This section discusses a German automotive company with manufacturing plants around the world. In the wake of digitalisation, the company opted to re-engineer and harmonise existing manufacturing execution systems (MES)[1] in their automotive production to build a stable foundation for continuous innovation in the light of Industry 4.0.
The enterprise sought to eliminate poor internal coordination in manufacturing across plants and to optimise processes on the shop floor. This objective was paramount because existing MES were heterogeneous and specific to each plant.
Consequently, processes spanning several plants were extremely difficult to implement. Therefore, it was beyond doubt that a process of harmonising IT systems across plants had to be put into motion, allowing the necessary preparation for Industry 4.0 architecture to be achieved in a set time frame. The new approaches also aimed to reduce processing costs, improve escalation management and enable mobile workflows on the shop floor through cross-plant swarm intelligence. Lastly, use of big data was intended to improve cross-plant pattern recognition of weak spots.
As Industry 4.0 is already changing the working environment and production processes are increasingly automated, machine operators were required to take on a new role.
This included their new function as escalation managers in the case of technical interferences.
The workplace itself became more mobile, being relocated to the shop floor, reducing processing time due to shorter journeys inside the plant and also enhancing the collaboration between technicians across plants through new online channels. As the possibilities of front-end technology, such as robotics in production and the latest forms of displaying information, expand every two to three years, new opportunities for production processes arise. The result here was a front-end development speed that clearly exceeded the actual backend-driven logic of business processes.
In order to improve the entire operating cycle and manage the different speeds of innovation, an API gateway was needed to detach front-end components from MES back-end components.
This API gateway[2] was based on REST (representational state transfer) services[3] granting access to the business logic of the MES for the different task-related front-end components. Loose coupling between these components allows for the establishment of evolution at different speeds, which empowers proper governance of IT at two speeds. This involves a fault-free, more predictable and robot-oriented development of the MES back-end, faster release cycles and demand-oriented front-end design.
Furthermore, instead of using inflexible single-tracked monoliths, task-oriented multi-channel applications have become key.
The front-end components are similar to the apps found on mobile phones. As such, these apps are presented in a task- and user-oriented design and are also managed through a cross-plant device-oriented app store and a virtual portal. This portal is not designed to be a classical portal solution. It is more like a valuable umbrella for all the small apps and provides mechanisms to share all kinds of contexts between the apps and outside systems.
As mentioned above, the workplace of the machine operators saw a shift from a rigid process with predefined activities to an event-oriented philosophy that can attend to escalations and disturbance signals in a more flexible manner. To approach the new challenges, task management was implemented. Via a specific entry on the task list, the system interprets its context and opens the respective app. The app recognises the possible disturbance signals and provides information to the appropriate business logic.
Comprehensive expert panels become obsolete and instead self-explanatory user interfaces take their place. The current context is not only determined by the occurrence of system events and the operators’ new role, but also their position on the shop floor itself. The task list adjusts according to the state of the machine that the operator is standing closest to. This dramatically simplifies the use of all supporting mobile solutions.
Moreover, data streaming[4] was chosen as the means to allow the gathering of data for analysis and predictive maintenance, and to create a foundation for event- and message-driven statistical process control (SPC)[5] of the machinery. In addition, the requirements for the revised IT systems were verified with user journeys, proofs of concept and architecture reviews. Interacting with stakeholders through user journeys helped significantly in the design of new interfaces. This approach will also be adopted in the future and merged with the approach of impact mapping to improve outcomes.
Lastly, the parallel use of classic and agile project management methods appears to be best. Both methods are combined in a multi-project management approach based on PRINCE2[6] and are supervised by product management on the shop floor.
The task of change facilitation was to convey an understanding of why it was necessary to pursue these projects using specific methods and to bring the entire team behind the product improvement.
_____
[1] Schmidt, A., Otto, B., Kussmaul, A.: ‘MES Services in the Automotive Industry’, University of St. Gallen, 2010.
[2] Gamblin, R., Williams, N.: ‘IBM z Systems Integration Guide for the Hybrid Cloud and API Economy’, IBM Redbooks, p. 20, 2017.
[3] Richardson, L., Amundsen, M.: ‘RESTful Web APIs’, O’Reilly Media, 2013.
[4] Psaltis, A.: ‘Streaming Data: Understanding the real-time pipeline’, Manning Publications, 2017.
[5] Celone, R. V.: ‘EZ SPC – Statistical Process Control Demystified’, CreateSpace Independent Publishing Platform, 2013.
[6] Bentley, C.: ‘PRINCE2: A Practical Handbook’, Routledge, 3rd ed., 2017.
