By Kevin Finnan, Systems Consultant, Yokogawa Electric Corporation
Two main types of platforms form the foundation of modern process automation: the distributed control system (DCS) and the programmable logic controller (PLC). Today’s decision process for a new or upgraded control system, however, requires more than simply choosing between the two. Contemporary control systems could also include other technologies, such as the safety instrumented system (SIS) and the supervisory control and data acquisition system (SCADA)—as well as Industrial Internet of Things (IIoT) functionality and other connectivity to enterprise business systems.
Taking these other technologies and required functions into account, along with additional factors as detailed below, will result in an improved decision-making process, leading to a successful implementation over the multi-decade lifecycle expected from modern control systems.
Selection Process and Key Criteria
The business case for modernizing a control system usually boils down to a timing issue. Modernizing too soon reduces the maximum value of the system over time, while modernizing too late can lead to excessive operational expenses and downtime.
Arriving at the decision point usually requires extensive research and analysis. The cost of migration to a new system must be weighed against the operational costs of shutdowns, equipment failures and plantwide performance degradation. Modeling these costs can determine appropriate financial metrics, such as net present value, return on investment or payback period.
Some control system suppliers can provide tools to assist in determining future CAPEX and OPEX costs, but the cost of maintaining versus replacing should be simple enough to model, even over a long-term planning horizon such as the useful life of the new control system, or the life expectancy of the entire plant (Figure 1).
Figure 1: Example financial analysis comparing operating expenses, initial CAPEX and savings between suppliers.
The decision process should be aimed at obtaining the best control system solution for the price, not necessarily the most advanced system on the market, while reducing expected equipment maintenance and associated expenses. Studies have shown four key challenges in the selection process:
Determining the exact criteria for selection
Determining the method for selection
Obtaining buy-in from all the multiple actors in the decision-making process, each often having their own biases and preferences for particular suppliers
The decision process should determine the lowest lifecycle costs to support a safe, reliable and profitable manufacturing enterprise. If this is used as the decision objective, the selection criteria can be clarified around this financial goal.
The company’s strategic outlook should be the overarching consideration in selecting a control system, with the flexibility and forward compatibility of the new control system key considerations for selection.
With these high-level factors decided upon, the next step is to look at competing platforms.
DCS vs. PLC: Factors to Consider
PLCs and DCSes were initially developed for entirely different applications. The PLC was created to control a large number of discrete I/O, with programming simulating hard-wired relay ladder logic. The DCS was developed to control large continuous processes requiring more sophisticated control schemes such as cascaded PID control loops, process optimization and advanced or multivariate process control.
The differences between PLCs and DCSes are less clear today, although significant differences remain, with mixed systems often appropriate. Some questions to be answered during the selection process include:
What is the product value of a batch or the product stream over a shift?
How complex is the plant start-up and shut-down sequence?
What is the relative mix of discrete and analog I/O?
Do the control requirements include advanced regulatory control, or only simple PID loops?
Must the process adhere to ISA18.2 or EEMUA 191 alarm management standards?
Will the process require a Safety Instrumented System (SIS)?
The BPCS (basic process control system) is the “brain” of the plant, and any selection criteria must include how well the BPCS will support commissioning, startup and operations
DCSes are typically capable of more advanced functions than PLCs, often using function blocks to combine multiple PID controllers, as in a cascade configuration. The key architectural element of the DCS is the ability to distribute control to individual nodes, isolating segments of the process to increase system performance and reduce response times.
DCSes and PLCs also have differing approaches to redundancy. DCS manufacturers typically offer fully redundant systems with “hot switchover” in the event of a fault. This level of reliability is more expensive, but redundancy and reliability are often primary considerations for selecting a DCS.
PLCs can also be used in a distributed architecture, and reliability can be improved by configuring backup PLCs running in “shadow mode”—if a failure is detected the backup PLC assumes control.
The DCS architecture is designed to implement management of change (MOC) in the controllers, which becomes increasingly important as the scale of the system increases. DCSes use databases to enable visibility and modifications of DCS components and control schemes across the plant. By contrast, PLC systems do not have a centralized database, making it more difficult to manage change.
Depending on the operation, a mix of PLCs and DCS’ may result in the best control infrastructure. Plants with PLCs for specialized equipment control, and whose operational elements are spread over a wide area, are examples of this type of infrastructure.
It is important to consider the capabilities of the prospective control system provider for delivering a broad set of integrated technologies, while at the same time ensuring future forward compatibility for both software and hardware.
