【作者】 黎敏；

【导师】 王秋莲；

【作者基本信息】 南昌大学 ， 工业工程（专业学位）， 2022， 硕士

【摘要】 工业是我国国民经济的主导产业、能量资源消耗和环境污染排放的重点领域。高能量消耗不利于机械制造业成本的控制,而且碳排放等污染造成居住质量下降。基于绿色制造理念,通过分析加工过程能耗特性,测算各加工时段能量效率,以冀能为实现机床节能目标提供理论依据,促进整体机械制造业绿色制造、节能减排水平的提升。本文提出一种基于GA-DBN的多品种小批量机械加工过程能量效率测算模型。研究内容包括:(1)针对与功率信号能量消耗密切相关的加工状态会在递归图上展现不同的纹理的特点,提出基于递归分析的多品种小批量机械加工过程功率信号二维图像表达方式。实验证明,该方法能够有效识别功率曲线上表示加工状态变动的点集合。(2)利用遗传算法(Genetic Algorithm,GA)全局搜索寻优的特点,优化深度信念网络(Deep Belief Network,DBN)模型初始化参数,并建立GADBN加工时段判定模型,以达到合理设置参数并进一步提高模型分类精度的目的。为后续能量效率优化、机床节能提供基础数据。(3)通过实例分析,将实时功率信号分别输入到DBN模型和GA-DBN模型中进行工件判定和加工时段判定。结果证明,GA-DBN模型相对于DBN模型,收敛速度更快,且判定精度更高。本文所提模型能够准确判定机械加工过程的各加工时段,精度达到95.86%,且由模型计算的加工过程能量效率与实际能量效率误差为3.96%,具有应用价值,能够为实现机械加工系统绿色制造目的提供理论依据。更多还原

【Abstract】 Industry is the leading industry of my country’s national economy,the key area of energy resource consumption and environmental pollution discharge.High energy consumption is not conducive to the control of the cost of machinery manufacturing,and pollution such as carbon emissions reduces the quality of living.Based on the concept of green manufacturing,by analyzing the energy consumption characteristics of the processing process,the energy efficiency of each processing period is measured,in order to provide a theoretical basis for realizing the energy saving target of machine tools,and promote the improvement of green manufacturing,energy saving and emission reduction in the overall machinery manufacturing industry.In this paper,a GA-DBN-based energy efficiency measurement model for multi-variety and smallbatch machining processes is proposed.Research content includes:(1)Aiming at the fact that the processing state closely related to the energy consumption of the power signal will show different textures on the recurrence plot,a two-dimensional image representation of the power signal in the multi-variety and small-batch machining process based on recurrence analysis is proposed.Experiments show that this method can effectively identify the point set on the power curve that represents the change of the machining state.(2)Using the characteristics of genetic algorithm(Genetic Algorithm,GA)global search optimization,optimize the deep belief network(Deep Belief Network,DBN)model initialization parameters,and establish a GA-DBN processing period judgment model,in order to achieve reasonable parameter settings and the purpose of further improving the classification accuracy of the model.Provide basic data for subsequent energy efficiency optimization and machine tool energy saving.(3)Through example analysis,the real-time power signal is input into the DBN model and the GA-DBN model respectively for workpiece judgment and processing period judgment.The results show that the GA-DBN model has faster convergence speed and higher judgment accuracy than the DBN model.The model proposed in this paper can accurately determine each machining period of the machining process,with an accuracy of 95.86%,and the error between the energy efficiency of the machining process calculated by the model and the actual energy efficiency is 3.96%,which has application value and can be used for green manufacturing of machining systems.The purpose is to provide a theoretical basis.更多还原