1 September 2006
What are they, and where are they going?
Let’s define hybrid systems ... again.
By Samuel Herb
They claimed to be programmable logic controls (PLCs) that acted like distributed control systems (DCSs).
Various people define these systems by functions, by industries served, by architecture, and even by no label at all. As a result, there is still confusion about the label hybrid systems.
First, we must look at the issues … what is the purpose of any control system?
Many involved overlook the obvious question. Everybody in the plant may realize the current system needs upgrading, but why is that? Specifically, what will it do differently than before?
We can improve production, product consistency, operating efficiency, safety of personnel, process, and product, and cost of operations, as well as total cost amortized over the life of the system. Another way to state this is the need for productivity … not all the latest cute features. Do not confuse the hype of cute features, however, with real technical advantages.
However, look at all the things working against productivity.
Your marketplace changes, government regulations change, social pressures emerge, changes occur in workforce, energy costs and availability fluctuates, and new aspects of environmental issues appear. None of these are technical. Nevertheless, the system must overcome them and still be flexible enough to overcome even newer issues that have not yet happened.
The tools on the road to productivity present us with some dilemmas. That includes the plethora of technology issues that affect whatever system we buy. We will not go into these in depth, but we can at least acknowledge them. Any control system design must be capable of addressing standards, fieldbuses, easy configuration, security, and the rest.
The real issue that has plagued companies both large and small is how to get the most out of what you have—the best use of ALL of your assets.
Diverse production worlds
To understand the idea of hybrid systems in the realm of productivity, we must first understand their predecessors and the design differences between factory automation and process control.
The origins of hybrid systems come from both the PLC world and the DCS world.
PLCs address the issues of factory automation in the discrete industries. DCSs address the issues of continuous control needed for the process industries.
These two categories of industries, though similar, still have some fundamental differences, which influence the design of the controllers that have been covered in previous articles (see InTech www.isa.org/link/WeaponMass).
For decades, those industries that had elements of both discrete control and process control had some very expensive options. Solutions were often complex hybrid kludges of both systems using equipment from more than one supplier, different operating systems, and all kinds of connectivity problems.
The advent of the microprocessor has significantly transformed industry as much as it has the home. Its pervasiveness meant the high volume of processor chips brought prices way down. Creative development in many different fields benefited all those fields in developing tremendous functionality. It is now possible to do many complex things inexpensively.
The hybrid kludge of functions became practical, and no longer a kludge.
So, how do we define hybrid? Well, that depends who you ask.
There are those who define hybrid as the marriage of the discrete functions, which were done so simply and far less costly with the PLCs with the sophisticated analogue continuous control capabilities of the DCSs.
There are those who define hybrid as the capability of batch functions that neither the original PLCs nor original DCSs could perform well without heavy engineering.
There are those who define hybrid as the industries in which the systems work and serve, like pharmaceutical, fine chemicals, food and beverage, and others.
There are those who define hybrid as the architectural marriage of the PLC simplicity and cost with the sophisticated operator displays, alarm management, and easy but sophisticated configuration capabilities of the DCS.
I prefer the last definition. To some degree, that fits what traditional DCS suppliers provide and as well as some PLC providers.
ARC (www.arcweb.com) prefers the industry definition. They found the need to identify the control system itself, so they have coined the term programmable automation controller (PAC). Some suppliers have noticed that term, while others use the name hybrid control. Some merely refer to a small DCS.
Value the hybrid system
You can get a 100 I/O point system with a $1,100 controller instead of $10,000. Project costs are frequently 30-40% lower than a DCS. Hybrids can offer less expensive I/O.
You can start with a rather small system and grow in smaller increments, saving in expansion costs. There is no cost penalty for sophisticated control function blocks within the IEC 61131 configuration tools.
Applications, firmware, software, manuals, and configuration tools can all exist on a removable flash memory card, which allows easy controller updates.
