- Email: [email protected]
Design of networked manufacturing systems for Industry 4.0
Available online at www.sciencedirect.com Available online at www.sciencedirect.com
ScienceDirect ScienceDirect
Available online atonline www.sciencedirect.com Available at www.sciencedirect.com Procedia CIRP 00 (2019) 000–000 Procedia CIRP 00 (2019) 000–000
ScienceDirect ScienceDirect
www.elsevier.com/locate/procedia www.elsevier.com/locate/procedia
ProcediaProcedia CIRP 00CIRP (2017) 81000–000 (2019) 1016–1021 www.elsevier.com/locate/procedia
52nd 52nd CIRP CIRP Conference Conference on on Manufacturing Manufacturing Systems Systems
Design networked manufacturing systems for Industry 4.0 CIRP Design Conference, May 2018, Nantes,for France Design of of 28th networked manufacturing systems Industry 4.0 a
b
c
b
Jelena Milisavljevic-Syed Milisavljevic-Syeda,, Janet K. Allen c, Farrokh Mistreeb Jelena Janetthe K. functional Allenb,, Sesh Sesh Commuri Commuri , Farrokharchitecture Mistree A new methodology to analyze and physical of Division of Industrial Design, School of Engineering, The University of Liverpool, Liverpool, L69 3GH, UK* Division of Industrial Design, School of Engineering, The Universityproduct of Liverpool, Liverpool, L69 3GH, UK* existing products for an assembly oriented family identification The Systems Realization Laboratory, The University of Oklahoma, Norman, Oklahoma 73019, USA The Systems Realization Laboratory, The University of Oklahoma, Norman, Oklahoma 73019, USA a a
b b c Department c
of Electrical and Biomedical Engineering, The University of Nevada, Reno, Nevada, 89557, USA Department of Electrical and Biomedical Engineering, The University of Nevada, Reno, Nevada, 89557, USA * Corresponding author. Tel.: +44 (0)151 795 7362; E-mail address: [email protected] * Corresponding author. Tel.: +44 (0)151 795 7362; E-mail address: [email protected]
Paul Stief *, Jean-Yves Dantan, Alain Etienne, Ali Siadat
École Nationale Supérieure d’Arts et Métiers, Arts et Métiers ParisTech, LCFC EA 4495, 4 Rue Augustin Fresnel, Metz 57078, France
Abstract
*Abstract Corresponding author. Tel.: +33 3 87 37 54 30; E-mail address: [email protected]
Advances in internet technologies have spurred global growth in manufacturing on an unprecedented scale. Distributed global manufacturing Advances in internet technologies have spurred global growth in manufacturing on an unprecedented scale. Distributed global manufacturing requires these industries to support flexible, large scale automation to meet dynamically changing user requirements and to allow interconnection requires these industries to support flexible, large scale automation to meet dynamically changing user requirements and to allow interconnection for data exchange between various components of the manufacturing enterprise. The proceeding needs are addressed with design of Networked Abstract for data exchange between various components of the manufacturing enterprise. The proceeding needs are addressed with design of Networked Manufacturing Systems (NMS) anchored in Industry 4.0. In this paper, we address the requirements for and the salient features of the design of Manufacturing Systems (NMS) anchored in Industry 4.0. In this paper, we address the requirements for and the salient features of the design of illustrated using an automotive paneltowards stamping process. InNMS today’s business environment, the trend product variety and customization is unbroken. Due to this development, the need of NMS illustrated using an automotive panel stampingmore process. © 2019 The Authors. Published by Elsevier Ltd. This is open access article under the BY-NC-ND license agile and reconfigurable production systems emerged to an cope with various products andCC product families. To design and optimize production © 2019 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license © 2019 The Authors. Published by Elsevier Ltd. (http://creativecommons.org/licenses/by-nc-nd/3.0/) systems as well as to choose the optimal product matches, product analysis methods are needed. Indeed, most of the known methods aim to (http://creativecommons.org/licenses/by-nc-nd/3.0/) This is an open access article under CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/) Peer-review under responsibility of the the on scientific committee of the 52ndproduct CIRP Conference on Manufacturing Systems. analyze a product or one product family the physical level. Different families, however, may differ largely in terms of the number and CIRP Conference Conference on on Manufacturing Manufacturing Systems. Systems. Peer-review under responsibility of the scientific committee of the 52nd CIRP nature of components. This fact impedes an efficient comparison and choice of appropriate product family combinations for the production Keywords: Networked manufacturing systems; adaptability; operability analysis; cyber-physical design and manufacturing (CBDM) system. A new methodology is proposed to adaptability; analyze existing products in view of their functional physical architecture. The aim is to cluster Keywords: Networked manufacturing systems; operability analysis; cyber-physical design and and manufacturing (CBDM) these products in new assembly oriented product families for the optimization of existing assembly lines and the creation of future reconfigurable assembly systems. Based on Datum Flow Chain, the physical structure of the products is analyzed. Functional subassemblies are identified, and a functional analysis is performed. Moreover, a hybrid functional and physical architecture graph (HyFPAG) is the output which depicts the 1. Framebetween of reference devices andsystem smart (digitized) manufacturing are the trends. similarity product families by providing design support to both,devices production and product designers.are An the illustrative 1. Frame of reference and smartplanners (digitized) manufacturing trends. Global competition requires enterprises to be able to respond to example of a nail-clipper is used to explain the proposed methodology. An Global industrial case study on two product families competition requires enterprises to of besteering able to columns respondof to Networked manufacturing systems (NMS) are complex these changing requirements in a cost-effective and timely thyssenkrupp Presta France is then carried out (NMS) to give a first the proposed approach. in a cost-effective and timely Networked manufacturing systems are industrial complexevaluation these of changing requirements facilitating one or more operations to assemble or manner while maintaining product quality [3]. ©facilitating 2017 The Authors. by Elsevierrequired B.V. one orPublished more operations required to assemble or manner while maintaining product quality [3]. manufacture a product. Typical examples of networked What are the requirements? To respond to these changes in Peer-review under responsibility of the scientific committee of the 28th CIRP Design Conference 2018.
