07/10/2014 13:35 | Interview with Andreas Deuter and Dr. Jens Dreyer, Phoenix Contact
Interview
Not only with the industry 4.0, the innovation potential of many products shifts more and more towards software. The development processes are, however, focused on the hardware. Andreas Deuter and Dr. Jens Dreyer sketch initiated at Phoenix Contact measures.
In the cover story, the software development is described as difficult to tangible and controllable project, so no classic control. Why is that?
” Software is the deciding factor. ” Jens Dreyer (right) (Image: Editorial IEE / Renate Schildheuer)
Andreas Deuter: There is not this a reason. From my professional experience, as head of development for automation software is so that the development in many companies, even at Phoenix Contact, the components business began. There is hardware-oriented development structures and manufacturing processes.
The topic of software does have won in the last 20 years about the automation technology in the business imperative. However, it was not systematically placed about with IT, but has grown with the increasing intelligence in the field devices gradually. Initially, there was a first processor in a device, which has just been programmed by a developer time. From these small cells out the issue has grown in importance. This may lack even today the full management attention, partly because it is difficult to make tangible and understandable, what actually constitutes software.
Hence, the unilateral costing of equipment and components. Software is in cost accounting are still considered part of the overheads of a company and not its meaning, ie its value and the expense based on separately. This piece of software of a device is not currently priced
So even at the cost calculation of a control
Andreas Deuter:.? In the development costs, the software flows naturally. However, only software externally, for which we pay a license fee has, as a ‘piece of software’ a price we can assign the product. In the calculation of production costs, it looks at the moment often made that the in-house developed software usually has no price
Jens Dreyer. This development costs are mostly in the overheads, which passed on the hardware, funded software development. Everything else is at the beginning of a project is not considered. This works as long as the productivity or cost share of the total software product is relatively low
Andreas Deuter:. But it stands out in recent years an ever increasing shift towards software from causing more and more effort -., and thus costs
How do you assess for Phoenix Contact products this weighting a
Jens? Dreyer: For a controller, for example, the software has several aspects: First, each controller has its firmware and the other a programming software package for engineering is necessary, with which I can use the controller only. Overall, the software is a high double-digit percentage of the overall functionality can therefore be assigned
Andreas Deuter: In addition, the innovation of many products today is in the software – during SPS almost. exclusively. Sure, over the life cycle of ten to 15 years, there are always hardware adjustments – for example triggered by discontinued components. Product developments, such as the implementation of a new Profinet version, OPC UA or improved diagnostic capabilities, all these innovations run largely on via software. This shows the growing importance of the topic
Jens Dreyer. I can confirm that. On the last Hannover Messe a potential customer came up to me and said: Let me see their engineering software, I am looking for a new controller. He did not want to see the hardware, because except for color and form but would be pretty much the same in his opinion. He was concerned mainly about how much effort have its software engineers to create the project. If the development environment enables him to save 30%, then control is interesting for him. The example shows the change: software for decision criterion
But what can be for improving moor by a software-change time
” The innovation lies in the software. ” Andreas Deuter (left) (Image: Editorial IEE / Renate Schildheuer)
Andreas Deuter: This affects not only our customers. We consider our own software development and try to constantly improve. Our developers are busy indeed so that the engineering system is the best, brings a change from C to C # ++ benefits? Would an integrated code analysis to further increase the quality? Here comes exactly on the topic of tangibility, measurability. What is an improvement? We can not believe it really, not yet
Jens Dreyer. Just in terms of objective measurements of improvements we have already noticed that there are still gaps. We and many others can not really prove that concrete measures the error rate has reduced considerably in the development or that our developers create a lot more functions within the same time or with the same effort. This measurability is missing
Andreas Deuter:. On such concepts we are working at the moment and have approaches from the traditional software industry analyzed. So the basic concept for a method by which we can make software development tangible arose, and want we also evaluate course.
As the concept of helping their clients?
Jens Dreyer: The reasoning can be transferred to our customers who use our development environment. Also program the developer in machine and plant and have to control the same problems, to measure the efficiency and – that’s the goal – to increase naturally. I see an analogy here, and the opportunity to benefit from our process later.
How do you measure now concretely the efficiency of programming? They said yes straight, you are at implementing and evaluating
Andreas Deuter:. First, we need the data that we generate in software development, in the first times have tapped. Previous audits by external experts in process quality have for example shown that our data are in five different pots. The requirements are there, the source code here, the test data in another database. And the errors are reportet in a completely different system. In short: The data that we need for a measurement were not taken consistently
Jens Dreyer. we have derived Step one: All the data in the software incurred development, we wanted to have in a single database, an Application Lifecycle Management. Meanwhile, work all our developers, testers, and also the marketing colleagues with this system, so we have all the data available.
Does this mean the measurability achieved?
software is the decision criterion. (Image: Editorial IEE / Renate Schildheuer)
Andreas Deuter: The basics are. We have a first reporting tool written test for the sake that analyzes how many software artifacts were created. How much code has been created? How many test cases we edit anyway? How many errors running during the development or how many mistakes come back after a release?
