An Ontology-based Validation Approach to Resolve Conflicts in Manufacturing Design Process

Transcription

1 An Ontology-based Validation Approah to Resolve Conflits in Manufaturing Design Proess Kazunari Hashimoto, Motoyuki Takaai, Yohei Yamane, Seiji Suzuki, Masao Watanabe, and Hiroshi Umemoto Researh and Tehnology Group, Fuji Xerox Co., Ltd. 6-1 Minatomirai, Nishi-ku Yokohama-shi Kanagawa Japan {Kazunari.Hashimoto,Motoyuki.Takaai,Yohei.Yamane, Abstrat. In design proess of manufaturing, designers analyze the fator of phenomena and apply solutions to phenomena (suh as resizing, hanging eletrial iruits, and modifiation of mass). There may be often onflits based on hain relations among solutions whih designers annot reognize. We fous on validating these solutions. In this paper, we propose an ontology for representing hain relations in design proess. To validate designers solutions, we propose an algorithm to detet onflits between a given solution and those for other design ases. In our ase study, we introdue our approah to produt development and demonstrate the effiay and pratiability of our proposed ontology. Keywords: Validating, Design Proess, Design Knowledge, Fault Tree Analysis, Failure Mode and Effet Analysis, SPARQL 1 Introdution The phases of the manufaturing design proess are as follows: Phase 1 deteting phenomena, Phase 2 fator analysis of phenomena, Phase 3 appliation of solutions based on those fators, and Phase 4 implementation of solutions. In this paper, phenomena means troubles whih ours in produt design, and solution means designers s approah improving their troubles. Designers need to share design knowledge to enhane the effiieny of the design proess. To prevent problems arising in design proess, Failure Mode and Effets Analysis (FMEA) and Fault Tree Analysis (FTA) are often employed. There are previous studies based on FMEA and/or FTA for managing the design knowledge with semanti web tehnology whih employ several different methods; i.e., apturing and modeling funtional knowledge [1], modeling the proess of FMEA [2], and onstruting an ontology-based FTA system for applying industrial proess [3] [4] [5]. Most of these studies are foused on relations between phenomena, and generally do not examine the relationships between phenomena and solutions, or relations between solutions.

2 2 Kazunari Hashimoto et al. Designers works based on the design proess for a produt (hereinafter, referred to as design ases), an be used to determine the onflit between solutions of design ases through ommuniations suh as douments and meetings. Conflits mean potential ontraditions between solutions suh as resizing, hanging eletrial iruits, and modifiation of mass and, as mentioned above, designers often annot reognize onflits due to hain relations between solutions and omponents for a partiular produt. This an mean that the effet of solution is halved, or a solution produes the opposite of its intended effet. Therefore, it is ruial that designers onfirm that the applied solution for a given design ase does not onflit with the solutions of other design ases. In this paper, we propose an ontology for representing hain relations in design proess. We then propose an algorithm for deteting onflits between possible solutions by inferring the trae of instane in stage. Finally, we introdue an approah to produt development and demonstrate the effiay and pratiability of our proposed ontology. 2 Proposed Approah Fig. 1. The simplified version of proposed ontology 2.1 The Arhiteture of Ontology There are multiple dependenies on the state of omponents among phenomena, solution, and omponents. The state of omponents indiates the strutural relation between omponents and the funtional relationship between omponents suh as power transmission or transmission of movement. Phenomena and solution are represented by these omponent states. Fator analysis and appliation of solution are both represented by the state of omponents.

3 An Ontology-based Validation Approah to Resolve Conflits 3 We propose an ontology that represents these multiple dependenies in the design proess. Fig.1 illustrates the simplified version of the proposed ontology. Independent entity lass represents an independent onept with phenomena lass, solution lass, and omponent lass as sub lasses. Dependent entity lass represents the dependene between Independent entity lasses with design proess lass, state of omponents lass, fat analysis lass, appliation of solution lass, and influenes between solutions lass as sub lasses. State of omponents lass represents the dependene between omponent lasses. Fat analysis lass, appliation of Solution lass, and influenes between solutions lass represent dependenies among phenomena lass, solution lass and state of omponents lass. Design proess lass is represented by links to phenomena lass, fat analysis lass, and appliation of solution lass. Influenes between solutions lass an represent the hain relation between solutions and validate solutions by the following algorithm. 2.2 Validating Solutions We assume that designers validate the solutions whih they apply to phenomena. Let DC be design ase, and DK be design knowledge. Let Comp be instanes of omponent lass, and Sol be instanes of solution lass. The instanes of Comp DC and Sol DC represent omponents and a solution in the design ase. and Sol DK represent omponents and a solution in and Sol DC. Via instanes of state is linked to Sol DC and Comp DK is linked to Sol DK. The instanes of Comp DK design knowledge, orresponding to Comp DC of omponents, Comp DC The appliation of solution means linking Sol DC to Comp DK. to Sol DK and linking Comp DC Designers must validate that the solution for the urrent ase does not onflit with the solutions for existing ases. Therefore, we propose the following algorithm for deteting onflits between solutions: For ase A and ase B, an algorithm for deteting the onflit between and SolDP is as follows: Sol DP A B Step1 Inferring the solution trae SolT rae DK A B from SolDK A to SolDK B. is omposed of instanes of solution lass and influenes be- SolT rae DK A B tween solutions lass. It is suffiient that at least one instane of influenes between solutions lass represents a ontradition relation. Therefore a onflit exists between SolA DK and SolDK B. Step2 Inferring the omponent trae CompT rae DK A B from CompDK A to CompDK B. CompT rae DK A B is omposed of instanes of omponent lass and state of omponents lass, linked to SolT rae DK A B. Step3 Inferring the omponent trae CompT rae DC A B from CompDC A to CompDC B. It is suffiient that eah instane of CompT rae DK A B of CompT rae DC A B for omponent lass. is linked to an instane The proposed algorithm enables us to detet onflits by inferring the ontradition trae from hain relations, whih the proposed ontology represents in stages. The inferene proess in eah step is implemented using Jena API [6] and SPQRQL query.