manufacture a product. Typical examples of networked What are the requirements? To respond to these changes in manufacturing systems (NMS) are automotive machining, the market, in the context of Industry 4.0, it is required for an manufacturing systems (NMS) are automotive machining, the market, in the context of Industry 4.0, it is required for an Keywords: Assembly; Design method; Familylithography identification processes, etc. packaging processes, semiconductor NMS to facilitate flexible production in the system. packaging processes, semiconductor lithography processes, etc. NMS to facilitate flexible production in the system. NMS have characteristics of both mechanical systems and What is the challenge? The challenge is to design an NMS NMS have characteristics of both mechanical systems and What is the challenge? The challenge is to design an NMS control systems. One of the problems to be addressed is that can accommodate dynamic changes, and respond to control systems. One of the problems to be addressed is that can accommodate dynamic changes, and respond to dimensional variations in the finished product due to imprecise unexpected disturbances while managing complexity and variations in the finished product due to imprecise of unexpected while managing complexity and 1.dimensional Introduction the productdisturbances range and characteristics manufactured and/or fixturing of parts, deterioration of tools, or sensors over time. uncertainty. fixturing of parts, deterioration of tools, or sensors over time. assembled uncertainty. in this system. In this context, the main challenge in Dimensional variations can occur in one or more stages of Why it is important to address this challenge? Integrating Dimensional variations can occur in one stages of of modelling Why it and is important to now address Integrating Due to the fast development in or themore domain analysis is not this onlychallenge? to cope with single manufacturing. These errors accumulate and propagate through flexibility in selection/determination of design parameters manufacturing. These errors accumulate through flexibilitya limited in selection/determination of design communication and an ongoing trend and of propagate digitization and products, product range or existing productparameters families, the process and affect the final product quality. The degradation affects not only the dimensional quality of the product but also the process and affect the finalenterprises product quality. The degradation affects only of the producttobut also digitalization, manufacturing are facing important but also not to be ablethe to dimensional analyze and quality to compare products define of quality impacts process productivity, efficiency, and result in the cost of the process. Resolving this issue provides a design of quality impacts process productivity, efficiency, result in new the cost of the process. Resolving this issue design challenges in today’s market environments: a and continuing product families. It can be observed that provides classical aexisting increased process cost [1 - 3]. engineer with better design alternatives and the opportunity to increasedtowards processreduction cost [1 - 3]. engineerfamilies with better design alternatives the opportunity to tendency of product development times and product are regrouped in functionand of clients or features. Why are networked manufacturing systems needed? Modern select an NMS configuration that is robust to variation. Why are networked manufacturing systems Modern However, select anassembly NMS configuration thatfamilies is robust to variation. shortened product lifecycles. In addition, there needed? is an increasing oriented product are hardly to find. manufacturing is characterized by global, distributed assembly Integrating flexibility in an NMS is necessary to accommodate manufacturing is characterized distributed Integrating flexibility in anlevel, NMSproducts is necessary accommodate demand of customization, beingbyatglobal, the same time in assembly a global On the product family differtomainly in two and manufacturing systems that are greatly impacted by wide unexpected modifications in the manufacturing process. The and manufacturing systems that greatly impacted wide main unexpected modifications the manufacturing competition with competitors all are over the world. Thisbytrend, characteristics: (i) the in number of componentsprocess. and (ii) The the variations in customer needs/preferences. Further, we are at the mechanical and control systems are interrelated and cannot be variations in customer Further, at the type mechanical and control are interrelated and cannot be which is inducing theneeds/preferences. development from macrowetoaremicro of components (e.g. systems mechanical, electrical, electronical). threshold of the 4th industrial revolution where mass selected in isolation. Simultaneously addressing both thresholdresults of the 4th industrial revolution mass selected isolation. Simultaneously both markets, in diminished lot sizes due to where augmenting Classicalinmethodologies considering mainlyaddressing single products customization, globalization, interconnectivity between mechanical and control systems at design-time can support customization, globalization,to low-volume interconnectivity between mechanical control systems at design-time can support product varieties (high-volume production) [1]. or solitary, and already existing product families analyze the To cope with this augmenting variety as well as to be able to product structure on a physical level (components level) which © 2019 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc© 2019 Thepossible Authors. Published by Elsevier Ltd. This is an article under the CC BY-NC-ND (http://creativecommons.org/licenses/by-ncidentify optimization potentials in open theaccess existing causes difficultieslicense regarding an efficient definition and nd/3.0/) nd/3.0/) production system, it is important to have a precise knowledge comparison of different product families. Addressing this Peer-review under responsibility of the scientific committee of the 52nd CIRP Conference on Manufacturing Systems. Peer-review under responsibility of the scientific committee of the 52nd CIRP Conference on Manufacturing Systems.
2212-8271 © 2019 The Authors. Published by Elsevier Ltd. This is an©open article Published under theby CC BY-NC-ND 2212-8271 2017access The Authors. Elsevier B.V. license (http://creativecommons.org/licenses/by-nc-nd/3.0/) Peer-review under responsibility of scientific the scientific committee theCIRP 52ndDesign CIRPConference Conference2018. on Manufacturing Systems. Peer-review under responsibility of the committee of the of 28th 10.1016/j.procir.2019.03.244
2
Jelena Milisavljevic-Syed et al. / Procedia CIRP 81 (2019) 1016–1021 Jelena Milisavljevic-Syed et al./ Procedia CIRP 00 (2019) 000–000
flexible production and improve the design of the NMS [3, 4]. Further, such a design can help maintain functionality and performance of an NMS in the face of disturbances or when requirements change [3]. Lastly, the ability to switch configurations and adapt to changes / disturbances is important to maintain and (re)establish connectivity between elements of NMS [5]. Making it happen? A critical review of over 250 publications is included in [3]. In Table 1, in the context of adaptability, operability and reconfigurability, we include a summary that differentiates our work from that of others, namely, current approaches for designing manufacturing systems are primarily
1017
for large-scale manufacturing of a product in the most efficient manner. These processes cannot easily handle the design of robust NMSs when the requirements are partially known, product requirements change/vary in a dynamic manner, or when a system failure occurs because of a breakdown of interconnectivity between elements of the NMS. As an answer to these challenges, in this paper, we propose a computational framework, Design for Dynamic Management (DFDM), to support flexible, operable, and rapidly configurable NMSs. This framework is foundational for further digitization of an NMS, Fig. 1. The use of DFDM is illustrated using an automotive panel stamping process.
Table 1. Rationale for development of computational framework for dynamic management.
Operability
Method Material planning (MRP), Kanban control systems, Hybrid strategies [9] Stochastic modeling of manufacturing systems [10]
Reconfigurability
Analysis Gap Question Answer Material planning and push, pull, and While these methods are What is the method that hybrid control strategies well known in the literature facilitates adaptable and Stochastic modeling of manufacturing and have found extensive concurrent design in the realization of networked systems where the production function and use in the analysis of manufacturing systems, they manufacturing systems in the inventory control Effect of variability in process times and the are not suited for use in the context of Industry 4.0? size of buffers between stages on the overall design of NMS to achieve a performance in terms of production rate and specified quality of the average in-process inventory manufactured product. Inherent operability Set-point controlled - outputs to be While these methods are What is the method that of a process without controlled at a desired value well known in the literature facilitates dynamic change Set-interval controlled - outputs to be and have found extensive over time/change in the specific control use in the plant design, they realization of networked controlled within a desired range structure Safe operating Operable regions and tolerable disturbances are not suited for use in the manufacturing systems in the with the available inputs design of NMS. Applicable context of Industry 4.0? regions and tolerable disturbances with the Minimum-time formulation evaluates the if design engineer has domain knowledge. operability of the process without available inputs Minimum-time considering a preassigned feedback control problem controller formulation [11] Time-optimal control problem formulated as a nonlinear programming problem Reconfigurable The core characteristics and design principles Verified with simple What is the method that manufacturing system of RMS reacting system. For facilitates rapidly changing (RMS) [12] The structure recommended for practical RMS complex systems such market requirements in the Reconfigurable Module selection and the assembly applications would require realization of networked the solution of larger manufacturing systems and manufacturing tool relationships among modules in RMT optimization problems and remains it competitive in the (RMT) [6] multivariable controllers. context of Industry 4.0? Computational Framework, Design for Dynamic Management (DFDM)
Aspect Adaptability
Fig.1. Mental model for DFDM.
