1、2023，45（2）DOI：10.13836/j.jjau.2023045江西农业大学学报 Acta Agriculturae Universitatis Jiangxiensishttp:/气吸圆盘式排种器智能设计与优化系统研究韩晓娟，赖庆辉*，赵庆辉，李沛航（昆明理工大学 现代农业工程学院，云南 昆明 650500）摘要：【目的】气吸圆盘式排种器广泛应用于大型精密播种机，能够有效提高农作物播种效率、降低农业生产成本，促进我国农业生产发展。针对排种器传统设计方法中计算量大、设计效率低、研发周期长、试验和制造成本高等问题，构建了一套气吸圆盘式排种器智能设计与优化系统。【方法】基于对气吸圆盘式排
2、种器的设计，采用基于本体式和产生式相结合的知识表示方法，建立设计知识库；采用基于规则和实例相混合的推理机制构建推理机；以Visual Studio编程软件为平台，利用VB.Net为开发语言对系统的人机交互界面进行设计；采用动态超链接库模式对SolidWorks软件进行二次开发，基于ADO.Net技术建立智能设计系统与MySQL数据库之间的链接；以吸室真空度、吸孔直径和吸孔个数为优化变量，合格指数、重播指数和漏播指数为评价指标，构建了DEM-CFD气固耦合的仿真平台，通过开展三因素五水平二次回归正交旋转组合仿真试验，建立优化变量与合格指数、重播指数、漏播指数之间的数学模型。利用MATLAB软件并
3、基于NSGA-II算法对此数学模型进行多目标优化并得到一组Pareto非劣解集，选取合格指数最高的一组解对排种器模型进行结构优化，并对优化后的排种器模型进行仿真验证，进而将优化后的模型存入模型库。【结果】利用该系统智能设计并优化了气吸圆盘式大豆排种器，获得了最佳的结构模型和作业参数。与优化前相比，合格指数提高1.23%，重播指数降低18.44%，漏播指数降低39.76%。【结论】该智能设计与优化系统可以有效提高气吸圆盘式排种器的设计知识的重用率，缩短研发周期并根据用户需求自动生成排种器模型，大大提高了气吸圆盘式排种器的设计效率。关键词：气吸圆盘式排种器；智能设计；参数化建模；DEM-CFD耦合
4、仿真；多目标优化中图分类号：S223.2 文献标志码：A 开放科学（资源服务）标识码（OSID）：文章编号：1000-2286（2023）02-0467-15Research on the Air Suction Disc Seed Metering Device s Intelligent Design and Optimization SystemHAN Xiaojuan，LAI Qinghui*，ZHAO Qinghui，LI Peihang（Kunming University of Science and Technology，College of Modern Agricultur
5、al Engineering，KunMing 650500，China）Abstract：Objective Widely used in large precision seeders，air suction disc seed metering device can effectively improve crop sowing efficiency，reduce agricultural production costs and promote the development of agricultural production in China.To solve the problem
6、s of large amount of calculation，low design efficiency，收稿日期：20221122 修回日期：20230205基金项目：国家自然科学基金项目（51975265，52165031）Project supported by the National Natural Science Foundation of China（51975265，52165031)作者简介：韩晓娟，硕士生，orcid.org/0000-0001-9503-8957，han_；*通信作者：赖庆辉，教授，博士，博士生导师，主要从事农业机械装备与计算机测控研究，orcid.o
7、rg/0000-0003-0885-6463，。韩晓娟，赖庆辉，赵庆辉，等.气吸圆盘式排种器智能设计与优化系统研究 J.江西农业大学学报，2023，45（2）：467-481.HAN X J,LAI Q H,ZHAO Q H,et al.Research on the air suction disc seed metering device s intelligent design and optimization systemJ.Acta agriculturae universitatis Jiangxiensis,2023,45(2):467-481.江 西 农 业 大 学 学 报第
8、45 卷long research and development cycle and high test cost in the traditional design method of seed metering device，an intelligent design and optimization system of air suction disc seed metering devices was constructed.MethodBased on the design of an air suction disc seed metering device，the knowle
9、dge representation method based on ontology and production was adopted to establish the comprehensive design knowledge base.To construct the inference engine，a reasoning mechanism which combines RBR（rule-based reasoning）and CBR（case-based reasoning）was employed.The user-machine interface of the syst
10、em has been designed by using Visual Studio programming software as the platform and VB.Net as the development language.Solid Works soft ware was redeveloped by using dynamic hyperlinked library mode，and the link between the intelligent design system and MySQL database was established based on ADO.N
11、et（ActiveX Date Objects）technology.Taking the vacuum degree of the suction chamber，the diameter of the suction hole，and the number of suction holes as optimization variables and the qualified index，reseeding index，and missed-seeding index as evaluation indexes，a DEM-CFD gas-solid coupling simulation
12、 platform was constructed.Through the quadratic regression orthogonal rotation combination simulation test of three factors and five levels，a mathematical model was established between the optimization variables and the evaluation indexes.The mathematical model was multi-objectively optimized by usi
13、ng MATLAB software and the NSGA-IIalgorithm（Non-dominated Storing Genetic Algorithm-II），leading to a set of Pareto non-inferior solutions.The solution with the highest qualified index was selected to optimize the structure of the seed metering model，which was simulated and verified，before being stor
14、ed in the model base.Result The system was used to intelligently design and optimize the air suction disc soybean seed metering device，and the best structural model and operating parameters were obtained.the qualified index had an increase of 1.23%，the replay index had a decrease of 18.44%，and the m
15、issed broadcast index was dropped by 39.76%.Conclusion The intelligent design and optimization system can effectively enhance the reuse rate of the design knowledge of the seed metering device，significantly reduce the research and development cycle，and automatically generate the model of the seed me
16、tering device according to the user s requirements，thus significantly improving the design efficacy of the air suction disc seed metering device.Keywords：air suction disc seed metering device；intelligent design；parametric modeling；DEM-CFD coupling simulation；multi-objective optimization【研究意义】排种器是精密播种机的核心部件，气吸圆盘式排种器适合高速作业，具有不伤种，对种子适应能力强等优点，大型高速精密播种机广泛采用1。气吸圆盘式排种器在设计过程中，需综合应用机械学、颗粒力学、流体力学、多刚体运动学等学科知识。传统设计方法存在设计周期长、效率低、设计知识重用率低等问题。随着计算机技术快速发展，智能化设计可有效解决传统设计方法的缺点。因此，探索气吸圆盘式排种器智能设计与优化十分必要。【前人研究进展】目前，
