Ulrich Sendler
Independent technology analyst and author
Flantinstr. 12
D-80689 München
www.ulrichsendler.de
Summary
In the last third of the 20th Century application software developed into a decisive driver for new kinds of product development and production and thus into the pivotal motor for industrial innovation. Since roughly the beginning of the new millennium, information technology has become the central factor in the product itself. Embedded software has overtaken the mechanical aspects with regard to its importance for product innovation. Intelligent, so-called cybertronic products are increasingly capable of communicating with each other via both the Internet and other networks.
With this services and functions for use, security, maintenance and many still not known applications that are delivered via such interconnected products, in the future with regard to their importance will stand beside if not above the “actual” product features. The connectivity of a car with other patrticipants of the traffic and the environment for example may already soon be as important as driving itself.
The character of the products is changing in a basic way. This change enforces a change in methods, tools and processes of their development and manufacturing. The domains must interconnect with each other intelligently, while product data, data of the customers, publicly available data of the market have to be accessible throughout the whole value chain. And this has to be so intelligent that everybody is able to access them without knowing them yet and without special knowledge of specific IT-systems. But up to know development and production are still focused on non-interconnected products with a minimal part of software.
The new kind of cybertronic products make it necessary to tear down the central barriers: the established separate educational streams for product developers in university faculties, the established separation between specialist domains in industry, the established compartmentalization of transaction software and engineering software and also the walling-off of monolithic application systems in all areas must be overcome. There is a lot of afford that is still undone.
This understanding specifically leads to three theses:
- Intelligent, software-based products require new multidisciplinary methods, models, tools and processes in product development and production.
- Without professional management of the entire system lifecycle, i.e. without systems lifecycle management (SysLM), intelligent, internetworked and intercommunicating products can neither be developed nor produced, and they cannot be modified quickly enough to meet changing needs.
- Without optimization and systematization of the engineering processes in the early phases of conceptual, functional and behavioral modeling and their interconnection with SysLM solutions, complex networked cybertronic products cannot be integrated into product lifecycle management.
1. Introduction: Status quo
It would be wrong to speak of a paradigm change. Science, industry and providers of IT tools have been working on methodologies and tools as an answer to the challenges of multidisciplinary system development for some time. Examples from the science and research world include the approach of “model-based smart systems engineering” by the Fraunhofer IPK under the direction of Prof. Rainer Stark [SF11]; the “W-Model” proposal by the Institute for Data Processing in Construction (DIK) from TU Darmstadt under the direction of Prof. Reiner Anderl [ANR12]; the “interdisciplinary product development – model-based systems engineering” [EGZ12] concept proposed by Prof. Martin Eigner, TU Kaiserslautern, and on the informatics side, the contribution by Prof. Manfred Broy, TU Munich, and others on “architectures of software-based functions in the vehicle – from requirements to implementation” [BRR11].
The demand for an end-to-end approach in the product and production system process chain (Illustration 1) is also not new. Recently (April 2012), “Conceiving products and production systems in an integrated sense – modeling and analysis at an early stage of product creation” [GLL12] was published.
Illustration 1: Product creation as part of the product lifecycle
Providers of engineering software are working on the integration of the software components into lifecycle data management. Individual tools for functional simulation are on the market and integration of the systems for PLM and MES is part of this development.
New in SysLM is the approach of linking all existing procedures, solutions and tools in such a way that they can lead to sustainable commercial success. Only in this way can it be seen for what tasks suitable methods, models and tools are still missing – in other words where the research community and the manufacturers still have work to do. And only from the perspective of an end-to-end lifecycle management can the future educational and training requirements of engineers be formulated and set up.
Systems lifecycle management is not to be understood here as a new term for an extended PLM. The primary goal is not the management of data, but above all the management of the processes and the organization. This means that the new approach also offers the opportunity of learning from the mistakes made in over-emphasizing the importance of IT tools in past decades and of assigning newer tools a more appropriate role.
The most important goal of the lifecycle management system is the creation of the prerequisites for a better, more effective and more efficient cooperation between the specialist disciplines, for example between development and production, between manufacturer and customer, between scientific research and industry.
2. The digitalization of product development
Digitalization of product development began with the introduction of computer-aided design (CAD) into mechanical and electrical/electronics technology. The growing complexity of the digital product structure compelled the adoption of product data management (PDM). This created a basis that allowed the entire product creation process to be supplied with this data. According to Prof. Frank-Lothar Krause, this includes the taking into account both production and the preparation of production as well as product development, while product development alone ends when production starts [KS97]. Product lifecycle management established itself as the high-level term of choice for the management of product data and engineering processes across their entire lifecycle in a company that has been extended to include supplier chains.
In the 80s and 90s, a new area of industrial product development arose that was concerned with software embedded in the product. The term application lifecycle management (ALM) has come to be commonly accepted for its administration and version management.