Jelena Milisavljevic-Syed et al./ Procedia CIRP 00 (2019) 000–000 1018
3
Jelena Milisavljevic-Syed et al. / Procedia CIRP 81 (2019) 1016–1021
Possible way forward? DFDM Framework that is proposed is flexible and can be adapted for the design of dynamic systems other than NMSs. Some possible theoretical analyses that can be addressed using DFDM are identified in Quadrant 1 in Fig. 1, and possible applications are identified in Quadrant 3 in Fig. 1. In the future, DFDM can be extended to Cloud-based Design and Manufacturing, Quadrant 2 in Fig. 1, and to applications in smart networked systems, Quadrant 4 in Fig. 1. Some possible extension of DFDM in various industries are the following: Cyber-physical design. Model-based and Model-Free intelligent decision-based design systems. We speculate that there are applications in steel making processes. . Cyber-physical manufacturing. Design of reconfigurable intelligent manufacturing systems. We speculate that there are applications in additive manufacturing, and semiconductor lithography processes, etc. Cyber-physical product and material. Integration of materials, products, and manufacturing processes. We speculate that there are applications in the design of smart sports equipment, etc. Cyber-physical-social system design. Model cyber-social design decision network in the realization of intelligent cyber-physical-social systems. We speculate that there are applications in the smart healthcare system, etc. The remainder of this paper is organized as follows. In Section 2, we present a brief description of DFDM. In Section 3, we present a panel stamping process as an illustrative example. In Section 4, we present the mathematical model of the cDSP formulation. In Section 5, we discuss the flexible and functional design of a panel stamping process in cyber-physical design and manufacturing. The closing remarks are presented in Section 6 2. Computational framework for design of networked manufacturing systems With the computational framework of Design for Dynamic Management (DFDM) some of the thorny problems in the design and analysis of NMS can be addressed. With this framework the needs of Industry 4.0 can be addressed in the future. The computational framework of DFDM is presented in Fig. 2 and the steps in implementing this framework are presented in Sections 2.1 – 2.3. 2.1. Step 1 - Adaptable Concurrent Design Is the system flexible? As a response, we propose Adaptive Concurrent Realization of Engineering Systems (ACRONES) to fit n-stage NMSs with a set of constraints and goals guided by rapidly changing requirements. ACRONES is based on the Stream of Variation (SoV) method to model the NMS and the compromise Decision Support Problem (cDSP). The cDSP is used to manage the structure and information of decisionmaking within adaptability models with and without
uncertainty. For more information about the ACRONES see [3, 4]. 2.2. Step 2 - Operability Analysis What is the functionality and dynamic performance of the system? Based on the output information we obtain from Step 1, Section 2.1, we frame operability and disturbance spaces. Hence, we propose a Steady-State Operability Model (SSOM) to analyze system functionality, and Dynamic Operability Model (DOM) to analyze the dynamic performance of the system. These models are based on Operability Analysis (OA), cDSP, and Minimum Time Control (MTC) to determine a response of a system undergoing dynamic changes. For more information see [3].
Fig. 2. Computational framework design for dynamic management.
2.3. Step 3 - Reconfiguration Strategy What is the reconfiguration strategy? Based on the information from Step 2, Section 2.2, we can determine whether the elements in the system are connected and if not how the system should be reconfigured. We hypothesize that different configuration strategies support the repeated systemic reconfiguration among elements in the system, particularly for the Reconfigurable Machine Tools (RMTs), the Reconfigurable Inspection Systems (RISs), and Reconfigurable Manufacturing Systems (RMSs). Reconfiguration strategies are based on Decision Trees (DTs) which are used to transform RMT/RIS configurations from various module libraries into one comprehensive representation, with the cDSP and Game Theory (GT) to explore interactions between RMT and RIS. For more information see [5, 6]. In the proposed computational framework, both consumer and producer preferences are integrated to design a cyberphysical-product-service system of low-cost and high-quality which is adaptable to dynamic market demands. The current status of our work on DFDM is on its uses in cyber-physical design and manufacturing while in future we anticipate accommodating cyber-physical products and materials, and cyber-physical-social systems, Quadrant 2 in Fig. 1.
4
Jelena Milisavljevic-Syed et al. / Procedia CIRP 81 (2019) 1016–1021 Jelena Milisavljevic-Syed et al./ Procedia CIRP 00 (2019) 000–000
1019
3. Illustrative example
4. The mathematical model
Design of an NMS has several possible applications, see Quadrant 3 in Fig. 1, (e.g., assembling processes, design of a chemical plant, design of a transmission box, etc.). However, in this paper we illustrate design of an NMS with an automotive panel stamping process. The challenges in the design of NMS are illustrated through a two-dimensional panel stamping process in which four parts are assembled across three stations, Fig. 3. Each part is restrained by a set of fixtures (Pi), and each position of a part is measured by a coordinate sensor (Mi). At Station 1 Parts 1 and 2 are assembled. In Station 2 Parts 1 and 2 are made into a Subassembly 1 that are assembled with Parts 3 and 4. At Station 3 all parts are assembled in final assembly. Each part at each station has up to 2 sensors that are measurement points. The total number of measuring points is Mi (i=1 – 20). Fixture locators can be active and inactive. There are two types of fixture locators, namely, a 4-way pin, and a 2-way pin.
In this paper, we address a two-part problem: (1) provide flexibility in the selection and determination of values of design parameters (tool and sensor attributes) that give satisfying (robust) solutions regarding the process cost and quality of product; and (2) provide for the determination of input ranges that give functional system design and satisfying dynamic performance of a system in the presence of changes in requirements. The problem characteristics are the number of work stations, N, the number of parts, n_p, the number of sensors, M_Pi, the number of tools, P_i, and the potential fixture failures in the process, m_r. The requirements for the process are diagnosability, controllability and operability and these must be full. Expected variations, y_k, must lie between -0.8 and 0.8 millimeters. The assumptions are that the expected variations parameters follow Gaussian distribution with zero mean and known covariance. The overall objective is to minimize the process cost and maximize product quality. The first part of the problem is simulated with the adaptability model, Section 4.1, and the second part of the problem is simulated with the operability model, Section 4.2. 4.1. The adaptability model
Fig.3. Two-dimensional panel stamping process [7, 8].
The adaptability model is based on the SoV model that we use to model an NMS and identify mechanical and control system drivers and is further partitioned inot smaller process decision and performance measurement cDSP models, the upper part of Fig. 4. The cDSPs models are used to build in flexibility in the selection/determination of design parameters and manage the structure and information of decision-making. For more information about the adaptability model see [3, 4].
Fig.4. Adaptability and operability model. .
