Knowledge Modelling Of Manufacturing Equipment Capability And Its Application

Manufacturing equipment capability(MEC)is the ability that equipments reflect during the processing of products.Manufacturing capability is described as a kind of ability generated from the process of curtain manufacturing activities,acting on the configuration and integration of manufacturing resources,giving expression to the level of enterprise for accomplishing a targeted task and expected goal.From the respect of the granularity of resources,manufacturing capabilities can be divided into resource unit level,workshop level,factory level and enterprise level.The manufacturing capability of resource unit level primarily reflects the capacity exhibited by individual manufacturing resources.Manufacturing equipment is the main resource for performing production and processing,and MEC is the main embodiment of manufacturing capacity at the resource unit level.In addition to the inherent functional attributes of the manufacturing equipment itself,MEC also includes various types of the information of soft resource involved in the execution of manufacturing activities,such as strategies in shopfloor,operational skills,real status constraints and so on.Based on the National Natural Science Foundation of China,in this paper,knowledge modelling of MEC is studied to support large-scale collaborative manufacturing in cloud-based environment,taking two types of machining equipments(CNC machine and industrial robots)as the objects,which are applied widely in the most of current industrial enterprises.For the manufacturing capability modelling of CNC machine,a knowledge modelling method based on schema mapping is applied to generate the model automatically.And for the capability of industrial robots,a knowledge modelling method based on dynamic description logic is studied to construct the model.Among this,the specific definition of both the simple and complex actions are given in the model,and a semantic description of energy consumption is proposed using the interval status of action.Aiming at the automatic update and maintenance of the models,a self-learning evolution method based on the multi-level parallel association mining is proposed to discover correlative knowledge from technical documents.The primary content is listed as the following:(1)A knowledge modeling framework of MEC is studied.After analyzing the application requirements for the dynamic sharing and intelligent allocation in cloud manufacturing mode,the knowledge scope for MEC is determined according to the current research result about manufacturing capability.The related concepts of MEC and the relationships between them are collected and arranged.After then,the knowledge modeling framework is built integrating the theory and method of knowledge modelling.(2)Aiming to the manufacturing capability of CNC machine tools,the knowledge modeling method based on schema mapping is studied.According to the knowledge scope of MEC,the knowledge model for the capability of CNC machine tools is constructed by integrated a general data model named STEP-NC with status information in the shopfloor and domain expert knowledge.The schema mapping EXPRESS-OWL is studied to translate STEP-NC into OWL model automatically.For the limitation of EXPRESS-OWL,the terminologies in the model are further defined using ontology language.The SWRL rules are defined to describe the expertise in STEP-NC,which is omitted during the mapping procedure.Afterwards,the associated rules between machine tool and machining task of part are constructed to describe machine tool capability dynamically by rule reasoning.Finally,the constructed model is verified by consistency check and rule inference with CNC machine instances and rules.Meanwhile,the model is compared to other nine models that have similar application background to illustrate the characteristic and performance.(3)Aiming to the capability of industrial robot,the knowledge modeling method based on dynamic description logic is studied.The related concepts for industrial robot capability and the correlations between them are further collected according to the knowledge scope of MEC to construct the knowledge model.In this knowledge model,the dynamic description logic is applied to supply the explicit definitions to the simple and complex actions of industrial robot.In order to depict the energy consumption of these acitons,a semantic description based on interval status of action is proposed to give the segment identification to support the energy efficient application of industrial robot.Then,three type of semantic rules are defined for the judgement of industrial robot capability against task information.Finally,the consistency of the built model is verified by the description of industrial robot and the measured energy consumption and also the rule inference with task instance.Meanwhile,the model is compared with other five models to illustrate the characteristics and the limitation.(4)For the update and maintance of knowledge models of MEC,a self-learning method based on multi-level parallel associations mining is proposed to discover implicit knowledge automatically from a kind of semi-structured technical documents in the field of manufacturing.Among this,an extraction method based on semantic weight for important concept node is proposed according to the structural analysis of the built knowledge model.The dataset is constructed by the extracted data of key concepts in the model for effective mining from the technical documents.Then,a multi-level parallel association mining method is proposed by integrating Map/Reduce and Apriori algorithm to explore knowledge for complementing the knowledge base.For the sake of the performance of the mining approach,the dataset is handled in advance using the hierarchical structure of the built knowledge model.At last,through the implementation and performance comparison of the method,the mining time and results are analysied and explained.(5)Based on the constructed knowledge models,a knowledge base management system for manufacturing equipment capability is developed to supply a web-based,sharable and co-developable knowledge base.The main functions in the system include graph visualization and maitanence of the knowledge models,and knowledge retrival and matching of manufacturing equipment capability.Then,the knowledge base is applied on cloud manufacturing equipment platform to provide knowledge service for task matching of manufacturing equipment.