Manufacturing Knowledge Graph: A Connectivism to Answer Production Problems Query With Knowledge Reuse

Abstract

Manufacturing knowledge (MK) is enjoying a ''new golden age'' in the academic domain, marked by vast reuse to support product-related production problems (PPs) solving decision making for manufacturing enterprises in the industry sector. However, the practice of MK reuse and research is fragmented and insufficient, which cannot be mature to provide a systemic solution for that a decision-maker has to consider the involving issues: how MK can be used earlier and rightly; what kind of practical problems can be solved? In order to answer those interconnecting issues, this paper firstly proposes a connectivism framework to clarify the compressive relationship of problem-to-problem, knowledge-to-knowledge and problem-to-knowledge with knowledge integration, knowledge matching, and problem-solving layers. Then, based on the framework, an ontology-based MK graph (MKG) is constructed with a unified MK-filter to collect and integrate multifactor and multilevel MK, and a graph-oriented meta-knowledge model (MKM) is proposed to represent the details between the knowledge entities (i.e., concept and instance), which also shows the contribution to knowledge reasoning. After that, driven by a structure temporal query (i.e., 5W2H), a semantics-based knowledge computation is developed to compute the intrinsic term similarity (IS) and relational term similarity (RS) between two knowledge entities in the MKG. Finally, a case study is taken to demonstrate the effectiveness and performance of the proposed methods. INDEX TERMS Production problems (PPs), manufacturing knowledge (MK), manufacturing knowledge graph (MKG). The associate editor coordinating the review of this manuscript and approving it for publication was Shih-Wei Lin. reuse percentage averages a 28% for manufacturing application [3].The Defense Acquisition University(2011) proposes that the acquisition, operation and support of product could account for 60-80% of the product life cycle cost [4]. These claims highlight that the marginal value of more MK reused by enterprises is the decrease in production cost of product lifecycle, and the increase in effective production time, efficiency and profit permitted by a smaller amount of production problem-solving time. However, from the angle of systematic view of PPs solving with MK, a lack of connective vision among problem solvers has limited the availability of industry-mature solutions. So, the question still needs to be answered: how to use MK, timely, in the product life cycle, to provide a more mature and accurate solution of decision-making for problems solving? While, previous researches have tried to identify the chance in exploring the knowledge reuse for

Social Production System: A Three-Layer Smart Framework for Implementing Autonomous Human-Machine Collaborations in a Shop Floor

“…The current research based on KG mainly includes knowledge representation, acquisition, question answering, and knowledge recommender, etc. [27]. In the field of industrial applications, General Electric (GE) proposed a semanticbased multi-source heterogeneous big data storage and access framework (SemTK) and verified the ease of use and benefits of constructing and querying poly store knowledge graphs with SemTK via four industrial application cases [28].…”

Towards Semantic Integration of Machine Vision Systems to Aid Manufacturing Event Understanding

“…Technologies empowering online semantic representation of system states, such as Semantic Sensor Network (SSN) by World Wide Web Consortium (W3C) [ 43 ], facilitate the introduction of expert experiences from analysis models, simulation platforms, and more recently, Semantic CPPS [ 44 ]. For example, establishing knowledge graphs (KGs) capturing declarative manufacturing knowledge (also referred to as manufacturing ontology) [ 45 , 46 , 47 , 48 ] and multimodal data in a machine-readable format for representation, storage, and further reuse [ 49 , 50 ]. Further integration can be expected to adapt communications, such as semantic gateway as service (SGS) [ 51 ], in the manufacturing applications.…”

A Heuristic Grafting Strategy for Manufacturing Knowledge Graph Extending and Completion Based on Nature Language Processing: KnowTree

“…As an extending of the previous work to answer product related production problems supported by the "Innovation method fund of China" project [3], an original MKG is selected to be extended with 2483 entities connected through 5054 fact triples in 56 named relations, which partly focuses on the white goods (WG) sector. Meanwhile, WG sector, in China, is one of the leading pioneers of introducing industrial engineering and abundant engineering cases have been studied and published.…”

“…Under industry 4.0, contemporary manufacturing enterprises both big and small require new generation information technologies to enable efficient production operation by reducing time and cost of building and extending knowledgebased-systems functionality and capability [1], and being able to make better decisions. Knowledge graph (KG) as a form of structure knowledge system has drawn great research attention from both academia and industry [2,3]. When applied in manufacturing, manufacturing knowledge graph (MKG) as a domain system makes it possible for intelligent search, question-answering and decision-making driven by production problem query and knowledge reuse based on the engineering case study.…”