Much innovation has emerged in the area of less expensive, but sophisticated, control systems that smaller companies can afford. Systems today can offer some basic control and I/O redundancy.
Fieldbuses are available. They can have 10/100 Ethernet control network. They are available with rapid spanning tree protocol on a mesh network. They can have peer-to-peer communication between controllers. Some have a real-time clock with network synchronization.
More capabilities are usually available for easier interconnecting with equipment from other suppliers. They work with traditional SCADA applications.
How does a hybrid system differ?
Traditionally, PLCs performed stand-alone, fast acting, highly repetitive tasks, like those found in the automotive industry.
Communications, redundancy, complex operator interfaces were unnecessary as part of their heritage. Diagnostics stayed within the device itself, not as a system.
As systems emerged through system integrators, engineering was unique to each installation, and operator interfaces were add-ons. This too has changed significantly with development by some suppliers who sell into process applications.
These particular suppliers have evolved to the hybrid controllers, often called PACs, strengthening the use of that term.
Now, networked, and with more so-phisticated operator interfaces, the focus tends to be on the controller functions, with increasing process capabilities.
As to DCSs, they have an advantage with large networks over large plant operations. They typically have more sophisticated fault tolerant strategies rather than just some redundancy.
There is usually a more sophisticated alarm handing capability over large systems. Operator interfaces typically will easily manage very large range of capabilities for the many diverse performance functions of a plant.
Often the equipment has a more rugged look and feel. They will usually have more selections and density of I/Os. They typically have more integrated plant-wide application capabilities.
DCSs emerged from expensive central computers. Inexpensive PLCs had been limited in process control functions, requiring expensive installation costs to make up the difference.
The term, hybrid systems, describes a middle ground. However, what has transpired for the traditional single loop controller user? Today his/her sneered at small islands of isolation can inexpensively address and integrate with sophisticated DCS capabilities at prices equivalent to single loop controllers.
The small user can finally respond to economic pressures by linking his set of islands to his desktop for business reporting. Yet that same system can seamlessly grow beyond those limits to a large system only as much and complex as that user’s own budget allows. What is important from earlier efforts … it is affordable, easily engineered, and configured in the same way, no matter the size. There is no different system for very small, mid-sized, or very large systems. The control system does not have to be custom. No up-front investment “just in case” he wants to expand later.
Keys to the highway are the ECS
Even the small user must make the best use of all his assets. He also cannot afford to have islands of automation. His system should really be an enterprise control system (ECS) that takes all of his assets into account.
It will become the only way one can be competitive in one’s own business, but it must be affordable.
Enterprise control systems enable manufacturers to develop solutions that span their entire business without concern for constraints traditionally imposed by crossing the boundaries of the different classes of systems typically found in their plant.
ECSs are specifically for breaking down the traditional barriers that have resulted in the islands of automation and information. The elimination of the barriers to effective solutions has resulted from four specific characteristics of ECSs.
First is full plant floor interoperation. The islands of automation start at the plant floor. Many plants bear the unenviable load associated with a variety of different DCSs and PLCs and other intelligent systems that just do not work well together. Any business solution that requires crossing these systems becomes expensive and time consuming to implement. ECSs offer full interoperation throughout the plant floor by providing a variety of low-cost software interfaces to all proprietary systems and open standards for later systems that support communication standards.
This allows the entire intelligent plant floor to work as a single system environment even when the base technologies have come from multiple suppliers.
The second ECS characteristic is full open communication access throughout the business enterprise. Automation systems have offered a degree of openness, but truly interoperating with the business-level systems can be a costly and time-consuming activity. ECSs support all the latest open standards, such as ISA-95, Mimosa, OPC, Open O&M, and the like, as well as supplier-provided standards, such as NetWeaver from SAP and Biztalk from Microsoft. By incorporating these standards into the ECS infrastructure, the cost and effort required to drive business solutions throughout the enterprise lessen by an order of magnitude. What could have cost millions of dollars with a traditional DCS could now be only in the thousands. Return on that investment could be in months, not years, critical for smaller businesses.
