面向知识自动化的磨矿系统操作员脑认知特征与控制效果的相关分析

发布时间：2018-06-12 14:40

本文选题：知识自动化 + 操作控制水平　；参考：《自动化学报》2017年11期

【摘要】：面向知识型工作自动化,研究了流程工业生产过程中操作人员的脑认知特征与操作控制水平之间的关键,建立了一种基于操作员脑网络特征的操作熟练程度隐性知识的显性化模型.采用关注信号瞬时相位、基于希尔伯特变换的相位锁方法,构建了脑功能网络(Functional brain network,FBN).基于磨矿系统操作员脑功能网络的图论参数与社区连接强度,建立了特征空间,采用支持向量机与神经网络进行特征分类.结果表明,在高频区,熟练操作员(熟手)的脑功能网络连接强度明显高于不熟练操作员(生手):在低频部分则生手的脑功能网络连接强度略高,其特征分类准确率为87.24%.磨矿系统操作过程中形成的溢流粒度(Grinding particle size,GPS)曲线可以初略地反映操作人员的熟练程度,本文在深入分析了其溢流粒度曲线与操作员脑网络特征的基础上,发现相对于溢流粒度曲线操作员的脑网络特征可以更全面地描述操作控制水平(特别在操作开始时间段),采用脑网络特征识别操作控制水平在时间上超前于溢流粒度曲线识别方法.本研究对于将知识工作者的认知特征引入到流程工业控制中,具有一定的借鉴意义.
[Abstract]:In this paper, the key factors between the cognitive characteristics of the operator and the level of operation control are studied in the process of process industry production, which is oriented to the knowledge type work automation. A dominant model of tacit knowledge of operational proficiency based on operator brain network features is established. Based on the phase locking method of Hilbert transform, the functional brain network is constructed by using the instantaneous phase of the concerned signal. Based on the graph theory parameters of the brain functional network of grinding system operator and the intensity of community connection, the feature space is established and the feature classification is carried out by using support vector machine and neural network. The results showed that in the high frequency region, the brain functional network connection intensity of the skilled operators (proficient hands) was significantly higher than that of the unskilled operators (unskilled operators: in the low frequency part, the brain functional network connection intensity was slightly higher, and the accuracy of the characteristic classification was 87.24%). The overflow granularity curve formed during the operation of the grinding system can reflect the proficiency of the operator at first. Based on the in-depth analysis of the overflow granularity curve and the characteristics of the operator's brain network, the overflow granularity curve and the characteristics of the operator's brain network are analyzed in this paper. It is found that the brain network features can describe the operation control level more comprehensively than the overflow granularity curve operator (especially at the beginning of the operation, the brain network feature is used to identify the operation control level ahead of the overflow level in time. Granularity curve recognition method. This study can be used as a reference for knowledge workers to introduce their cognitive characteristics into process industry control.
【作者单位】：东北大学机械工程与自动化学院;东北大学流程工业综合自动化国家重点实验室;曼彻斯特大学自动化中心;
【基金】：国家自然科学基金(51505069,61621004) 辽宁省高等学校创新团队项目(LT2014006) 流程工业综合自动化国家重点实验室开放基金(PAL-N201304)资助~~
【分类号】：TP18

【相似文献】