1020
Jelena Milisavljevic-Syed et al. / Procedia CIRP 81 (2019) 1016–1021 Jelena Milisavljevic-Syed et al./ Procedia CIRP 00 (2019) 000–000
5.1. Flexible design of panel stamping process
Table 2. The compromise decision support problem (cDSP) formulation. Given (Parameters) Total number of operational stations Number of stamping parts in the process Number and position of fixture points Potential number of sensors and position Dimensional quality (size of variations) boundary values are set Find (System Variables) Total number of sensors and sensing stations Use of PT`s control actions Sensing penalties Sensors distribution schemes that are diagnosable, controllable, and costeffective Satisfy (Constraints) Tooling constraints Sensing constraints
5
N np
Pi [-]; (x, z) [mm]
MPi [-]; (x, z) [mm]
Yk [mm]
MPi,M ; MS,M (M=D,C,E)
PT P
Mi,k
Use of programmable tooling, Eq. 3.6 in [3] Number, position and distribution of sensors, Eq. 3.6 in [3] 100% or partial, see Eq. 3.3 in [3] 100% or partial, see Eq. 3.5 in [3]
Process diagnosability Process controllability Satisfy (Bounds) MPi,M (M=D,C,E) , Section 4.2 in [3] Lower and upper number of sensors Lower and upper number of sensing MS,M (M=D,C,E) , Section 4.2 in [3] stations Use of programmable tooling PT, Section 4.2 in [3] Lower and upper limit of sensing P, Section 4.2 in [3] penalties − + Deviation variables di , di ≥ 0, Section 4.2 in [3] Goal 1: Maximize process Chapter 4, Section 4.2 in [3] adaptability Goal 2: Maximize process operability Chapter 5, Sections 5.1 and 5.2 in [3] Minimize (Deviations) Minimize deviation function Z = min ∑3i=1(Wi d−i + Wi d+i ); ∑3i=1 Wi = 1; Wi ≥ 0; d−i , d+i ≥ 0; d+i ∙ d−i = 0 (i = 1 − 3), Section 3.2.1 in [1]
4.2. The operability model
The operability model is used to analyze the functionality and dynamic performance of the system as the system requirements change. Operability analysis (OA) is accomplished in two stages: steady-state operability analysis and dynamic operability analysis. In this paper, we focus on steady-state operability analysis. Steady-state operability is used to analyze how different requirements, driven by customer preferences, change system functionality. System functionality is analyzed using a Steady-State Operability Model (SSOM) that is based on OA and the cDSP construct, as shown in the lower part of Fig. 4. For further details of the steady-state operability model see [3]. In Section 5, we present the flexible and functional design of the panel stamping process. 5. Flexible and functional design of the panel stamping process The specification of system variables and their effects on the overall cost and quality of the NMS are difficult to ascertain before implementation. For the sake of illustration, five sensors are assumed to be available for use in the design of the panel stamping process. The results are obtained using a MATLAB simulation.
In this section, we answer the question “How does the selection of design parameters (sensors and sensing stations) influence the cost of the process and dimensional quality of the product?” In Fig. 5 a), it can be seen that when we are using 2 sensing stations the cost of the process is lower than when we are using 3 sensing stations. Further, in Fig. 5 b) and c), the expected size of variations is lower when we are using 2 sensing stations than 3 sensing stations. With the use of 2 sensing stations in the process, we design a low-cost and high-quality system in which all requirements are satisfied. Our recommendation is that “less is more” with a reduced number of sensing stations we lower the cost of the process and obtain a high quality of the product. Further, the dimensional quality is not improved by an increase in the number of sensing stations but rather with a proper distribution of adequate numbers of sensors.
Fig.5. Cost of the process and dimensional quality of the product.
5.2. Functional design of panel stamping process In this section, we answer the question “Is the system functional in the presence of disturbances?” In order to analyze system functionality we need to obtain information regarding Available Input Space (AIS), and Desired Input Space (DIS). AIS are inputs of the process which are able to change over a certain range, identified as the feasible design space. AIS is obtained by exercising the adaptability model, shown in Section 4.1 and Fig. 6, a). DIS is the set of input values required to reach the entire Desired Output Space (DOS). DIS is framed according to customer preferences, Section 4.2 and Fig. 6, b). The Operability of the system is full (a system is functional) where customer preferences (DIS) intersect with ranges of change (AIS) in the presence of disturbances (natural uncertainty) where system can stabilize, Fig. 6, c). In summary, the solution scheme shown in, Fig. 4, is an elegant and efficient way to explore the solution space and identify both flexible and functional design in cyber-physical design and manufacturing. Exercising the example gives the opportunity of visualizing the dependence of the cost on the number and location of the sensing stations in the process, Fig. 5, a).
6
Jelena Milisavljevic-Syed et al./ Procedia CIRP 00 (2019) 000–000 Jelena Milisavljevic-Syed et al. / Procedia CIRP 81 (2019) 1016–1021
1021
Fig.6. System operability.
However, dimensional quality is improved not by increasing the numbers of sensing stations in the process, Fig. 5 b) and c), but rather by the appropriate selection of the design parameters and sensor distributions. Furthermore, the system is functional where customer preferences intersect with feasible design space even in the presence of disturbances as long system can stabilize, Fig. 6, c), under dynamic change of the requirements. 6. Closure In this paper, we present a computational framework, DFDM, suited for theoretical analysis of Networked Manufacturing Systems and for addressing the design needs of multi-stage manufacturing processes. In the context of Industry 4.0, DFDM can be used to design NMSs that are adaptable to dynamic market demands, robust to network/connectivity failures, and capable of mass customization. In the future, DFDM can also be used for design of cloud-based cyberphysical and manufacturing systems, based on adaptability, operability, and reconfigurability. In this paper, the use of DFDM in flexible and functional design was illustrated with an automotive panel stamping example. Within the International System Realization Partnership (ISRP) we are in process of CBDM platformization with possible application of DFDM in cyber-physical design, cyberphysical manufacturing, cyber-physical product and material, and cyber-physical-social system design. Further, we speculate that DFDM will transit to industry and be applied in the steel making processes, additive manufacturing, design of smart healthcare system, etc. For more information about current projects please visit www.liverpool.ac.uk/engineering/staff/jelenamilisavljevic-syed. Acknowledgements Jelena Milisavljevic-Syed acknowledges financial support from the John and Mary Moore Chair and the L.A. Comp Chair for her graduate studies at the University of Oklahoma, Norman, Oklahoma.
References [1] Milisavljevic J. Architecting networked engineering systems. Master thesis. School of aerospace and mechanical engineering. University of Oklahoma 2015. Norman. Oklahoma. USA. [2] Ding Y, Ceglarek D, and Shi J. Diagnosability analysis ofmultistationmanufacturing processes. Journal of dynamics systems, measurement, and control 2002c; 124:1–13. [3] Milisavljevic J. Architecting networked engineering systems. Doctoral dissertation. School of aerospace and mechanical engineering. University of Oklahoma 2018. Norman. Oklahoma. USA. [4] Milisavljevic-Syed J, Allen JK, Commuri S, Mistree F. A method for the concurrent design and analysis of networked manufacturing systems under uncertainty. Engineering optimization 2018. GENO-2016-0753.R2. [5] Shang X, Milisavljevic-Syed J, Wang G, Allen J K, Mistree F. a key feature-based method for configuration design of reconfigurable inspection system. Journal of intelligent manufacturing 2018. JIMS-D-18-00505; under review. [6] Wang G, Shang X, Yan Y, Allen JK, Mistree F. A tree-based decision method for the configuration design of reconfigurable machine tools. Journal of manufacturing systems 2018;149:143-169. [7] Ding Y, Kim P, Ceglarek D, Jin J. Optimal sensor distribution for variation diagnosis in multistation assembly processes. IEEE transactions on robotics and automation 2003; 19. 4:543-556. [8] Ding Y, Ceglarek D, Shi J. Design evaluation of multi-station assembly processes by using state space approach. Journal of mechanical design 2002a; 124. 3: 408-418. [9] Krishnamurthy A, Suri R, Vernon M. Re-examining the performance of mrp and kanban material control strategies for multi-product flexible manufacturing systems. International journal of flexible manufacturing systems 2004; 16. 2: 123-150. [10] Altiok T. Performance evaluation of manufacturing systems. Springer Verlag; 1997. [11] Subramanian S, Uztürk D, Georgakis C. An optimization-based approach for the operability analysis of continuously stirred tank reactors. Industrial and engineering chemistry research 2001; 40. 20: 4238-4252. [12] Koren Y, Heisel U, Jovane F, Moriwaki T, Pritschow G, Ulsoy G, Van Brussel H. Reconfigurable manufacturing systems", CIRP annals 1999; 48. 2: 527-540.
