He counts the days. Helmut Müller is ringing in the final weeks of his working life. In three weeks, he will be celebrating his retirement, after which the maintenance technician will close his black notebook for the last time - his personal knowledge management tool, his collection of data on the many injection molding machines in the plant. In it, he has noted down all the special features, all the pitfalls, every 'cough' of the machines - over 30 years. The boss always admired him for it, the junior probably also smiled at him.

The notebook remains in the company after Müller leaves, but who can interpret the notes, who understands the data, who knows Helmut's tricks, the youngsters ask themselves. In short, the company faces an enormous challenge, especially as machines and systems are becoming increasingly complex mechatronic systems. These increasingly require close integration between the various trades of mechanical engineering, electrical engineering, information technology and pure IT at ERP level.

In order to get to grips with the increasing complexity of such projects, a module and model-oriented approach is currently establishing itself in engineering; and it is precisely this approach that enables the implementation of future-oriented knowledge management. The concept of modules means that machines and systems are broken down into logical functional units. With prefabricated software modules - for example for filling a container, for feeding parts or for sending an alarm e-mail - the engineering services provided for this can be used several times. Existing systems can also be easily and efficiently expanded with new functions. At the same time, a modular approach enables iterative development and verification processes that accelerate engineering. If the software modules are encapsulated, tested and compiled units that can exchange data via clearly defined interfaces, the machine and plant manufacturer no longer needs to know what the module contains in detail. They can therefore purchase the function or software module as a tradable good as required and integrate it into their solutions.

The trades are already using abstracted models to depict reality. In mechanical engineering, for example, these are CAD plans, in electrical engineering circuit diagrams. In model-based engineering, individual sub-functions are broken down into function modules with corresponding interfaces, resulting in a structured and simplified representation of reality. Depending on the task, there are various options for implementation, for example the Matlab/Simulink model for the control loop or an IEC 61131 program for sequence control.

Focusing on the holistic process

Many suppliers of automation technology have already started to support such a module- and model-based engineering process, partly because some machine manufacturers are explicitly demanding it. After all, they want and need to provide their customers with answers as to what the knowledge ladder of the future will look like.

In general, however, the industry is still strongly rooted in conventional ways of thinking, especially when it comes to the process superstructure. This includes the aspect that the members of the various trades tend to focus on their individual tasks within the scope of their projects rather than having a holistic view of the goal - i.e. the overarching function. For example, when it comes to implementing the 'drilling' function, in simple terms the mechanic usually sets out to design the drilling machine using CAD, the electrical engineer creates a circuit diagram and the automation engineer programs the control system. And only then are the trades linked together.

However, the project should be approached the other way round by first defining the drilling function within the technical process: In what context is drilling to take place, what material is involved, what are the environmental conditions here and so on? This idea of holistic system engineering has not yet been practiced as a rule. However, it is essential in order to be able to grasp complex interrelationships more easily and quickly.

Knowledge is created from data: Machine data is summarized in components and logically assigned.

© Bachmann

In all likelihood, there will not be a standardized 'global language' that all trades can use to communicate directly with each other. A superordinate tool is therefore needed to coordinate the processes. The Systems Modeling Language (SysML), a graphical modeling language for complex systems, already offers an initial approach. In the future, systems engineering will be much more broadly defined and will have to be more process-oriented. What is needed is a dedicated process language that is driven by the perspective of the system manufacturer and should be understandable for participants from all disciplines.

The first steps in this direction have been taken with the Namur Recommendation (NE) 148 in process engineering. With this document published in 2013, the association is pursuing the goal of drastically reducing the time between product idea and market launch and giving operators of process plants more flexibility to adapt to the ever faster changing requirements of customers and markets. To this end, the recommendation examines the entire life cycle of a process plant in the context of a modular plant design and formulates the resulting suggestions and framework conditions for the realization of automation technology.

Bachmann is currently working on the further development of a similar concept that lays the foundation for a solution in mechanical engineering. It consists of a superordinate level for process monitoring and control, which communicates with process-oriented intelligent and autonomous function modules. It is important to create a standardized interface via which all software modules communicate so that users no longer need to program them manually.