The third characteristic of ECSs that provides manufacturers with specific advantages over older technologies is they support asset performance management (APM). APM is a new methodology for operating process plants that combines the traditionally separate maintenance management and operations management software tools into a cohesive view, so assets are available when actually needed and their use can be managed effectively and controlled together as one system. This combined with the ability to implement real-time accounting and performance measurement systems enables manufacturers to move up from closed-loop process control to closed-loop business control.
The fourth key characteristic of ECSs that provides significant value to a company is a single engineering environment across all plant floor domains. Traditionally, engineers may have had to learn many different sets of engineering and configuration software just to engineer plant floor systems such as DCSs, safety systems, MES systems, PLCs, and HMIs. ECSs move the efficiency of system engineering to a new level by providing one state-of-the-art engineering environment that operates across multiple plant floor domains. This improves engineering efficiency while significantly reducing cost.
More than a cheap DCS, these hybrid control systems can be an inexpensive stepping-stone towards managing the business well.
What is important is these elements can come aboard in manageable increments to meet tight budgets. I guess the term Hybrid Control System becomes a bit misleading now. It is more than a PLC that acts like a DCS. But then again, DCSs and PLCs have grown past their initial role as well. They are all stepping-stones to becoming enterprise control systems, when they can operate as a single system to make a business more productive.
I have trouble with the term SCADA
In recent years, the use of the term SCADA has been applied increasingly to what are really data acquisition systems that, these days, also do control.
They consist of programmable logic controllers (PLCs) with a personal computer (PC) operator workstation. The proper term for these had been data acquisition and control. This type of system has not been the definition of SCADA for over five decades.
Used on a large scale, real supervisory control and data acquisition (SCADA) systems work for control actions and information gathering from beyond the plant. These SCADA systems have not traditionally worked in process control but rather in the starting and stopping of remote units, such as those found in remote power transformers or remote water or gas pumps on pipelines.
The time delays on these SCADA systems have usually made it necessary not to rely on monitoring and controlling the details of process itself from a distance. The supervisory control portion merely turned specific units on or off, or they bypassed units that were not working due to storm damage, an accident, or the like.
Power distribution companies must live with these conditions all the time.
Any communications in such remote transmissions as a SCADA system have to allow for long time delays between the request for action and the action to happen. In addition, a concern is frequent unexpected interruptions of any communicated signal.
This generally precludes any continuous process action, which needs a more responsive operation. For decades, unique technologies accommodated these control limitations.
To confuse the issue even more, since the late 1950s, Supervisory Control in the process world meant to drive the set points of control loops using a central computer. An operator could override this action at the controller itself by switching from supervisory to automatic, to manual.
The practice of referring to a PC connected to one or more PLCs, which has been happening since the mid-1990s, adds still another layer of confusion.
In manufacturing plants, there is the need for a variety of control schemes. There are traditional uses for PLCs, PCs, and process controllers.
One could argue the placement of controllers remotely from the operator make them a SCADA system. If so, what do you call a DCS? I do not believe merely adding a PC on a PLC makes it a SCADA.
Therefore, we must distinguish these different uses of the term SCADA by identifying them by their more direct terminology.
ABOUT THE AUTHOR
Samuel Herb (email@example.com) is a life senior member of ISA and a PE. He has an electrical engineering degree and belongs to the Industrial Computing Society. He is principal of JAOMAD Consultancy in Pennsylvania.
Weapon for mass production: Distributed microprocessing enables control of hybrid operations, without the traditional complexity, cost, and limits. www.isa.org/link/WeaponMass
Understanding Distributed Processor Systems for Control, ISA Press, 1999. www.isa.org/link/DistributedProcessorBK
Barriers bite the dust: DCS, PLC, and PC technologies converge in new breed of hybrid controllers. www.isa.org/link/Barriersdust