4 4 Kazunari Hashimoto et al. 3 Case Study The proposed method was introdued to one design ase of our developing produt. In Phase 1, the designer found phenomena in whih the omponent Cable 50 extends out of the omponent Sensor 20. In Phase 2, the designer analyzed fators whih indiated that the apaity of Sensor 20 was too small. In Phase 3, the designer applied the solution of dereasing the thikness of Sensor 20, thereby inreasing its apaity. The designer then validated the onflit between the solution of the urrent ase and the solutions of existing ases. Fig.2 illustrates a use-ase for deteting the onflit between solutions using our proposed method. Results showed that our proposed method ould aurately detet the onflit between (S10) dereasing the thikness of Sensor 20 and (S20) inreasing the mass of Frame 10. Fig. 2. A use-ase of deteting the onflit between solutions using proposed approah In Step1, our method inferred the solution trae of design knowledge as follows: (S11) dereasing thikness of omponent (IS13) (S31) (IS32) (S21)Inreasing Mass of Component. It was suffiient that this solution trae inluded the ontradition relation (IS32)Reverse Effet. In Step2, orresponding to the above solution trae, our method inferred the omponent trae of design knowledge as follows: (C11) (SC13) (C31) (SC32) (C21). In Step3, orresponding to the above omponent trae, our method inferred the omponent trae of the design ase as follows: (C10)... (C20). Finally, the link (C21) (C20) was inferred using (C31) (C10), (C31) (C32)

5 An Ontology-based Validation Approah to Resolve Conflits 5 (C21). When the omponent trae from (C10) to (C20) existed, the onflit between (S10) and (S20) was indiated. As mentioned above, this ase study learly shows that the proposed method an detet onflits between solutions and represent the potential hain relation whih designers ould not identify. This hain relation was defined as follows: (S10) (S11) (IS13) (S31) (IS32) (S20). As a result, our method ould detet potential onflits between the solutions of 10 design ases. 4 Conlusion In this paper, we foused on validating solutions in design ases for produt manufaturing. We proposed an ontology for representing hain relations among solutions and omponents, and proposed an algorithm for deteting onflits between their solutions. Our method realizes deteting the potential ontradition among designers solution in the initial design phase, so that they an redue design troubles in the post-proess, and the entire design man-hour is redued. In our ase study, the proposed method was applied to our developing produt and ould detet potentials onflit between the solutions of 10 design ases. Designers must ontinuously enhane their design knowledge, and we believe our our proposed method is a useful tool for the enhanement of that knowledge. When onflits whih designers an grasp are not deteted by the proposed algorithm, the area between solutions inludes either erroneous instanes or missing instanes. Designers an then identify the areas whih require enhanement in terms of design knowledge. It is hoped that the proposed method presented in this paper will ontribute to improved solution validation in manufaturing design. Referenes 1. Yoshinobu, Kitamura.: Roles of Ontologies of Engineering Artifats for Design Knowledge Modeling, International Seminar and Workshop Engineering Design in Integrated Produt Development, 59 69, (2006) 2. Zobia, Rehman., Claudiu, V, Kifor.: An Ontology to Support Semanti Management of FMEA Knowledge, International Journal of Computers Communiations & Control, 11(4): , (2016) 3. Ioannis, M, Dokas., Cork, Ireland.: Ontology to Support Knowledge Representation and Risk Analysis for the Development of Early Warning System in Solid Waste Management Operations, International Symposium on Environment Software Systems, (2007) 4. Martin, Melik-Merkumians., Alois, Zoitl.: Ontology-based Fault Diagnosis for Industrial Control Appliations, IEEE onferene on ETFA, (2010) 5. Allan, Veneslau., Raphaela, Lima., Luiz, Affonso, Guedes., Ivanovith, Silva.: Ontology for Computer-aided Fault Tree Synthesis, IEEE Conferene on ETFA (2014) 6. Apahe Jena.