Knowledge repository across system and operator boundaries

In addition to the idea of focusing on the holistic process, another important aspect of the optimal engineering tool is the expansion of the module concept. The idea behind this is that software modules are extended beyond the time of commissioning and beyond system and owner boundaries to include knowledge about the specific use and the engineering tool thus develops into a knowledge store. In this way, every 'cough' of the machine is ultimately saved.

The new type of data management ensures that only authorized persons, each equipped with target group-specific usage rights, have access to the data collected worldwide.

© Bachmann

So much for the theory. But what does current practice look like? Capital goods such as presses have a service life of 30 years. During this period, they are repeatedly rebuilt. The operator's primary objective over all these years is to maximize the productivity and availability of the machine or system. The standard procedure of Müller and his colleagues is currently preventive maintenance. This means that wearing parts such as the main cylinders of a die casting machine or the bearings of a roller are replaced preventively according to certain cycle specifications, regardless of the actual state of wear. This approach can lead to these highly stressed components being replaced too early, too late or at an inopportune time, resulting in unnecessary costs.

In contrast, condition-based maintenance, which is based on information about actual use and actual wear, offers clear advantages. However, the inhibition threshold for switching from preventive to condition-based maintenance is still high in the industry because of the high effort and costs involved. This fear is often unfounded, as intelligent software concepts that use the data largely provided by existing hardware can already be implemented today without much additional effort. Bachmann is working on such a concept and is also integrating the visionary idea of using data aggregation to intelligently reuse information on actual machine usage across different customers.

This approach offers the machine manufacturer various added values: the usage data provides valuable facts and suggestions for the optimal design of future machines. This is an advantage that should not be underestimated, given that currently only 20 to 50% of the functions developed for a machine are actually used later on. In addition, the usage data enables the machine manufacturer to provide end customers with targeted advice on the best possible use of their machines. For example, if a machine manufacturer knows that the drill bit of 70% of all the drills of one type in use breaks after 1000 drill holes, then this is an empirical value that he can pass on with recommendations for replacing the drill bit or drilling unit.

As the machine manufacturer is now in a position to offer the end customer better planning of service calls, extended warranty periods and new service offers are also conceivable. These could be provided in exchange for permission to access data. The end user also benefits from the fact that their machine or system not only informs the service or maintenance personnel in the event of severely worn components, but can also initiate reduced operation, for example by waiting for a more favorable time for the necessary maintenance measures. The topic of standardized interfaces is also becoming increasingly important in this context. Bachmann relies on standards that are already available, such as OPC-UA or HTML5.

Bachmann is currently gaining initial experience with the extended modular approach in engineering with the new Fleet Management System (FMS) software module. This collects, stores, aggregates and analyzes data from the sensors and actuators installed in the machine - such as the number of pump starts or valve switching cycles, pump start/run-up times or even the running performance of cylinders. The concept is based on the VDMA standard sheet 24582 for monitoring functions in pneumatics, electrics, hydraulics and mechanics.

Module extension using the example of fleet management

he information is made available to other local modules, the visualization and the MES system via a system variable interface. In this way, warning messages such as "Sand filter is clogged. Clean filter" can be forwarded directly to operations management and linked to the system visualization so that, for example, operation can be shut down due to increased wear until the filter can be replaced.

In addition, the FMS supports distributed data management, which allows the monitoring of many systems even across company boundaries with different authorizations. This means that the machine manufacturer can query the data of machine X worldwide not only at customer A, but also at customers B and C, so that he receives reliable, comparable information about the different uses of the machine. This is made possible by the 'private cloud' principle based on Atvise technology. This combines the advantages of the globally available cloud for data storage and data exchange with the option of assigning usage rights to specific target groups using an easy-to-use, preconfigured web portal. Flexible, scalable user interfaces then allow convenient evaluation of the data depending on the various user requirements, from development to service. And then Müller's maintenance card will finally be a thing of the past!

Author: Georg Scharf is Product Manager Engineering at Bachmann Electronic.