ScienceDirect ScienceDirect
Available online atonline www.sciencedirect.com Available at www.sciencedirect.com Procedia CIRP 00 (2019) 000–000 Procedia CIRP 00 (2019) 000–000
ScienceDirect ScienceDirect
www.elsevier.com/locate/procedia www.elsevier.com/locate/procedia
ProcediaProcedia CIRP 00CIRP (2017) 81000–000 (2019) 1016–1021 www.elsevier.com/locate/procedia
52nd 52nd CIRP CIRP Conference Conference on on Manufacturing Manufacturing Systems Systems
Design networked manufacturing systems for Industry 4.0 CIRP Design Conference, May 2018, Nantes,for France Design of of 28th networked manufacturing systems Industry 4.0 a
b
c
b
Jelena Milisavljevic-Syed Milisavljevic-Syeda,, Janet K. Allen c, Farrokh Mistreeb Jelena Janetthe K. functional Allenb,, Sesh Sesh Commuri Commuri , Farrokharchitecture Mistree A new methodology to analyze and physical of Division of Industrial Design, School of Engineering, The University of Liverpool, Liverpool, L69 3GH, UK* Division of Industrial Design, School of Engineering, The Universityproduct of Liverpool, Liverpool, L69 3GH, UK* existing products for an assembly oriented family identification The Systems Realization Laboratory, The University of Oklahoma, Norman, Oklahoma 73019, USA The Systems Realization Laboratory, The University of Oklahoma, Norman, Oklahoma 73019, USA a a
b b c Department c
of Electrical and Biomedical Engineering, The University of Nevada, Reno, Nevada, 89557, USA Department of Electrical and Biomedical Engineering, The University of Nevada, Reno, Nevada, 89557, USA * Corresponding author. Tel.: +44 (0)151 795 7362; E-mail address: [email protected] * Corresponding author. Tel.: +44 (0)151 795 7362; E-mail address: [email protected]
Paul Stief *, Jean-Yves Dantan, Alain Etienne, Ali Siadat
École Nationale Supérieure d’Arts et Métiers, Arts et Métiers ParisTech, LCFC EA 4495, 4 Rue Augustin Fresnel, Metz 57078, France
Abstract
*Abstract Corresponding author. Tel.: +33 3 87 37 54 30; E-mail address: [email protected]
Advances in internet technologies have spurred global growth in manufacturing on an unprecedented scale. Distributed global manufacturing Advances in internet technologies have spurred global growth in manufacturing on an unprecedented scale. Distributed global manufacturing requires these industries to support flexible, large scale automation to meet dynamically changing user requirements and to allow interconnection requires these industries to support flexible, large scale automation to meet dynamically changing user requirements and to allow interconnection for data exchange between various components of the manufacturing enterprise. The proceeding needs are addressed with design of Networked Abstract for data exchange between various components of the manufacturing enterprise. The proceeding needs are addressed with design of Networked Manufacturing Systems (NMS) anchored in Industry 4.0. In this paper, we address the requirements for and the salient features of the design of Manufacturing Systems (NMS) anchored in Industry 4.0. In this paper, we address the requirements for and the salient features of the design of illustrated using an automotive paneltowards stamping process. InNMS today’s business environment, the trend product variety and customization is unbroken. Due to this development, the need of NMS illustrated using an automotive panel stampingmore process. © 2019 The Authors. Published by Elsevier Ltd. This is open access article under the BY-NC-ND license agile and reconfigurable production systems emerged to an cope with various products andCC product families. To design and optimize production © 2019 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license © 2019 The Authors. Published by Elsevier Ltd. (http://creativecommons.org/licenses/by-nc-nd/3.0/) systems as well as to choose the optimal product matches, product analysis methods are needed. Indeed, most of the known methods aim to (http://creativecommons.org/licenses/by-nc-nd/3.0/) This is an open access article under CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/) Peer-review under responsibility of the the on scientific committee of the 52ndproduct CIRP Conference on Manufacturing Systems. analyze a product or one product family the physical level. Different families, however, may differ largely in terms of the number and CIRP Conference Conference on on Manufacturing Manufacturing Systems. Systems. Peer-review under responsibility of the scientific committee of the 52nd CIRP nature of components. This fact impedes an efficient comparison and choice of appropriate product family combinations for the production Keywords: Networked manufacturing systems; adaptability; operability analysis; cyber-physical design and manufacturing (CBDM) system. A new methodology is proposed to adaptability; analyze existing products in view of their functional physical architecture. The aim is to cluster Keywords: Networked manufacturing systems; operability analysis; cyber-physical design and and manufacturing (CBDM) these products in new assembly oriented product families for the optimization of existing assembly lines and the creation of future reconfigurable assembly systems. Based on Datum Flow Chain, the physical structure of the products is analyzed. Functional subassemblies are identified, and a functional analysis is performed. Moreover, a hybrid functional and physical architecture graph (HyFPAG) is the output which depicts the 1. Framebetween of reference devices andsystem smart (digitized) manufacturing are the trends. similarity product families by providing design support to both,devices production and product designers.are An the illustrative 1. Frame of reference and smartplanners (digitized) manufacturing trends. Global competition requires enterprises to be able to respond to example of a nail-clipper is used to explain the proposed methodology. An Global industrial case study on two product families competition requires enterprises to of besteering able to columns respondof to Networked manufacturing systems (NMS) are complex these changing requirements in a cost-effective and timely thyssenkrupp Presta France is then carried out (NMS) to give a first the proposed approach. in a cost-effective and timely Networked manufacturing systems are industrial complexevaluation these of changing requirements facilitating one or more operations to assemble or manner while maintaining product quality [3]. ©facilitating 2017 The Authors. by Elsevierrequired B.V. one orPublished more operations required to assemble or manner while maintaining product quality [3]. manufacture a product. Typical examples of networked What are the requirements? To respond to these changes in Peer-review under responsibility of the scientific committee of the 28th CIRP Design Conference 2018.
