Cover Image

PAPERBACK
$45.00



View/Hide Left Panel

Business Process Management Systems to Optimize Manufacturing

CHRISTIAN WILL
Dassault Systèmes

Challenging economic trends, rising value chain complexities, and intensified global competition are driving the manufacturing industry to upgrade its execution systems. And advances in cloud computing, big data, social collaboration technologies, and mobility are prompting society in general toward the digitally connected enterprise and value chain, which must ultimately satisfy the demands of a better-educated and socially aware consumer. These market dynamics and technology advances pose challenges but also offer opportunities to those who successfully leverage and incorporate them into their mainstream. For manufacturers, the challenges are encouraging a fundamental reassessment of their current and future factories.

This paper introduces key concepts behind a manufacturing execution system built on a process-centric software architecture designed to meet these challenges. After introducing the concept of business process management (BPM), I explain how it supports business intelligence by incorporating assisted and automatic decision making into the manufacturing processes. I then explore opportunities for embedding emerging technologies in the BPM approach.

MODEL-DRIVEN DEVELOPMENT

For business information technology (IT) groups, the use of models has played a key role in enabling both technical and nontechnical professionals to work together to debate and define the business processes and requirements of a system. Such models motivated these groups to create new applications, refactor existing ones, or help guide vendor selection and potential customization of an off-the-shelf application. Modeling languages and tools proliferated during the



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 45
Business Process Management Systems to Optimize Manufacturing Christian Will Dassault Systèmes Challenging economic trends, rising value chain complexities, and intensified global competition are driving the manufacturing industry to upgrade its execu- tion systems. And advances in cloud computing, big data, social collaboration technologies, and mobility are prompting society in general toward the digitally connected enterprise and value chain, which must ultimately satisfy the demands of a better-educated and socially aware consumer. These market dynamics and technology advances pose challenges but also offer opportunities to those who successfully leverage and incorporate them into their mainstream. For manu­ facturers, the challenges are encouraging a fundamental reassessment of their current and future factories. This paper introduces key concepts behind a manufacturing execution system built on a process-centric software architecture designed to meet these challenges. After introducing the concept of business process management (BPM), I explain how it supports business intelligence by incorporating assisted and automatic decision making into the manufacturing processes. I then explore opportunities for embedding emerging technologies in the BPM approach. Model-Driven Development For business information technology (IT) groups, the use of models has played a key role in enabling both technical and nontechnical professionals to work together to debate and define the business processes and requirements of a system. Such models motivated these groups to create new applications, refactor existing ones, or help guide vendor selection and potential customization of an off-the-shelf application. Modeling languages and tools proliferated during the 45

OCR for page 45
46 FRONTIERS OF ENGINEERING 1990s, with various methods springing into popularity only to be replaced by oth- ers. Today, the most notable surviving method is the unified modeling language (UML). Since 2000 the market for such tools has seen a steady decline, because of UML’s complexities as well as the transition to agile, lightweight development methods such as extreme programming. But the complexity of systems grew with the advent of XML and Web services, the shift toward componentizing application functionality into service-­ oriented architectures, and the seemingly endless possibilities of transforming them into solutions. Whereas HTML or Hyper Text Markup Language served the first-generation of the Internet well by enabling the creation of Web pages and other content to be displayed in a Web browser, XML, or the Extensible Markup ­Language, brought on the next generation by introducing a set of rules for encoding documents in a format that is both human-readable and machine-readable. XML also formed the basis for describing Web service interfaces, which businesses and applications would soon use to exchange information over the Web. The increased complexity in turn led to a renewed popularity in model-driven approaches. With advances in graphical software modeling tools, models became the de facto stan- dard for IT programmers and users to develop and maintain applications and the primary vehicles for managing systems throughout their life cycles. From Modeling Processes to Executing Them: Introducing BPM Around the time when model-driven development tools based on the UML reached their peak (ca. 2002), a competing camp emerged seeking to transform models to a machine-readable form that could be executed at run time. This camp sought to prevent programmers from touching the underlying code once a process was authored, thus greatly reducing the need for specialized programming skills to manipulate a solution. The result was the emergence of a now large number of software vendors who deliver packaged software known as business process management suites (BPMS). BPM opened the door to nonprogrammers such as business analysts, industrial engineers, and even business users to participate in system development from design through implementation and throughout the life cycle. The rest is mostly history; virtually all of the largest platform houses and dozens of start-up firms currently compete in this segement of software industry. As BPM vendors sought to model primarily processes instead of entire system architectures, specialized markup languages began to take form. To rep- resent and make models transportable across tools, BPM established the business process model and notation (BPML). And to represent an executable process and make it transportable, XML Process Definition Language (XPDL) soon became the de facto standard. As the BPM market grew, a few vendors chose to offer a core set of ­embedded technical and functional components aligned to a specific business context, tar-

