基于人工神经网络的釉料厚度沉积率模型拟合软件模块的开发与实现

釉料厚度沉积率模型是机器人自动施釉的重要依据。为了进行陶瓷产品的机器人离线编程作业的轨迹自动规划,实现施釉过程中工件釉料厚度的精确性和均匀性,以釉料厚度沉积率模型为研究对象,采用MFC开发了釉料厚度沉积率模型软件功能模块。提出基于神经网络算法拟合沉积模型方案;利用COM形式的VC与MATALB的混合编程实现模型拟合、模型分析功能;设计动态链接库,使模型拟合结果产生word报表。通过试验进行试验验证,结果表明,其最大误差在5um范围之内,从而验证了模型的正确性和有效性。软件界面友好,符合工程实际,提出的方法有助于提高施釉机器人釉料厚度的控制精度,为后续的陶瓷自动施釉离线编程轨迹规划的软件编程和仿真实现提供了具体可操作的模型依据与方法指导。

SOFTWARE MODULE DEVELOPMENT AND REALIZATION FOR GLAZE DEPOSITION FITTING BASED ON ARTIFICIAL NEURAL NETWORK

The model for the deposition rate of glaze lays an important basis for automatic glazing of ceramics.For the automatic track-planning of the offline programming robot working on ceramic production line to ensure the accuracy and uniformity of glaze thickness,the software module fitting the glaze deposition model is developed with MFC.The fitting scheme is proposed based on neural network.Model fitting and analysis are realized by hybrid programming with VC and MATLAB based on COM.DLL file is developed to produce word report of the model.The software module is tested by experiments.The result shows that the error of the thickness model is within 5μm.The model built by the software module is correct and effectual.The software’s UI is friendly and satisfactory for project practice.The method increases the control over the accuracy of glazing thickness.The paper provides a specific theoretical and methodological support for the realization of process planning and simulation system in ceramic glazing.It will make the future developed system meet the actual processing requirement.