Only in recent years has a debate developed – mostly on the research side, but also in parts of large-scale industries – that explores the possibility of adapting methods taken from aerospace and from the software industry’s systems engineering (SE) methodology for use in the general manufacturing industry.
An example of this is the already-mentioned model-based systems engineering approach of TU Kaiserslautern (Illustration 2). It expands the V-Model, the left-hand side of which represents the creation of a system up to discipline-specific model formulation and simulation. At the same time, this illustration shows the necessity of managing the system data across the entire life cycle with the help of a PLM backbone.
Systems Engineering is only one of the approaches that attempts to solve the problem of the multiple disciplines of today’s products. Research projects mostly aim at specific aspects of system development.
Even global digital tools have considerably improved the quality of product development and the products themselves in all the individual application areas, although limitations are clearly shown in regard to the realization of an end-to-end data chain across the entire lifecycle.
Illustration 2: Model-based systems engineering (Sources: Eigner, Gilz, Zafirov, 2012)
3. The digitalization of production and order processing
On the production side, the programmable logic controller (PLC) brought computer support into the manufacturing facility. Instead of relays and failsafes with hardwired controllers, control devices appeared that could be preprogrammed for one specific purpose but could be reprogrammed for other purposes. All steps became programmable – the planning and development of the system and its control, the design of the user interface for the system or machine (i.e. the human machine interface (HMI)), safeguarding of the system and its controller; fault diagnosis, safety mechanisms, commissioning and maintenance, and even the control center itself.
Standard protocols were developed in the 80s for communication between, on one side, the controlling elements of the machines and systems, and on the other for the controller units. The so-called Fieldbus defined who communicated what and to whom at the field level. The industry subsequently agreed on further standards that helped to regulate communication between the production level (the shop floor) and the control system level (the top floor). Fieldbus systems could be integrated using PROFINET (Process Field Network) and information from the field level queried in real time.
It was, however, not possible to speak of end-to-end production (Illustration 3). In general, a mix of systems from different manufacturers could be found and every component had its own control unit. These controllers almost never talked the same language even if they were developed using the same system.
Illustration 3: Automation pyramid (Source Friedl, 2010)
In spite of the progress in production productivity and quality that digitalization has brought, it is becoming increasingly clear that the integration of all process steps in manufacturing companies must be firmly placed on the agenda, since the special optimization of individual manufacturing steps is seen as a limitation.
To complete the picture, a brief look at order and project processing is necessary. Enterprise Resource Planning (ERP) has become the term of choice here. Even though often only one system is used for these tasks, after 30 years no integration is in sight between the systems that control the transaction and those that have been implemented for product development and production. The bill of materials (BOM) continues to be the document that causes the disagreements, with an oft-discussed question being that of the “leading system”, i.e. whether PLM or ERP should carry the responsibility for the BOM.
4. The system needs “simplexity”
From the perspective of the overall company, the many software tools have at the same time become a central element with additional complexity (Illustration 4). The company as a whole is now discussing “simplexity”. This term is an artificial word formed from the English-language terms simplicity and complexity and expresses the goal of being able to treat a complex matter simply. With regard to industrial product creation, the term first turned up in a working thesis presented last year at a German Academy of Technical Sciences (acatach) workshop on the subject of “Smart engineering – interdisciplinary product creation”. It appeared there as “Simplexity for products and processes. The product and process complexity demanded by the market must be responded to with solutions that reduce the inner complexity….”. [AES12]
Illustration 4: The complexity of the industrial IT world (Source: Author, 2012)
This approach applies to the overall view of the development and manufacturing processes for multidisciplinary systems. Their inner complexity, as well as the complexity of the product system itself, will in future undoubtedly continue to increase.
Industry needs a high degree of engineer and technician specialization for intelligent technical product systems, and highly specialized digital development and manufacturing tools are also needed. What is missing is an end-to-end approach in order to master the complexity of these systems and the processes behind them.
The absence of this end-to-end approach is particularly obvious when one takes a look at the strategic questions of a manufacturing company (Illustration 5). The power of innovation often relies on the company being able to recognize the demands of the market at an early stage. The interlinking of the differing IT systems in the various company departments is not provided today.
Converting this into successful products, on the other hand, requires a high degree of interlinking of the operational organization units with one another and with partners. With the structures and the systems available up until now, this can only be realized with enormous effort.
Illustration 5: Strategic interlinking (Source: Author, 2012)
5. Systems Lifecycle Management SysLM
The complexity of the products and processes described affects the entire manufacturing and process industry independently of the size of the company. Indeed, many sectors and companies looking for solutions do so with varying intensity, but up until the present this search does not promise success because the fundamental approaches are:
- Too strongly linked to known procedures, models and organizational structures
- Concentrate on getting results quickly and thus only ever address minor sub-aspects
The reason for this is that the problems are primarily seen at the operational level. It is therefore mainly those responsible for a particular department – development, specialist sections, production, organizational development, sales etc. – who are seeking to solve these with some urgency.
