There are a large number of applications in the process technology environment that are difficult to implement: cascaded controls, fast torque controls or even the implementation of a complex mathematical formula in a control application. Although such applications can usually be programmed in IEC 61131 code for a PLC, this is not necessarily convenient. If the programmer has created a difficult formula according to IEC 61131 in structured text, for example, it is often unrecognizable. Accordingly, optimization or troubleshooting is also not easy. The next step is to test the complex function. Obstacles also arise here if there is no clear overview. This is because improving a control system proves to be time-consuming even for less difficult tasks.

The challenges mentioned above are, so to speak, classics that have been occupying programmers for a long time. However, new topics are being added. In the process industry, automation is increasingly based on modular approaches, especially since a standardized software interface has found its way into the industry with MTP (Module Type Package) in accordance with VDI/VDE/NAMUR 2658. The individual modules now have their own controller, which takes over the controls. This reduces the load on the central control system. However, the programming environment in which the user implements the control system cited as an example changes: previously the control system environment, now general control technology.

Matlab Simulink software can be used to implement such complex tasks more simply and elegantly in control technology. The tool is a highly developed tool with which both mathematical and control engineering solutions can be developed, tested and converted into code. The idea for Matlab was born back in the 1970s in the USA by Cleve Moler at the University of New Mexico. The software was then turned into a commercial product in the mid-1980s. From the 2000s onwards, Matlab became more widespread, partly because the tool was included in university curricula.

The subsequent expansion of Matlab to include the Simulink module means that the programs created can now be simulated and blocks of them used to model systems. This allows the systems to be visualized clearly and graphically. Numerous so-called toolboxes are available for the software, which allow the use of ready-made function modules such as the 'Signal Processing Toolbox' or the 'Fuzzy Logic Toolbox'.

As already mentioned, such solutions have been around since the 1980s. What is new is that the user can now easily combine the system with an industrial controller. Where previously pure microcontrollers were available as a runtime environment or C code could be generated, the PLC has now been added as a 'target' for the code. This means that a complex controller can now be designed in Simulink, tested in the entire frequency range under the influence of disturbance variables, optimized with the help of Simulink tools and finally exported as a program for the PLC controller.

Export to the PLC system

Figure 2: Selection of the AXC F 2152 PLCnext controller as a target. A target is the 'target' controller on which the application is to run.

© Phoenix Contact

This opens up a wide range of possibilities for the user that are not usually available in an engineering environment for controllers. In particular, testing and improving the controllers proves to be a tedious process in which the engineer is grateful for any relief. Matlab Simulink includes all the classic methods of control engineering, such as the Bode diagram, root locus curves or the locus curve of the frequency response, which simplifies code development. Another advantage of the software is the use of graphical blocks for designing the controller. For example, a gain can be added simply by drag & drop, whereby the design takes place in a level more suitable for control engineering than in a PLC environment(Fig. 1).

If the programmer has planned the controller in a conventional PLC environment, he is confronted with the problem of testing. Although input signals can be simulated in such an environment, this proves to be difficult over the entire frequency range, as a conventional PLC system does not provide a function for this. With the extensive possibilities of Simulink, commissioning in the field can be considerably shortened if the programmer has already been able to check the behavior of the controller in advance. Once the controller has finally been modeled and tested, the innovation in the system comes into play: the designed controller can be exported to suit the respective PLC system. Targets - so-called PLC coders - are available for the TIA Portal from Siemens, B&R, Codesys, Phoenix Contact and several other manufacturers. Connection to hard real-time systems with the complete I/O portfolio and fieldbus systems is now also easier. From a modeling perspective, Matlab Simulink has thus developed into a tangible and useful tool for automation technology(Fig. 2).

With PLCnext Technology, Phoenix Contact has launched a new, open controller generation based on the Linux operating system, which offers various advantages. The AXC F 2152 controller, for example, with two independent 800 MHz processors, offers a good platform for running a Simulink model on one core and executing standard tasks on the other core, for example.