OCR for page 45
BUSINESS PROCESS MANAGEMENT SYSTEMS 47 geting particular roles, processes, and associated workflows. Apriso’s FlexNet BPM product, for example, targets manufacturing operations and brings together engineers from material control, quality, maintenance, and production with IT to configure solutions. How BPM Helps Accelerate Manufacturing Excellence Programs Evolution of Organizational Development Methods and the Role of IT Those familiar with manufacturing excellence approaches such as Lean and Six Sigma, both still in widespread use, may recall the intentional absence of any dependence on IT systems such as those for enterprise resource planning (ERP). Others may remember the hugely popular albeit short-lived wave from the early 1990s known as business process reengineering (BPR). While IT played an impor- tant role in BPR, it too was short lived. In many respects BPR was largely a shock-wave approach to push Western companies to quickly respond to threats from overseas competitors that exhibited superior performance in a number of key manufacturing performance metrics and were eroding the US manufacturing base. Those who “survived” the BPR wave by revolutionizing their manufacturing methods soon had to focus on managing and evolving them. By the turn of the millennium, BPR gave way to continuous improvement methods—and business process management was born (Fingar 2006). BPM readily adopted the main tenets of BPR, which called for using models of current business processes as a starting point for business analysis, redesign, and continuous improvement. With a solid foundation as an organizational devel- opment method predicated on evolution, not revolution, BPM advocates knew they had to rely on current IT infrastructures for years to come; IT budgets were severely constrained after 2000 and many companies had yet to show a return on investment on their huge ERP investments. From Enterprise Resource Planning to Business Process Management The search for technologies that could integrate with existing systems, shape functionality to the needs of the as-is environment, support global coordination with local operations within the factories, and serve as the vehicle for monitoring performance and targeting process improvement, led to BPM (Figure 1). The BPM approach also aligned well with emerging technologies such as Web services and integration tools that could readily incorporate the functionality of ERP and other traditional applications as well as describe data and interfaces in a neutral manner using XML technologies. BPM technologies are ideally suited to accelerate process improvement, standardization, and excellence programs because they can translate the outcome

OCR for page 45
48 FRONTIERS OF ENGINEERING FIGURE 1  Continuous process improvement on a business process management (BPM) platform. of process modeling efforts into an executable form that limits or eliminates the need to code or customize a core application. Process knowledge that is captured in the model is thus kept current relative to the system’s implementation, whereas in almost all other approaches models are retired in stacks of paper or on book- shelves after an initial implementation and become artifacts for a firm’s historians. Recognizing that a generic BPM technology is not an application or a solu- tion in a particular business context, the opportunity exists to further accelerate manufacturing excellence programs by preconfiguring and prepackaging many of the core elements needed for a particular industry. Such packages are inte- grated in the BPM software and contain a unified model of the business entities involved, a core set of fine-grained business services to accommodate the busi- ness entities, and a collection of one or more process fragments applicable to

OCR for page 45
BUSINESS PROCESS MANAGEMENT SYSTEMS 49 the various business areas in that industry (e.g., production, material handling, machine maintenance, and human capital management). A library of preconfig- ured, process-centric assets eliminates the need to start from an empty page and provides a solid starting point from which the organization can accelerate its manufacturing excellence programs.1 Making Processes Intelligent Processes configured to run in a BPM system can exhibit operational intel- ligence in one of two ways: by guiding users through the complexities of their process or by autonomously taking some (presumably optimal) action. Both are based on the system’s “knowledge” of the particular problem domain to which it is applied. Such knowledge must necessarily be configured into a process. Although the same approach is used in traditional applications, what differentiates a BPM application is that such intelligence is not coded into the solution. With BPM, intelligent behavior may be added to a single process controlling a work center or automated manufacturing cell, or an end-to-end chain of processes spanning the entire factory, enterprise, or value chain. An example of the former might be the highlighting of a particular sequence for processing work orders to meet competing objectives such as minimizing material consumption, maximizing machine utilization, or focusing on the most important customer. An example that spans the entire factory floor would enhance the just-in-time supply of component materials to production in a manner that accommodates the production schedule while ensuring the efficient use of floor space. The extent to which intelligence can be enabled is limited only by what the manufacturing excellence team incorporates in its processes and by the kinds of optimization and intelligence tools available in a BPM platform. Often, the BPM can integrate popular tools that are available in the marketplace, such as business intelligence (e.g., analytics, data mining), simulation, statistical analysis, artifi- cial intelligence (e.g., genetic, inference/rules-based), and operations research (e.g., smart math or constraint-based optimization). Such tools can be tightly integrated into the BPM’s design and run-time engine, or loosely coupled through a service bus or other means. Hybrid approaches are also possible. For decision support scenarios, the user may receive guidance in the form of inline feedback. For example, a make-to-order engine manufacturer might con- figure its assembly operation to conduct background checks of an engine’s digital configuration to alert an operator about deviations from an authorized bill of mate- rial (e.g., substitutions or engine configuration options) or other checks that need to be performed. Or the system might retrieve specific fail-safe data for an engine 1  Some have referred to the approach of combining BPM technologies with manufacturing excel- lence programs as outlined by McClellan (2012). This is in stark contrast to traditional approaches, which are data-driven and application-centric.