PLCs have traditionally been preferred for fast response times, but this distinction between PLCs and DCSes has blurred because a DCS can now be used to control fast-responding processes.
Security, Safety and Simulation
The use of standardized Windows and Linux operating systems has enabled greater compatibility among system components, but this same standardization has opened both types of control systems to cyber vulnerabilities like Stuxnet and Flame. Cybersecurity has thus become one of the most important selection criteria.
Security practices and policies for the BPCS and SIS differ in implementation for DCS and PLC-based control systems, and cybersecurity for control systems differs greatly from general-purpose IT systems. IT system priorities are, in order, confidentiality, integrity and availability, while in a control system, availability, integrity and confidentiality are critical—and in that order.
A variety of schemas can be used to defend the BPCS. Some manufacturers isolate the HMI and control network using a secure gateway, and adopt a proprietary protocol for the control network. There are good guides from NIST, ISA and other organizations for recommended security practices.
Control systems typically separate the BPCS from the SIS. The SIS is a separate system used to safely shut down the BPCS and the process. Some suppliers provide a consistent operator interface between the BPCS and the SIS, although each system should have independent event historians, asset management and network communications (Figure 2).
The concept of a “digital twin” has led to the increased adoption of operator training simulators (OTS). The OTS can be applied throughout the lifecycle of the control system, from conceptual design through commissioning and ongoing operations. As a high-fidelity model of the system, the OTS synchronizes with the plant control system to predict plant internal states and responses. The development of the OTS starts in the early planning stages, so its value can be realized in both CAPEX and OPEX reductions.
Figure 2: This system architecture diagram shows an implementation of a safety instrumented system.
Future Control System Architectures
The IIoT offers tremendous promise in many different guises. For example, the application of artificial intelligence, machine learning and data analytics applied in near real-time using multi-node cluster computing including graphical processing units (GPUs) has disruptive potential. Using analytics in a cloud computing environment, artificial intelligence and machine learning can be applied across an entire plant for asset management, energy management, predictive maintenance, process optimization and safety management.
The concept of “edge devices” has changed design architectural thinking. Edge devices operate at the perimeter of the control network, close to the process. Analytics and computing control at the edge is called “fog computing” to distinguish from cloud computing architectures. Fog computing can also be used to isolate the edge from the Internet, and can be used to protect intellectual property.
Several new architecture initiatives have emerged from the Open Group and the NAMUR Open Architecture initiative. The Open Group is developing a standards-based, open, secure, interoperable process control architecture to address both technical and business issues for process automation.
The NAMUR Open Architecture describes a pyramid (Figure 3), similar to the four levels of automation identified in the ISA95 standard. NAMUR standards will address content exchanged among levels, including a safe and secure protocol for handling communications.
Figure 3: The NAMUR open architecture automation pyramid can be used to determine safe and secure protocols for handling communication among the four layers.
Other Supplier Selection Criteria
There are several other important considerations when choosing a BPCS or SIS supplier.
How well does the supplier’s strategic outlook align with the company’s? What is the supplier’s product development roadmap, how do their development plans incorporate new architectures and cloud computing, and to what extent will they provide backward compatibility for new system components?
How likely is the supplier to support the business case? Will the supplier maintain an agile project execution methodology on a global basis? Will the supplier be capable of designing and implementing the project on schedule, including commissioning? Can the supplier act as the main automation contractor if necessary? And, is the supplier’s system designed to minimize maintenance post-implementation?
To what degree will the supplier help reduce commissioning time? Does the supplier’s equipment design address quick commissioning?
Does the supplier provide progressive compatibility for future upgrades?
When purchasing a system, the company is entering into a long-term partnership with the supplier. Are the company’s and the supplier’s strategies and business models aligned? Which supplier can best support the system over the extended lifecycle of a modern control system? Which supplier will best support new technologies and architectures?
Answers to these and other questions can be found in a White Paper titled Control System Selection Key Criteria.
About the Author
Kevin Finnan is a System Consultant at Yokogawa Corporation of America. He was previously an independent consultant serving the automation and measurement industries, Vice President of Marketing for CSE-Semaphore, and Director of Marketing at Bristol Babcock. He has over 30 years of experience with oil & gas measurement, process automation and SCADA systems, and has been active as an author and presenter for numerous industry publications and events.
Check out our free e-newsletters
to read more great articles.