These first evaluations give us a good feeling that we get valid data. Now we have the whole course test for resilience and control data over a certain time for consistency and correctness. Here we are, in my view a very good way to say, we can measure.
it Does not need these functions in the classical PLC programming system?
Andreas Deuter: This is one of my current tasks, the software development process, as we now live it in product development, introduce in the application software teams with which we drive our solutions business. The programmers who create PC-Worx programs now work with the same lifecycle management system as our software developers. As of today. But why not individual aspects of lifecycle management integrated directly into the engineering system?
If you continue to seize the term engineering, again very different construction sites and interfaces are added, straight toward CAD, cabinet design or classical electrodynamics planning, through to the mechanics. Are there points of contact
Andreas Deuter: These are in principle different data containers and the similar situation, as we here in the standard software development had so far: Every trade union stores its data in itself. The major problem is to make the connection between these data. If it is possible, so to implement a lifecycle management system for the engineering of a plant, then you are one step further. And in this direction think many companies. The basic idea of these concepts is, data – no matter where and when they occur – to centrally manage the entire process. This is at least in Lifecycle Management Software definitely a success factor, and why not also for the planning of entire plants.
How does it look in your view the ideal development or engineering process from a software ?
Deuter Andreas: Well would it be only once to take the time to understand the requirements, in the process, whether in a software project or in the engineering of a plant to be implemented. Sometimes I catch myself doing that, I immediately begin to program at a task that has somewhere to do with software. But this is short-sighted. It is better to make yourself first about the requirements thoughts. This is to question an arduous process, which is supposed to do? In the long run, it is cheaper to implement than something that later turns out to be useless. Only, this way is not so clear, because a short-term result is missing.
The process itself is described in the theory of the planning phase is in a design phase, the programming phase in which quite agile methods can come up with short feedback loops are used. Then follows the end the acceptance and quality assurance phase. The craft stuff is there and described. It may have to be applied even more.
In the software Phoenix Contact is working with various service providers, about their daughter KW-Software. Go your development partner to the outlined path with
Jens Dreyer : The majority of our software developer for four years working under the umbrella of the Center Industrial IT (CIIT) in Lemgo, as well as our colleagues of KW-Software. In some Phoenix-Contact-specific projects we are working there together across teams. Of course, we use exactly the same methods. Service provider for our software now follow our processes and integrate even directly in our Lifecycle Management System. Our partners, we can not impose their processes, but we make sure that they harmonize with our
Andreas Deuter:. For the next generation of our engineering platform, a .NET and C # based Automation Framework, we have created internal process descriptions and first brought all software developers working on the project, on this uniform knowledge
Jens Dreyer. The particular constellation in Lemgo is one reason why KW-Software renamed on 1 January 2015 and as Phoenix Contact Software GmbH is part of the business unit Control system. The staff of KW-Software at Lemgo becomes merged with the software developers by Phoenix Contact. The management take over the current KW-Software’s managing director Andreas Orzelski and Detlev Kuschke, Head of Development Director in Control Systems.
How is the status of this Automation Framework?
“Our goal is, and the efficiency in software development to enhance our customers.” Dr. Jens Dreyer (right) (Image: Editorial IEE / Renate Schildheuer)
Jens Dreyer: The status is very good. We have this year’s Hannover Messe for our smart logic relay a compact engineering tool released – Logic +. This is almost the first software of this Automation Framework. In parallel, we’ve been talking with key customers, will look like the next steps
What influence does the subject industry. 4.0 – the teeth of various components and modules from many manufacturers -. Onto the software development
Deuter Andreas: Industry 4.0 is sustained change the topic of software, both the manufacturers and the companies that work according to industry 4.0. Often there is also a mix. We even develop products for industry 4.0 and put the concepts well into our production. The complexity to develop an industry-4.0 system, which is supplied by several manufacturers is, of course, very large. As with software development topics are reuse, standards and common components real productivity driver. This must, therefore, much more into focus than before and with him the topic of software.
That is, the main issue is then once again interfaces, their standardization, implementation and testing. At the same time the individual function modules are getting smaller. What impact does this have on the programming
Jens Dreyer: analog to the machine and the software has to be built more decentralized. If a machine module is replaced, the software must be able to communicate on the new module with the software of the other modules. Even if all the modules provides a manufacturer, the software must be modular. These interfaces and the communication must be precisely defined. This in turn requires clear structures and responsibilities of the modules in the software.
What is your vision for a master interface for such decentralized approaches from, OPC UA?
Jens Dreyer: It should now look like from all over afterwards. OPC UA is scalable from small appliances to controllers and ERP systems. So suitable for communication in the classical automation pyramid from the bottom to the top. And yet coupled with Ethernet mechanisms. We ourselves are working to incorporate such mechanisms. Visit us at the SPS IPC Drives or at the next Hannover Messe, then we can gladly demonstrate.
SPS IPC Drives 2014
Hall 9 , Stand 310
(sk)