OCR for page 45
50 FRONTIERS OF ENGINEERING configuration to provide additional context to both the automation layer and the operator. Fail-safe data might range from the specification of the items and order of the component materials to be consumed, to the tool calibrations and settings for each step of the assembly operation. Standard work ­ nstructions might be aug- i mented to display configuration-specific deviations through multimedia visual aids, thereby minimizing the need for operator intervention on the computer terminal. Decision support may also manifest in the form of supervisorial dashboards that provide data on the throughput and timeliness of a process, end-to-end perfor- mance across the entire factory, or process abnormalities. In addition, the system may alert production staff through email, texting, and other notification channels. Automated system actions can also be implemented by invoking secondary busi- ness processes (subordinate business processes that are utilized to accomplish a more narrowly focused human-system or system-system function in a common manner), external Web services or applications, or triggering action in a plant’s automation layer (e.g., smart machines or programmable logic controllers). The ability to identify and handle exceptions and to incorporate intelligent decision making in BPM-controlled manufacturing processes is essential. By revealing recurring process exceptions, this capability becomes the core driver for prioritizing and engaging in continuous improvement efforts and opens the door to real-time adjustments of process parameters to improve outcomes. Such adjust- ments range from changing product data (e.g., to authorize substitute materials when a shortage occurs) to eliminating tasks, adding approval steps, redirecting a work item to another work center, or applying alternative business rules. The timescale for such exception handling may involve long-running transactions that last hours, days, or even months. Opportunities for BPM Technology Connecting the Virtual and the Real: The Role of Simulation As market dynamics dictate shorter product life cycles, increasingly complex supply chains, and rising costs to bring products to market, simulation may be useful to replace physical prototyping of new product introductions and manufac- turing processes. While the cost to physically prototype and test new products and processes increases, at a minimum at the rate of inflation, the cost and capabilities of the virtual will continue to decrease. For the use of BPM technologies in manufacturing, a natural place to begin is by looking at the simulation capabilities of product life cycle management (PLM) systems. Although there are many applications of PLM simulation, from design through production to sustenance, those applicable to plant operations include the ability to simulate (1) a sequence of actions and alternatives for a single work center, (2) ergonomics associated with the physical actions of an operator, and (3) the interaction among processes throughout the factory.

OCR for page 45
BUSINESS PROCESS MANAGEMENT SYSTEMS 51 Because PLM systems can manipulate models of processes, they are ideal for simulating how individual workstations and cells perform activities, from the consumption of component materials to the productive steps performed by an operator or machine to the movement of materials out of a cell. In addition, PLM systems can increasingly simulate human ergonomics to ensure safe and efficient movements in keeping with work instructions. If the representation of process models is standardized or based on industry standards, the sharing of PLM simulation capabilities can be readily integrated into BPM technologies. In particular, simulation and optimization tools can be used in tandem to improve machine control systems, factory scheduling, and decision support. The challenge of introducing simulation involves not so much technological capacity as the fact that organizations implementing BPM have been busy for quite some time implementing it and using the more mature analytical frame- works available in the marketplace. So the introduction of simulation is a work in progress. From Analytics to Big Data: Gaining Insights from Both Structured and Unstructured Data Analysis of the manufacturing history of a product to search for the root causes of a quality issue, or clues on how to improve production for a part, or any other pattern of interest requires the storage of large amounts of data. The leading database vendors on the market have done a great job of introducing relational database and business intelligence technologies that structure data in a suitable format for viewing and analysis. But these technologies store data in a predefined format. Thus, although large volumes of data can be stored for decades, the data are structured and the repository is usually a relational database, providing limited ability to search and detect patterns in unstructured data that comprises much of what is stored by an organization or across the Internet. In the case of manufacturing operations data, details about the as-built product, process, or manufacturing history are often dropped for a number of reasons—storage costs, or ERP was deemed the “system of record” and its data- base design could not accommodate such details, or the manufacturing execution system (MES) transactional database could not retain data for any length of time because of performance concerns. Such data would eventually be archived or purged—as would the ability to analyze or gain insight from it later. With the maturity of today’s business intelligence frameworks, it is possible to extract details from the MES’s transactional database and retain them for years in a reporting and analytics data store at a reasonable total cost of ownership. Yet even relational databases and business intelligence data stores cannot keep up with the rapidly growing volumes of data generated by sensors, machines, and other devices in a factory’s automation layer. In other cases, the data are highly varied