However, since the main problem is concerned with insufficient communication, cooperation and synchronization of all involved, the highest level of company management must first of all accept this task as their own. The measures to be taken, and in which sequence, can only be subsequently determined. (Illustration 6)
Illustration 6: Systems Lifecycle Management SysLM (Source: Author, 2012)
This problem is at the moment assigned a high priority neither by the management level in industry nor by society itself. This may possibly be because Central European industry has in recent years been extremely successful on the world market with intelligent products. This success is, however, endangered if the lack of integration of the specialist departments and the processes is not fundamentally addressed. The current success rests entirely on the high degree of specialization and the second-to-none capabilities of engineers, technicians and scientists in Western Europe.
Even in the mid-term, industry will need new methods, models, organizational structures and processes as well as appropriate tools.
Methods
Conventional methods of development and production are mainly based on the high degree of specialization of individual specialist departments. For intelligent products, the specialization must be supplemented by methods that permit those participating to relate more strongly than at present to other specialists, to achieve better coordination and to take more into account the relationships between one’s own work steps and those steps being carried out upstream, downstream or in parallel.
Models
In many parts of industry, models have become an important means of speeding up processes as well as increasing both their quality and their safety. However the many models that are currently in use were only ever intended to support specific processes and work steps. They are not designed to accommodate multidisciplinary development or to support the entire chain of processes.
Industry is missing practical approaches that:
- Effectively link the many existing models with one another
- Create a joint data model for development and production with a suitable granularity
- Define models that permit functional simulation of product and manufacturing realistic enough to largely make hardware prototyping obsolete
- Generate models that allow end-to-end use of the data across the entire lifecycle of the product and the production system
Organizational structures and processes
Today’s organizational structures and processes came about in most industries to cater for products based on the specialization of specific disciplines. They are based on the assumption that product development and production are to a great extent separate closed processes. But both of these assumptions no longer apply to modern products.
If the existing structures are not to become a barrier to further development, they must become subservient to the end-to-end approach of the multidisciplinary development and manufacture of technical systems. The barriers between the specialist departments must become more permeable. In the end, the successful creation of such structures can be best expressed by taking on personal responsibilities for systems lifecycle management.
Tools
The tools used today are designed to support normal processes and ways of working. Tools – particularly IT tools – that support multidisciplinary development and manufacture only exist in rudimentary form.
With regard to tools, the problem is a threefold one:
- Existing tools must communicate better with one another
- Tools for many kinds of tasks in all areas are missing
- For all tools and their interlinking or integration, the following applies – their operation must be enormously simplified
The challenges described above cannot be solved in the short term. Long-term commitment and large investments, particularly by the manufacturing industry, will be needed. Such investments will however already start to pay back in the medium-term due to a growing certainty that Central European products can maintain, and possibly extend, their leading position on the world market.
The prerequisite for implementation of SysLM is a general and fundamental rethink. Neither the individual components of the product nor the individual tasks carried out by a process, but the product as a system within systems and the end-to-end process from concept through to recycling, are the primary goals of optimization.
To achieve the central goals of Systems Lifecycle Management necessitates intensive research that will require industry to work even more closely with science. Here, both sides can count on the growing attention in recent years that is also becoming evident at the level of the German political establishment, for example with the “Industrie 4.0” campaign started at the beginning of 2012.
Literature
[SF11] Eco tracing – a systems engineering method for efficient tracelink modelling, Stark, R., Figge, A., 2011, conference contribution, published in Publica at IPK
[ANR12] Das W-Modell – Systems Engineering in der Entwicklung aktiver Systeme, Anderl, R. Nattermann, R., Rollmann, Th., PLMportal: The Science and Research Positions, Volume 2012
[EGZ12] Interdisziplinäre Produktentwicklung – Modellbasiertes Systems Engineering, Eigner, M., Gilz, T., Zafirov, R., PLMportal: The Science and Research Positions, Volume 2012
[BRR11] Architekturen softwarebasierter Funktionen im Fahrzeug: von den Anforderungen zur Umsetzung, Broy, M., Reichart, G., Rothhardt, L., Informatik-Spektrum Vol. 34, No. 1, Springer Verlag 2011
[GLL12] Produkte und Produktionssysteme integrativ konzipieren – Modellbildung und Analyse in der frühen Phase der Produktentstehung, Gausemeier, J., Lanza, G., Lindemann, U., Carl Hanser Verlag 2012
[KS97] Das virtuelle Produkt, Management der CAD-Technik, Krause, F.-L., Spur, G., Carl Hanser, Munich Vienna, 1997
[AES12] Interdisziplinäre Produktentstehung, R. Anderl, M. Eigner, R. Stark, in Smart Engineering – Interdisciplinary Product Creation, acatech discussion series, April 2012, Contributors: R. Anderl, M. Eigner, R. Stark, U. Sendler, Springer-Vieweg