yincheng1, Gao Chunhua1
1.College of Engineering, China Agricultural University, Beijing 100083, China; 2.College of Mechanical and Control Engineering, Guilin University of Technology, Guilin, Guangxi 541004, China; 3. Beijing Key Laboratory of Optimization Design of Modern Agricultural Equipment, Beijing 100083, China; 4.College of Information and Electrical Engineering, China Agricultural University, Beijing 100083, China; 5.Chinese Academy of Agricultural Sciences, Beijing 100083, China
Abstract:In order to clarify the influences of operational parameters on drop size distribution issued by a greenhouse air-assisted mist sprayer, a series of spraying experiments are conducted, then two response surface models about the drop size distribution coefficient are correlated to the operational parameters by using quadratic polynomial regression and BP neural network method to obtain the optimum operational parameters combination. It is shown that the correlation coefficient and root-mean-square error are 0.987 1 and 0.134 0 respectively for the BP neural network model, compared with 0.928 2 and 0.215 9 for the response surface model established by the quadratic polynomial regression, suggesting the former is the best model. According to the quadratic polynomial response surface model, the minimum drop size distribution coefficient is as high as 1.47%, compared with 1.58% experimental measurement. Based on the BP neural network model, however, the minimum drop size distribution coefficient predicted by genetic algorithm is just 1.21% against 1.28% experimental observation. Ob-viously, the BP neural network method combined with genetic algorithm exhibits a better accuracy than the response surface method in the operational parameters optimization of air-assisted mist sprayer.
钟丽平,, 祁力钧,, 冀荣华, 袁雪, 梅银成, 高春花. 基于RSM和GA的温室风送式弥雾机喷雾效果优化分析[J]. 排灌机械工程学报, 2014, 32(9): 788-794.
Zhong Liping,, Qi Lijun,, Ji Ronghua, Yuan Xue, Mei Yincheng, Gao Chunhua. Analysis of optimum drop size distribution of greenhouse air-assisted mist sprayer by using response surface method and genetic algorithm. Journal of Drainage and Irrigation Machinery Engin, 2014, 32(9): 788-794.
[1]王永菲,王成国.响应面法的理论与应用[J]. 中央民族大学学报:自然科学版,2005,14(3): 236-240. Wang Yongfei,Wang Chengguo.The application of response surface methodology[J]. Journal of the CUN:Natural Sciences Edition,2005,14(3): 236-240.(in Chinese)[2]李烜,何雄奎,仲崇山,等. 荷电雾滴沉积效果的多因子响应面模型[J].高电压技术,2007,33(2):32-36. Li Xuan, He Xiongkui, Zhong Chongshan, et al. Effect of different spray factors on charged droplet deposit using response surface methodology[J].High Voltage Engineering, 2007,33(2):32-36.(in Chinese)[3]袁雪,祁力钧,王虎,等. 温室摇摆式变量弥雾机喷雾参数响应面法优化[J].农业机械学报,2012,43(4):45-50. Yuan Xue, Qi Lijun, Wang Hu,et al. Spraying parameters optimization of swing, automatic variables and greenhouse mist sprayer with response surface method[J].Transactions of the Chinese Society for Agricultural Machinery,2012,43(4): 45-50.(in Chinese)[4]林惠强,肖磊,刘财兴,等. 果树施药仿形喷雾神经网络模型及其应用[J].农业工程学报,2005,21(10): 95-99. Lin Huiqiang,Xiao Lei,Liu Caixing,et al.Neural network model for profile modeling spray of chemical to fruit trees and its applications[J].Transactions of the CSAE, 2005,21(10): 95-99.(in Chinese)[5]张富贵,洪添胜,肖磊,等. 果树冠幅的检测机理研究[J].农业工程学报,2008,24(4):25-29. Zhang Fugui,Hong Tiansheng,Xiao Lei,et al.Detection mechanism for fruiter crown diameter[J]. Transactions of the CSAE,2008,24(4):25-29.(in Chinese)[6]Tahmasebi M, Rahman R A, Mailah M,et al. Roll movement control of a spray boom structure using active force control with artificial neural network strategy[J]. Journal of Low Frequency Noise, Vibration and Active Control,2013,32(3):189-202.[7]Vitalijs Komasilovs, Egils Stalidzans, Vitalijs Osadcuks,et al. Specification development of robotic system for pesticide spraying in greenhouse[C]//Proceedings of the 14th IEEE International Symposium on Computatio-nal Intelligence and Informatics, 2013.[8]Hmyers R H. Response surface methodology—Current status and future directions[J].Journal of Quality Technology,1999,31(1):30.[9]郭勤涛, 张令弥, 费庆国. 用于确定性计算仿真的响应面法及其试验设计研究[J].航空学报,2006,27(1):55-60. Guo Qintao, Zhang Lingmi, Fei Qingguo. Response surface method and its experimental design for deterministic computer simulation[J]. Acta Aeronautica et Astronautica Sinica, 2006,27(1):55-60.(in Chinese)[10]杨晓华. 智能算法及其在资源环境系统建模中的应用[M].北京:北京师范大学出版社,2005.[11]阎平凡,张长水. 人工神经网络与模拟进化计算[M].北京:清华大学出版社,2000.[12]李俭川,秦国军,温熙森,等.神经网络学习算法的过拟合问题及解决方法[J].振动、测试与诊断,2002,22(4):260-264. Li Jianchuan,Qin Guojun,Wen Xisen,et al. Over-fitting in neural network learning algorithms and its solving strategies[J].Journal of Vibration Measurement and Dia-gnosis,2002,22(4):260-264.(in Chinese)[13]徐敬,王秀坤,胡家升.基于神经网络的无源多传感器属性数据关联[J].系统仿真学报,2003,15(1):127-131. Xu Jing, Wang Xiukun, Hu Jiasheng. Multiple passive sensors feature data association based on neural networks[J].Journal of System Simulation, 2003,15(1):127-131.(in Chinese)[14]Bingöl D, Hercan M, Elevli S, et al. Comparison of the results of response surface methodology and artificial neural network for the biosorption of lead using black cumin [J].Bioresource Technology,2012, 112: 111-115.[15]哈根.神经网络设计[M].北京:机械工业出版社, 2002.[16]Blanco A, Delgado M, Pegalajar M C. A real-coded genetic algorithm for training recurrent neural networks[J].Neural Networks, 2001, 14(1):93-105.