OCR for page 45
52 FRONTIERS OF ENGINEERING (i.e., unstructured) and contained in various textual document types, log files, blog entries, and other content stored in collaboration portals or email systems. Enter the world of big data. Until recently only the likes of Google, Face- book, Yahoo, and Microsoft could afford the systems necessary for storing and performing searches of such data. However, big data technologies have become both affordable and usable enough for the average organization and IT skillset. And the big software platform vendors are rushing to introduce, or have already introduced, search engine technologies that work with both unstructured and relational data stores (e.g., Microsoft’s SharePoint FAST and Semantic Search, or Apache Lucene), making it possible to store “live,” unstructured data right in the database engine and query it at will. Inclusion of the capacity to search unstructured data in a BPM platform is of immediate benefit to users. It is truly rare to find a process that does not link to at least several unstructured document types. In the aerospace and defense industries, the collection of volumes of unstructured data for as-designed, as-built, as-tested, or as-maintained products has long been a mainstay, as customers demanded such artifacts and were willing to pay large sums to have them. With today’s advances in big data technologies, and the low cost of storage and computing power, keep- ing such details about products and processes is possible for all industries—and they are beginning to see the value and opportunities in the ability to analyze data they couldn’t dream of touching a few years ago. From Processes to Practices: Encouraging Behavior to Drive Innovation and Agility As many businesses become process-centric and excel at sharing, implement- ing, and improving processes across their global manufacturing base, one might ask what happens when a business becomes too efficiency-focused? What if it wants to intentionally deviate from standard processes to discover more effective ways of manufacturing a new product? To do this, the business might want to encourage open debate between engineering and manufacturing, between pur- chasing and quality, and among other groups. A key capability of next-generation manufacturing is the ability to respond promptly to changes in demand or consumer and market trends. Grieves (2011), in his book Virtually Perfect, mentions innovation and other unstructured “practices” (e.g., whiteboard, blogging, crowd sourcing, and gamification) that an organiza- tion should encourage to achieve corporate goals and develop an agile culture. To help organizations engage in unstructured practices, IT groups often introduce social collaboration and content management tools. The opportunity here is to determine the role of BPM in supporting such practices. Could it help accelerate the transition from new engineering or manufacturing innovations to well-documented and cost-effective processes during the ramp-up to a new product introduction?

OCR for page 45
BUSINESS PROCESS MANAGEMENT SYSTEMS 53 Today, there are significant opportunities for introducing novel BPM approaches to content and human-interaction management. Under content man- agement is the ability to retrieve, create, update, modify, and correlate unstruc- tured content around the context of a process; support video, audio, text, and social streams; and enable content organization around the processes to which content relates. Under human-interaction management is the ability to manage shared work queues and enable advanced visualization, individual and group collabora- tion, support for virtual communities, generating user experiences based on a user’s role within the context of his current activity, the ability to show the on-line presence of users related to the current activities and initiate voice, messaging, and other forms of interactive communication, and managing off-line notifications to affected users—also within the context of an active and managed process. Where the Future of BPM May Lead With BPM technologies generating billions of dollars annually in revenue, one can readily conclude that they pose a significant and growing threat to the market for traditional software development tools. And although the technology can be considered distinct from developer tools because it targets nondevelopers, this is a state yet to be achieved. Most BPM products automatically generate a user interface in some form, but delivery of a responsive user interface that satisfies today’s demanding user on the wide range of fixed and mobile devices available requires developers skilled in Web 3.0 technologies. But the technology has certainly delivered on one of its key promises: the ability to take a continually changing model of a business’s processes and keep it synchronized with an executable form. When that executable form incorporates business logic from a library of preconfigured service-oriented architecture com- ponents and a consistent model of business data, BPM can deliver solutions that satisfy a wide range of business contexts, including manufacturing operations. The inclusion of business intelligence, big data, social, and other emerging technologies in BPM software will make this technology an enduring foundation for any process-centric business that has embarked on the journey from being single-plant and efficiency-focused to demand-sensing with the ability to adapt promptly as markets change. References Fingar P. 2006. Business Process Management: The Third Wave. Tampa FL: Meghan Kiffer Press. Grieves M. 2011. Virtually Perfect: Driving Innovative and Lean Products through Product Life Cycle Management. Brevard County FL: Space Coast Press. McClellan M. 2012. Improving manufacturing excellence: managing production processes across the value chain. White paper. Vancouver WA: Collaboration Synergies, Inc.

OCR for page 45