manufacture a product. Typical examples of networked What are the requirements? To respond to these changes in manufacturing systems (NMS) are automotive machining, the market, in the context of Industry 4.0, it is required for an manufacturing systems (NMS) are automotive machining, the market, in the context of Industry 4.0, it is required for an Keywords: Assembly; Design method; Familylithography identification processes, etc. packaging processes, semiconductor NMS to facilitate flexible production in the system. packaging processes, semiconductor lithography processes, etc. NMS to facilitate flexible production in the system. NMS have characteristics of both mechanical systems and What is the challenge? The challenge is to design an NMS NMS have characteristics of both mechanical systems and What is the challenge? The challenge is to design an NMS control systems. One of the problems to be addressed is that can accommodate dynamic changes, and respond to control systems. One of the problems to be addressed is that can accommodate dynamic changes, and respond to dimensional variations in the finished product due to imprecise unexpected disturbances while managing complexity and variations in the finished product due to imprecise of unexpected while managing complexity and 1.dimensional Introduction the productdisturbances range and characteristics manufactured and/or fixturing of parts, deterioration of tools, or sensors over time. uncertainty. fixturing of parts, deterioration of tools, or sensors over time. assembled uncertainty. in this system. In this context, the main challenge in Dimensional variations can occur in one or more stages of Why it is important to address this challenge? Integrating Dimensional variations can occur in one stages of of modelling Why it and is important to now address Integrating Due to the fast development in or themore domain analysis is not this onlychallenge? to cope with single manufacturing. These errors accumulate and propagate through flexibility in selection/determination of design parameters manufacturing. These errors accumulate through flexibilitya limited in selection/determination of design communication and an ongoing trend and of propagate digitization and products, product range or existing productparameters families, the process and affect the final product quality. The degradation affects not only the dimensional quality of the product but also the process and affect the finalenterprises product quality. The degradation affects only of the producttobut also digitalization, manufacturing are facing important but also not to be ablethe to dimensional analyze and quality to compare products define of quality impacts process productivity, efficiency, and result in the cost of the process. Resolving this issue provides a design of quality impacts process productivity, efficiency, result in new the cost of the process. Resolving this issue design challenges in today’s market environments: a and continuing product families. It can be observed that provides classical aexisting increased process cost [1 - 3]. engineer with better design alternatives and the opportunity to increasedtowards processreduction cost [1 - 3]. engineerfamilies with better design alternatives the opportunity to tendency of product development times and product are regrouped in functionand of clients or features. Why are networked manufacturing systems needed? Modern select an NMS configuration that is robust to variation. Why are networked manufacturing systems Modern However, select anassembly NMS configuration thatfamilies is robust to variation. shortened product lifecycles. In addition, there needed? is an increasing oriented product are hardly to find. manufacturing is characterized by global, distributed assembly Integrating flexibility in an NMS is necessary to accommodate manufacturing is characterized distributed Integrating flexibility in anlevel, NMSproducts is necessary accommodate demand of customization, beingbyatglobal, the same time in assembly a global On the product family differtomainly in two and manufacturing systems that are greatly impacted by wide unexpected modifications in the manufacturing process. The and manufacturing systems that greatly impacted wide main unexpected modifications the manufacturing competition with competitors all are over the world. Thisbytrend, characteristics: (i) the in number of componentsprocess. and (ii) The the variations in customer needs/preferences. Further, we are at the mechanical and control systems are interrelated and cannot be variations in customer Further, at the type mechanical and control are interrelated and cannot be which is inducing theneeds/preferences. development from macrowetoaremicro of components (e.g. systems mechanical, electrical, electronical). threshold of the 4th industrial revolution where mass selected in isolation. Simultaneously addressing both thresholdresults of the 4th industrial revolution mass selected isolation. Simultaneously both markets, in diminished lot sizes due to where augmenting Classicalinmethodologies considering mainlyaddressing single products customization, globalization, interconnectivity between mechanical and control systems at design-time can support customization, globalization,to low-volume interconnectivity between mechanical control systems at design-time can support product varieties (high-volume production) [1]. or solitary, and already existing product families analyze the To cope with this augmenting variety as well as to be able to product structure on a physical level (components level) which © 2019 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc© 2019 Thepossible Authors. Published by Elsevier Ltd. This is an article under the CC BY-NC-ND (http://creativecommons.org/licenses/by-ncidentify optimization potentials in open theaccess existing causes difficultieslicense regarding an efficient definition and nd/3.0/) nd/3.0/) production system, it is important to have a precise knowledge comparison of different product families. Addressing this Peer-review under responsibility of the scientific committee of the 52nd CIRP Conference on Manufacturing Systems. Peer-review under responsibility of the scientific committee of the 52nd CIRP Conference on Manufacturing Systems.
2212-8271 © 2019 The Authors. Published by Elsevier Ltd. This is an©open article Published under theby CC BY-NC-ND 2212-8271 2017access The Authors. Elsevier B.V. license (http://creativecommons.org/licenses/by-nc-nd/3.0/) Peer-review under responsibility of scientific the scientific committee theCIRP 52ndDesign CIRPConference Conference2018. on Manufacturing Systems. Peer-review under responsibility of the committee of the of 28th 10.1016/j.procir.2019.03.244
2
Jelena Milisavljevic-Syed et al. / Procedia CIRP 81 (2019) 1016–1021 Jelena Milisavljevic-Syed et al./ Procedia CIRP 00 (2019) 000–000
flexible production and improve the design of the NMS [3, 4]. Further, such a design can help maintain functionality and performance of an NMS in the face of disturbances or when requirements change [3]. Lastly, the ability to switch configurations and adapt to changes / disturbances is important to maintain and (re)establish connectivity between elements of NMS [5]. Making it happen? A critical review of over 250 publications is included in [3]. In Table 1, in the context of adaptability, operability and reconfigurability, we include a summary that differentiates our work from that of others, namely, current approaches for designing manufacturing systems are primarily
1017
for large-scale manufacturing of a product in the most efficient manner. These processes cannot easily handle the design of robust NMSs when the requirements are partially known, product requirements change/vary in a dynamic manner, or when a system failure occurs because of a breakdown of interconnectivity between elements of the NMS. As an answer to these challenges, in this paper, we propose a computational framework, Design for Dynamic Management (DFDM), to support flexible, operable, and rapidly configurable NMSs. This framework is foundational for further digitization of an NMS, Fig. 1. The use of DFDM is illustrated using an automotive panel stamping process.
Table 1. Rationale for development of computational framework for dynamic management.
Operability
Method Material planning (MRP), Kanban control systems, Hybrid strategies [9] Stochastic modeling of manufacturing systems [10]
Reconfigurability
Analysis Gap Question Answer Material planning and push, pull, and While these methods are What is the method that hybrid control strategies well known in the literature facilitates adaptable and Stochastic modeling of manufacturing and have found extensive concurrent design in the realization of networked systems where the production function and use in the analysis of manufacturing systems, they manufacturing systems in the inventory control Effect of variability in process times and the are not suited for use in the context of Industry 4.0? size of buffers between stages on the overall design of NMS to achieve a performance in terms of production rate and specified quality of the average in-process inventory manufactured product. Inherent operability Set-point controlled - outputs to be While these methods are What is the method that of a process without controlled at a desired value well known in the literature facilitates dynamic change Set-interval controlled - outputs to be and have found extensive over time/change in the specific control use in the plant design, they realization of networked controlled within a desired range structure Safe operating Operable regions and tolerable disturbances are not suited for use in the manufacturing systems in the with the available inputs design of NMS. Applicable context of Industry 4.0? regions and tolerable disturbances with the Minimum-time formulation evaluates the if design engineer has domain knowledge. operability of the process without available inputs Minimum-time considering a preassigned feedback control problem controller formulation [11] Time-optimal control problem formulated as a nonlinear programming problem Reconfigurable The core characteristics and design principles Verified with simple What is the method that manufacturing system of RMS reacting system. For facilitates rapidly changing (RMS) [12] The structure recommended for practical RMS complex systems such market requirements in the Reconfigurable Module selection and the assembly applications would require realization of networked the solution of larger manufacturing systems and manufacturing tool relationships among modules in RMT optimization problems and remains it competitive in the (RMT) [6] multivariable controllers. context of Industry 4.0? Computational Framework, Design for Dynamic Management (DFDM)
Aspect Adaptability
Fig.1. Mental model for DFDM.
