Gricha Raether
Product Manager
National Instruments Corp.
Austin, Tex.
|
Programmable automation controllers, like these from National Instruments Corp., complement and enhance PLC operations by performing special tasks above and beyond an average PLC's capabilities, such as machine vision and highspeed process data acquisition and analysis. |
|
This cFP-2120 Compact Fieldpoint processor from NI is an example of the new PACs developed for embedded control or distributed I/O applications. Typical uses for these controllers include measurement taking with data logging, pid loop monitoring and control, realtime situation analysis, and simple I/O operations such as valve actuation and motor control. Ethernet connectivity with a built-in Web server transmits information gathered by the system to any OPC client or HMI/Scada display. |
Programmable logic controllers, or PLCs, are ideal for discrete logic control. However, as PLC process control and automation complexity expands, shoehorning new requirements into these ubiquitous devices can prove difficult. There is a significant and still-growing need for devices with expanded processing power, advanced I/O interfaces, and more control capabilities. Craig Resnick of the ARC Advisory Group coined an appropriate name for these devices. He calls them programmable automation controllers, or PACs. PACs represent the next evolution in automation controllers made possible by exponential advances in computer technology. Thus, companies like General Electric, Rockwell Automation, Siemens, and National Instruments can pack more processing capability into small and highly rugged devices that require less power.
Differences between traditional PLCs and the newer PACs become apparent when comparing software-execution models. PLCs generally follow a linear execution of their programming code. The PLC reads values from input modules and updates its internal input registers. These input values are then used to create or modify other register values according to specific rules of logic dictated by the PLC program. Then the PLC updates its output modules with the newly calculated values and the process starts again.
All PLCs follow this input-process-output model in what is referred to as a scan cycle — an ideal model for discrete logic. In fact, IEC 61131-3 relates how programmable-control systems use ladder logic to reflect this model. Complex control challenges, such as performing several actions in parallel, require more-complex data-acquisition and processing architectures with greater processing power. PACs use advanced hardware architectures and high-level programming tools to concurrently process multiple simultaneous tasks.
Today's control processes rely on a myriad of signals and data, ranging from simple analog and digital I/O to highresolution image recognition and multiaxis motion-control commands. Applications involving high-speed and highprecision production, real-time machine-condition monitoring, and complex process control require deterministic execution of advanced analysis and processing algorithms alongside high-speed data-acquisition systems. High-end PLCs do exist with enough processing power to satisfy some of these requirements. However, many lack resources like floating-point processors and sufficient memory capacity necessary to handle these signals efficiently. Yet this technology is readily available in commercial off-the-shelf hardware developed for the PC industry. PACs integrate commercial hardware with real-time operating systems to provide a cost-efficient platform for high-performance automation.
PACs are not meant to replace PLCs but, rather, to complement them in existing applications. For example, PACs can easily optimize a local or specific part of a production line or process. PAC controllers integrate easily with existing systems thanks to their open architecture. For instance, an engineer may add a PAC to a production line for realtime thermal analysis. It updates the existing control system with the results of that analysis through standard Ethernet or shared registers in commercial gateways. PACs can perform any analysis, such as vibration monitoring, on equipment sending update signals to the overall system. If the PAC detects excessive or improper vibrations, it shuts the machine down before major damage occurs — possibly avoiding severe impact on plant capacity.
IEC 61131-3 presents a predetermined framework to create discrete logic control applications with ease. The standard removes worry about programming execution details making the only concern what goes in and what comes out. However, optimizing a process locally or performing more advanced control functions can be beyond the scope of IEC 61131-3. Programming such applications requires a more-thorough understanding of the subsystem needing optimization, expertise in programming languages that allows parallel task processing, and knowledge in the manipulation of advanced I/O and processing algorithms.
The rigidly defined structure of standard ladder-logic programming is not suited for the advanced nature of these applications. Fortunately, languages such as C/C++ and National Instruments LabView do provide the flexibility required because their execution code defines every detail — an advantage for more-advanced tasks but a challenge for the traditional PLC programmer. Not only must programmers configure and program the I/O, analysis, and control algorithms, they must also define the entire program structure.
This added complexity may overwhelm the simple programmer. Some advanced languages, such as NI LabView, help programmers by using identical development structures no matter what the task. With LabView, programmers use the same graphical interface for machine vision, advanced I/O, motion control, HMI/Scada, and even leading-edge technologies such as fieldprogrammable gate arrays.
Most PLCs today feature fixed-point processors because they are inexpensive and provide the computational power that discrete logic demands. The processors can even handle simple processing tasks like PID control. However, analyzing hundreds of video frames per second, calculating fast-Fourier transforms, performing order analysis routines, or solving advanced control algorithms such as model-free adaptive control, require significantly higher processing levels. PACs routinely include Pentium-class floating-point processors for these intensive calculations.
PACs come in several models, depending on the application. Some dedicate all their processing power to only one type of application, such as machine vision. Others have high-speed processors powerful enough to combine multiple tasks like motion, vision, and high-speed measurements. The combined power updates hundreds of PID loops in microseconds or creates millimeter-precision trajectories for multiaxis motion systems. Intensive computing power like this also requires high-speed inputs and outputs that keep up with program execution.
PACs evolved from the need for easy-to-use devices with processing power comparable to PCs and the reliability and ruggedness of PLCs. Most PACs sport industrial-grade specifications such as expanded temperature ranges from 40 to 70°C, 50-g shock, and 5-g vibration. Real-time operating systems give deterministic performance and operate for extended periods without crashing — a recurring problem with general-purpose operating systems like Windows.
Other industrial concerns include long-term device availability and the effects of commercial off-the-shelf part obsolescence. Vendors deal with this issue by designing PACs using interchangeable parts readily available from different sources. Backwards-compatible design of new versions preserve existing functionality with dropin replacements, and typically expand processing power and enhance features. Vendors usually know about part obsolescence with sufficient lead time to design replacements with no impact on product availability.
MAKE CONTACT:
National Instrument Corp., (800) 531-5066, ni.com/pac