The history of the automation industry has been a history of engineering and technical advances, yet sometimes these advances seem to barely keep pace with the increased demands from the corporations and organisations that use them. New field devices, more complex processes, a need to feed data into corporate computer networks—all these requirements and more can be a source of frustration for busy control engineers. As automation applications grow and change at an increasing pace, systems that once seemed powerful and future-proof gradually begin to look less capable. Even advanced control systems that have worked well for years are nonetheless limited by the design assumptions and manufacturing capabilities at the time they were specified.
A little history
Automation systems of the past were often developed to serve a specific purpose in a specific industry. For instance, Programmable Logic Controllers (PLCs) were originally developed to replace banks of physical relays that controlled and monitored discrete machines and equipment, such as, the on/off controls in a bottling plant. These relays were essentially switches that were either open or closed, and the ladder logic programming within the PLC was designed to mimic the action of the original relays.
In a largely discrete system, such as this, the job of the central controller is to systematically read the states of all input/output (I/O) points, solve the logic and then write back to the I/O, repeating this pattern over and over again quickly. PLCs are fast.
In contrast, many complex processes such as, oil refining and wastewater treatment use smaller quantities of on/off digital signals. These processes involve large numbers of variable analog signals for temperatures, pressures, variable pumps and the likes. A different type of system, the Distributed Control System (DCS), was developed to accommodate them.
In a DCS, the reading, writing, and logic solving is not concentrated in the central controller. Instead, much of it is distributed to smaller intelligent units, usually located near the device being monitored or controlled. The central controller provides overall supervision. Distributed intelligence was essential for process control because analog signals and the logic needed for them require much greater processing power. Converting electrical signals into degrees Celsius, as just one example, involves a complex mathematical formula. One controller simply could not handle all these analog tasks at once; but a division of labour enabled even very large systems to run smoothly.
Today’s automation applications are less narrowly defined than in the past. Today, both discrete control and process control are often required to get the job done. We can see this trend toward hybrid systems in the offerings of some of the largest DCS manufacturers, for example, Emerson’s DeltaV® and ABB’s Freelance®.
Control engineers, who have spent most of their time developing and running discrete systems, may now need to add process control for the first time. Because of the architectural differences between a PLC-based system and a DCS, approaching a hybrid system from the PLC side can offer some challenges.
Alternatives for process control
Adding process control to a PLC-based system presents problems of cost, integration, and system performance. A full DCS is hard to justify due to its size, expense, and the complexity of learning such a system. For companies doing small to medium-sized process control, for example, in pharmaceuticals, the food and beverage industry, or water and wastewater, a full DCS is, frankly, an overkill.
Using one of the newer hybrid systems from a DCS manufacturer may mean abandoning an existing system that’s working fine. Conversely, integrating a DCS with your existing system is difficult because these systems are proprietary. Traditionally, they were designed as closed systems, both for manufacturing and marketing reasons. PC-based process control is a possibility, but operating system reliability and the cost for industrially hardened hardware are concerns. Integration may also be difficult, involving engineering time and expense.
What about simply using the PLC for process control as well as discrete manufacturing? Newer PLCs offer analog functions, and PLC manufacturers have added functionality to meet process control needs. But both PC-based control and PLCbased systems lack the distributed intelligence that is the main advantage of a DCS. So, any substantial amount of analog logic will slow down scan times, increase network traffic and adversely affect system throughput. While slower scan times for analog devices may be acceptable, a PLC also running digital control cannot afford to slow down critical digital response time.
Supercharging the PLC system
What if one didn’t have to choose between a PLC-based system and distributed intelligence for process control, but could have the best features of both? What if one could supercharge their PLC system for process control? Here’s a way to do just that, by adding intelligent remote I/O to the existing PLC-based system.
Like a DCS, PLC-based systems used to be closed and proprietary. But today, many such systems—for example, Allen- Bradley®, ControlLogix® and CompactLogix® PLCs—use a common communication platform consisting of an Ethernet network and EtherNet/IP, an industrial protocol developed by A-B and currently supported by ODVA (Open DeviceNet Vendors Association).
The major advantage of EtherNet/IP is that it provides a widely used, standard conduit for communication between products from various vendors. So, a PLC that uses the EtherNet/ IP protocol can communicate easily with devices from other manufacturers. And this interoperability gives one choices.
Like the Information Technology (IT) department, which may use Microsoft® software and Dell® computers but chooses printers and peripherals from other vendors, you, as a control engineer, can use EtherNet/IP-capable hardware and software, such as, Allen-Bradley CompactLogix and ControlLogix PLCs and RSLogix, but choose I/O for specific purposes from another vendor.
PLC systems, by their very nature, do not include distributed control. As you are well aware, when you add analog I/O, you must also add new ladder logic to process that I/O. But additional logic and more I/O points eat up processing power, impacting overall system performance by increasing network traffic and slowing scan times.
One of the more exciting choices EtherNet/IP presents, however, is the possibility of augmenting a PLC-based system with I/O that’s designed for process control. You can choose to add the distributed intelligence of a DCS to a PLC-based system. SNAP I/O™ from Opto 22 can augment an A-B ControlLogix or CompactLogix system—or another PLC system that uses EtherNet/IP (such as MicroLogix 1100/1400)— by doing exactly that: providing intelligent remote I/O that offloads many I/O functions, especially those involving the heavy analog signal processing required by most process control applications.
With remote I/O that handles such functions as ramping, thermocouple linearisation, analog scaling, and proportionalintegral- derivative (PID) loop control, the PLC can continue to do its normal job with little impact.
The no-programming alternative
In a PLC-based system, communication with most remote I/O is through a bus coupler. In the past, putting intelligence at the I/O level meant buying another PLC and programming it, either in ladder logic or by learning a new programming language. Both require development time and expense.
The advantage of Opto 22 SNAP I/O is that it does not require programming. All I/O functions are built into the I/O, in a device called a brain. The brain provides communications, like a bus coupler, but it also provides automatic I/O processing. As soon as the I/O is configured, the brain immediately begins processing. Compatibility with A-B PLCs is guaranteed because SNAP I/O is EtherNet/IP conformance tested by ODVA
For process control applications, the built-in analog functions in SNAP I/O, that are especially useful, include PID loop control (up to 96 loops per brain), minimum and maximum values, analog scaling, calibration, ramping, totalising, engineering unit conversion, thermocouple linearisation, temperature conversion, watchdog timeout and output clamping. In addition, SNAP I/O solution offers serial and digital functions on the same I/O rack, including, multiple serial device control (RS 232/485), input latching, digital filtering, quadrature counting, high-speed counting, watchdog timeout, pulse generation, pulse measurement, time-proportional output and frequency and period measurement. The ability to add all these functions in remote I/O, with no programming and little impact on the overall system, may well be the supercharger your system needs in order to succeed with new process control tasks.