Jelena Milisavljevic-Syed et al./ Procedia CIRP 00 (2019) 000–000 1018
3
Jelena Milisavljevic-Syed et al. / Procedia CIRP 81 (2019) 1016–1021
Possible way forward? DFDM Framework that is proposed is flexible and can be adapted for the design of dynamic systems other than NMSs. Some possible theoretical analyses that can be addressed using DFDM are identified in Quadrant 1 in Fig. 1, and possible applications are identified in Quadrant 3 in Fig. 1. In the future, DFDM can be extended to Cloud-based Design and Manufacturing, Quadrant 2 in Fig. 1, and to applications in smart networked systems, Quadrant 4 in Fig. 1. Some possible extension of DFDM in various industries are the following: Cyber-physical design. Model-based and Model-Free intelligent decision-based design systems. We speculate that there are applications in steel making processes. . Cyber-physical manufacturing. Design of reconfigurable intelligent manufacturing systems. We speculate that there are applications in additive manufacturing, and semiconductor lithography processes, etc. Cyber-physical product and material. Integration of materials, products, and manufacturing processes. We speculate that there are applications in the design of smart sports equipment, etc. Cyber-physical-social system design. Model cyber-social design decision network in the realization of intelligent cyber-physical-social systems. We speculate that there are applications in the smart healthcare system, etc. The remainder of this paper is organized as follows. In Section 2, we present a brief description of DFDM. In Section 3, we present a panel stamping process as an illustrative example. In Section 4, we present the mathematical model of the cDSP formulation. In Section 5, we discuss the flexible and functional design of a panel stamping process in cyber-physical design and manufacturing. The closing remarks are presented in Section 6 2. Computational framework for design of networked manufacturing systems With the computational framework of Design for Dynamic Management (DFDM) some of the thorny problems in the design and analysis of NMS can be addressed. With this framework the needs of Industry 4.0 can be addressed in the future. The computational framework of DFDM is presented in Fig. 2 and the steps in implementing this framework are presented in Sections 2.1 – 2.3. 2.1. Step 1 - Adaptable Concurrent Design Is the system flexible? As a response, we propose Adaptive Concurrent Realization of Engineering Systems (ACRONES) to fit n-stage NMSs with a set of constraints and goals guided by rapidly changing requirements. ACRONES is based on the Stream of Variation (SoV) method to model the NMS and the compromise Decision Support Problem (cDSP). The cDSP is used to manage the structure and information of decisionmaking within adaptability models with and without
uncertainty. For more information about the ACRONES see [3, 4]. 2.2. Step 2 - Operability Analysis What is the functionality and dynamic performance of the system? Based on the output information we obtain from Step 1, Section 2.1, we frame operability and disturbance spaces. Hence, we propose a Steady-State Operability Model (SSOM) to analyze system functionality, and Dynamic Operability Model (DOM) to analyze the dynamic performance of the system. These models are based on Operability Analysis (OA), cDSP, and Minimum Time Control (MTC) to determine a response of a system undergoing dynamic changes. For more information see [3].
Fig. 2. Computational framework design for dynamic management.
2.3. Step 3 - Reconfiguration Strategy What is the reconfiguration strategy? Based on the information from Step 2, Section 2.2, we can determine whether the elements in the system are connected and if not how the system should be reconfigured. We hypothesize that different configuration strategies support the repeated systemic reconfiguration among elements in the system, particularly for the Reconfigurable Machine Tools (RMTs), the Reconfigurable Inspection Systems (RISs), and Reconfigurable Manufacturing Systems (RMSs). Reconfiguration strategies are based on Decision Trees (DTs) which are used to transform RMT/RIS configurations from various module libraries into one comprehensive representation, with the cDSP and Game Theory (GT) to explore interactions between RMT and RIS. For more information see [5, 6]. In the proposed computational framework, both consumer and producer preferences are integrated to design a cyberphysical-product-service system of low-cost and high-quality which is adaptable to dynamic market demands. The current status of our work on DFDM is on its uses in cyber-physical design and manufacturing while in future we anticipate accommodating cyber-physical products and materials, and cyber-physical-social systems, Quadrant 2 in Fig. 1.
4
Jelena Milisavljevic-Syed et al. / Procedia CIRP 81 (2019) 1016–1021 Jelena Milisavljevic-Syed et al./ Procedia CIRP 00 (2019) 000–000
1019
3. Illustrative example
4. The mathematical model
Design of an NMS has several possible applications, see Quadrant 3 in Fig. 1, (e.g., assembling processes, design of a chemical plant, design of a transmission box, etc.). However, in this paper we illustrate design of an NMS with an automotive panel stamping process. The challenges in the design of NMS are illustrated through a two-dimensional panel stamping process in which four parts are assembled across three stations, Fig. 3. Each part is restrained by a set of fixtures (Pi), and each position of a part is measured by a coordinate sensor (Mi). At Station 1 Parts 1 and 2 are assembled. In Station 2 Parts 1 and 2 are made into a Subassembly 1 that are assembled with Parts 3 and 4. At Station 3 all parts are assembled in final assembly. Each part at each station has up to 2 sensors that are measurement points. The total number of measuring points is Mi (i=1 – 20). Fixture locators can be active and inactive. There are two types of fixture locators, namely, a 4-way pin, and a 2-way pin.
In this paper, we address a two-part problem: (1) provide flexibility in the selection and determination of values of design parameters (tool and sensor attributes) that give satisfying (robust) solutions regarding the process cost and quality of product; and (2) provide for the determination of input ranges that give functional system design and satisfying dynamic performance of a system in the presence of changes in requirements. The problem characteristics are the number of work stations, N, the number of parts, n_p, the number of sensors, M_Pi, the number of tools, P_i, and the potential fixture failures in the process, m_r. The requirements for the process are diagnosability, controllability and operability and these must be full. Expected variations, y_k, must lie between -0.8 and 0.8 millimeters. The assumptions are that the expected variations parameters follow Gaussian distribution with zero mean and known covariance. The overall objective is to minimize the process cost and maximize product quality. The first part of the problem is simulated with the adaptability model, Section 4.1, and the second part of the problem is simulated with the operability model, Section 4.2. 4.1. The adaptability model
Fig.3. Two-dimensional panel stamping process [7, 8].
The adaptability model is based on the SoV model that we use to model an NMS and identify mechanical and control system drivers and is further partitioned inot smaller process decision and performance measurement cDSP models, the upper part of Fig. 4. The cDSPs models are used to build in flexibility in the selection/determination of design parameters and manage the structure and information of decision-making. For more information about the adaptability model see [3, 4].
Fig.4. Adaptability and operability model. .
