超高强度钢切削物理数据服务技术研究

Typical material cutting physical database is the data foundation to realize the intelligent manufacturing and precision manufacturing. With the analysis of the development of database technology and practical application for the cutting of ultra-high strength steel, the architecture of cutting physical data service platform will be designed and the database system oriented cloud manufacturing and big data will be developed. The cutting mechanism of ultra-high strength steel and the characteristics of cutting physical data will be analyzed to reveal the interaction rules of the cutting physical parameters and to form the cutting data feature specification and acquisition mode. The topology model based on the vector space and the cutting knowledge spectrum for multi-source heterogeneous cutting data and knowledge will be built. The data mining and visualization technology will be researched to find the process knowledge and the association rules in ultra-high strength steel cutting with the coupling control of the part shape and performance. Some scientific problems such as the design theory of cutting data service system oriented big data, the space-time mapping model and visualization technology of the complex high dimensional multivariate cutting data, the cutting data mining and knowledge discovery technology will be expounded theoretically. As a result, the design principle and cutting data specification can be used as the theoretical basis and guarding method for the typical material cutting database system with proprietary intellectual property rights.

典型材料切削物理数据库是智能制造实现的数据基础，是机械制造走向定量精准制造的必由之路。本申请拟以超高强度钢切削基础数据库技术发展和实际应用需求为牵引，以切削物理数据为核心，开展切削加工基础数据服务平台顶层架构设计，开发面向云制造、大数据的切削物理数据服务平台；通过研究超高强度钢切削机理，分析切削物理数据特征，揭示切削过程物理参量的作用规律，形成切削物理数据特征规范和获取模式；研究多源异构切削数据与知识拓扑结构建模，建立基于向量空间的切削数据知识拓扑结构模型和切削知识型谱；研究面向形性耦合控制的切削数据挖掘与可视化技术，挖掘形性一体化控制工艺知识与关联规则。拟从理论上阐明面向大数据的切削基础数据服务架构方法、复杂高维多元切削数据的时空映射表达与可视化、切削物理数据作用规律的聚类挖掘与知识发现等关键科学问题，为开发形成具有我国自主知识产权的典型材料切削数据库提供系统的基础理论依据和方法指导。

切削物理数据作为工艺信息的载体，由于动态时变、多源异构、耦合机制复杂等特点，难以保证加工表面的形性耦合精准控制，切削物理数据的分析与知识发现是实现智能制造的基础性关键技术。本项目针对超高强度钢切削过程中的多源异构切削物理数据知识描述困难、相互作用机理不清等共性问题，开展了面向云制造的切削数据服务平台顶层架构设计，多传感器融合的切削过程监测系统，多源异构切削数据获取规范与融合理论、量化表征与知识发现的研究；以切削数据元为单位，构造了多源异构切削数据间的约束矩阵集合，提出了空间时变切削信号的小波分析方法、加工表面纹理的图像灰度共生矩阵和功能性统计函数等量化表征方法，揭示了超高强度钢切削过程物理参量的作用规律与关联规则；基于时-频分析与深度学习算法，建立了基于累积和图的刀具破损失效分析模型、基于支持向量机的刀具磨损状态预测模型、基于深度置信网络的加工表面形性耦合控制的智能决策模型，开发了超高强度钢基础切削数据服务平台，可提供切削物理数据的结构化存储与分析、刀具磨损状态监测、表面完整性表征参数控制、切削参数推荐等智能化服务，从而为提高高端制造装备的加工效率和智能制造水平，提供基础理论依据和方法指导。