1020
Jelena Milisavljevic-Syed et al. / Procedia CIRP 81 (2019) 1016–1021 Jelena Milisavljevic-Syed et al./ Procedia CIRP 00 (2019) 000–000
5.1. Flexible design of panel stamping process
Table 2. The compromise decision support problem (cDSP) formulation. Given (Parameters) Total number of operational stations Number of stamping parts in the process Number and position of fixture points Potential number of sensors and position Dimensional quality (size of variations) boundary values are set Find (System Variables) Total number of sensors and sensing stations Use of PT`s control actions Sensing penalties Sensors distribution schemes that are diagnosable, controllable, and costeffective Satisfy (Constraints) Tooling constraints Sensing constraints
5
N np
Pi [-]; (x, z) [mm]
MPi [-]; (x, z) [mm]
Yk [mm]
MPi,M ; MS,M (M=D,C,E)
PT P
Mi,k
Use of programmable tooling, Eq. 3.6 in [3] Number, position and distribution of sensors, Eq. 3.6 in [3] 100% or partial, see Eq. 3.3 in [3] 100% or partial, see Eq. 3.5 in [3]
Process diagnosability Process controllability Satisfy (Bounds) MPi,M (M=D,C,E) , Section 4.2 in [3] Lower and upper number of sensors Lower and upper number of sensing MS,M (M=D,C,E) , Section 4.2 in [3] stations Use of programmable tooling PT, Section 4.2 in [3] Lower and upper limit of sensing P, Section 4.2 in [3] penalties − + Deviation variables di , di ≥ 0, Section 4.2 in [3] Goal 1: Maximize process Chapter 4, Section 4.2 in [3] adaptability Goal 2: Maximize process operability Chapter 5, Sections 5.1 and 5.2 in [3] Minimize (Deviations) Minimize deviation function Z = min ∑3i=1(Wi d−i + Wi d+i ); ∑3i=1 Wi = 1; Wi ≥ 0; d−i , d+i ≥ 0; d+i ∙ d−i = 0 (i = 1 − 3), Section 3.2.1 in [1]
4.2. The operability model
The operability model is used to analyze the functionality and dynamic performance of the system as the system requirements change. Operability analysis (OA) is accomplished in two stages: steady-state operability analysis and dynamic operability analysis. In this paper, we focus on steady-state operability analysis. Steady-state operability is used to analyze how different requirements, driven by customer preferences, change system functionality. System functionality is analyzed using a Steady-State Operability Model (SSOM) that is based on OA and the cDSP construct, as shown in the lower part of Fig. 4. For further details of the steady-state operability model see [3]. In Section 5, we present the flexible and functional design of the panel stamping process. 5. Flexible and functional design of the panel stamping process The specification of system variables and their effects on the overall cost and quality of the NMS are difficult to ascertain before implementation. For the sake of illustration, five sensors are assumed to be available for use in the design of the panel stamping process. The results are obtained using a MATLAB simulation.
In this section, we answer the question “How does the selection of design parameters (sensors and sensing stations) influence the cost of the process and dimensional quality of the product?” In Fig. 5 a), it can be seen that when we are using 2 sensing stations the cost of the process is lower than when we are using 3 sensing stations. Further, in Fig. 5 b) and c), the expected size of variations is lower when we are using 2 sensing stations than 3 sensing stations. With the use of 2 sensing stations in the process, we design a low-cost and high-quality system in which all requirements are satisfied. Our recommendation is that “less is more” with a reduced number of sensing stations we lower the cost of the process and obtain a high quality of the product. Further, the dimensional quality is not improved by an increase in the number of sensing stations but rather with a proper distribution of adequate numbers of sensors.
Fig.5. Cost of the process and dimensional quality of the product.
5.2. Functional design of panel stamping process In this section, we answer the question “Is the system functional in the presence of disturbances?” In order to analyze system functionality we need to obtain information regarding Available Input Space (AIS), and Desired Input Space (DIS). AIS are inputs of the process which are able to change over a certain range, identified as the feasible design space. AIS is obtained by exercising the adaptability model, shown in Section 4.1 and Fig. 6, a). DIS is the set of input values required to reach the entire Desired Output Space (DOS). DIS is framed according to customer preferences, Section 4.2 and Fig. 6, b). The Operability of the system is full (a system is functional) where customer preferences (DIS) intersect with ranges of change (AIS) in the presence of disturbances (natural uncertainty) where system can stabilize, Fig. 6, c). In summary, the solution scheme shown in, Fig. 4, is an elegant and efficient way to explore the solution space and identify both flexible and functional design in cyber-physical design and manufacturing. Exercising the example gives the opportunity of visualizing the dependence of the cost on the number and location of the sensing stations in the process, Fig. 5, a).
6
Jelena Milisavljevic-Syed et al./ Procedia CIRP 00 (2019) 000–000 Jelena Milisavljevic-Syed et al. / Procedia CIRP 81 (2019) 1016–1021
1021
Fig.6. System operability.
However, dimensional quality is improved not by increasing the numbers of sensing stations in the process, Fig. 5 b) and c), but rather by the appropriate selection of the design parameters and sensor distributions. Furthermore, the system is functional where customer preferences intersect with feasible design space even in the presence of disturbances as long system can stabilize, Fig. 6, c), under dynamic change of the requirements. 6. Closure In this paper, we present a computational framework, DFDM, suited for theoretical analysis of Networked Manufacturing Systems and for addressing the design needs of multi-stage manufacturing processes. In the context of Industry 4.0, DFDM can be used to design NMSs that are adaptable to dynamic market demands, robust to network/connectivity failures, and capable of mass customization. In the future, DFDM can also be used for design of cloud-based cyberphysical and manufacturing systems, based on adaptability, operability, and reconfigurability. In this paper, the use of DFDM in flexible and functional design was illustrated with an automotive panel stamping example. Within the International System Realization Partnership (ISRP) we are in process of CBDM platformization with possible application of DFDM in cyber-physical design, cyberphysical manufacturing, cyber-physical product and material, and cyber-physical-social system design. Further, we speculate that DFDM will transit to industry and be applied in the steel making processes, additive manufacturing, design of smart healthcare system, etc. For more information about current projects please visit www.liverpool.ac.uk/engineering/staff/jelenamilisavljevic-syed. Acknowledgements Jelena Milisavljevic-Syed acknowledges financial support from the John and Mary Moore Chair and the L.A. Comp Chair for her graduate studies at the University of Oklahoma, Norman, Oklahoma.
References [1] Milisavljevic J. Architecting networked engineering systems. Master thesis. School of aerospace and mechanical engineering. University of Oklahoma 2015. Norman. Oklahoma. USA. [2] Ding Y, Ceglarek D, and Shi J. Diagnosability analysis ofmultistationmanufacturing processes. Journal of dynamics systems, measurement, and control 2002c; 124:1–13. [3] Milisavljevic J. Architecting networked engineering systems. Doctoral dissertation. School of aerospace and mechanical engineering. University of Oklahoma 2018. Norman. Oklahoma. USA. [4] Milisavljevic-Syed J, Allen JK, Commuri S, Mistree F. A method for the concurrent design and analysis of networked manufacturing systems under uncertainty. Engineering optimization 2018. GENO-2016-0753.R2. [5] Shang X, Milisavljevic-Syed J, Wang G, Allen J K, Mistree F. a key feature-based method for configuration design of reconfigurable inspection system. Journal of intelligent manufacturing 2018. JIMS-D-18-00505; under review. [6] Wang G, Shang X, Yan Y, Allen JK, Mistree F. A tree-based decision method for the configuration design of reconfigurable machine tools. Journal of manufacturing systems 2018;149:143-169. [7] Ding Y, Kim P, Ceglarek D, Jin J. Optimal sensor distribution for variation diagnosis in multistation assembly processes. IEEE transactions on robotics and automation 2003; 19. 4:543-556. [8] Ding Y, Ceglarek D, Shi J. Design evaluation of multi-station assembly processes by using state space approach. Journal of mechanical design 2002a; 124. 3: 408-418. [9] Krishnamurthy A, Suri R, Vernon M. Re-examining the performance of mrp and kanban material control strategies for multi-product flexible manufacturing systems. International journal of flexible manufacturing systems 2004; 16. 2: 123-150. [10] Altiok T. Performance evaluation of manufacturing systems. Springer Verlag; 1997. [11] Subramanian S, Uztürk D, Georgakis C. An optimization-based approach for the operability analysis of continuously stirred tank reactors. Industrial and engineering chemistry research 2001; 40. 20: 4238-4252. [12] Koren Y, Heisel U, Jovane F, Moriwaki T, Pritschow G, Ulsoy G, Van Brussel H. Reconfigurable manufacturing systems", CIRP annals 1999; 48. 2: 527-540.