Back to the Future of PLC

Programmable logic controller technology has certainly reached maturity – it is already 60 years old. Which begs the question: will today's PLCs become “retirees” and will future versions of them go to the grave? This assumption seems reasonable given the rapid, sometimes exponential, advances in computer hardware, software, artificial intelligence, cloud services, and communications. These advances have gradually expanded information technology into the previously isolated realm of operational technology.

In light of these events, we present article Jeff Payne's article, published in Control Engineering magazine, discusses the future of industrial automation controllers and applications as PLCs have evolved in recent decades.

Stay true to your cause

The core purpose of PLCs remains the same as it always has been: to provide reliable control and monitoring of physical field devices, even in challenging environments. This has been achieved through the use of specialized processors, operating systems, and programming environments embedded in rugged platforms. However, economies of scale continue to drive the adoption of mainstream consumer and commercial technologies in PLCs wherever feasible. The “smaller, faster, better” principle has remained and will continue to be true, but mostly in the faster and better aspects as the trend toward further miniaturization has leveled off over the past decade.

Many of the benefits of advances in electronic components, processors, and solid-state memory—lower cost, smaller size, lower power consumption, and increased capability—have already been realized in PLCs and other industrial electronics. However, while there will continue to be small improvements in size, cost, and performance, the real advances will be in capability. At the moment, platform size is largely limited by the physical wiring required to interface with the PLC I/O modules. Traditional wired I/O is still necessary, but in many cases, field device communications are shifting to digital networks and distributed remotely using technologies such as IO-Link and wireless.

Multi-core processors integrated into PLC designs now allow deterministic control to be supplemented with extensive additional computing and communication functions. For over 20 years, the term programmable automation controller (PAC) has been widely used to describe an industrial controller with more advanced capabilities than a classic PLC.

While the PAC may have initially seemed like a separate product from the PLC, time has shown that automation engineers are less concerned with the name and nomenclature and much more interested in the performance and features available. Going forward, users will be willing to consider almost any type of hardware or operating system as an automation platform, which may continue to be called a PLC, although it will actually be something more if it can provide real-time control while providing advanced computing capabilities.

Combination of flexibility and logic

While Windows-based systems dominate the consumer and commercial computing world and are prominent in industrial visualization, this is not the case for real-time control. PLC/PAC platforms typically run a proprietary operating system, although some Linux-based options exist. In general terms, users must balance their desire for openness, which provides greater flexibility and lower product costs, with the need for industrial-strength reliability that has historically only been available through proprietary systems. These proprietary systems also provide a high degree of cybersecurity, although primarily at the expense of being somewhat unknown to hackers.

For many years, there has been a trend, or at least a strong interest, towards more open industrial systems, both in terms of hardware platforms and programming languages. Some end users have used standard Raspberry Pi and Arduino hardware to implement automation and data processing projects. Others have avoided such experiments with consumer-grade products due to concerns about reliability. But now several versions of these platforms have evolved into industrial-grade devices (Figure 1). Users are demanding the ability to combine a modern programming platform with proven industrial I/O devices.

With such a diversity of hardware, the next hurdle to openness was the homogenization of the programming environment. Classic PLCs used vendor-specific software that was difficult to port to other brands. The IEC 61131-3 standard introduced consistent PLC programming languages ​​and data types, but vendor-specific implementations still hampered code portability across brands. Eventually, the CODESYS integrated development environment (IDE) offered a more consistent way to create code using standard languages ​​for cross-platform deployment on industrial controllers.

However, none of these initiatives took into account the fact that programmers entering the workforce often preferred to write code in more modern, IT-based languages ​​such as C++ or Python.

Despite all these efforts towards openness and modern programming languages, it is safe to say that classic ladder logic will remain around for the foreseeable future. Ladder logic has a large installation base and remains a simple coding methodology preferred by many electricians, APCS technicians, and even developers. Its graphical style allows for basic troubleshooting and typical industrial automation functions, and its widespread adoption provides additional benefits.

Today, most hardware platforms support ladder logic, either native or implemented through other IDEs such as CODESYS, and many also support other coding methods that can be combined as needed. Different programming languages ​​have their own strengths and weaknesses for specific tasks, and most users enjoy developing their own version of the best tool for the task, balancing flexibility with complexity. An added bonus for users is that moving beyond proprietary languages ​​allows them to create a code library that can be deployed to any type of target hardware, minimizing rework.

The key point today and in the future is that users need automation platforms offered and supported by proven and experienced industrial vendors, with the ability to support any type of preferred programming language.

Communications

Some of the recent advances in industrial automation have been related to improved communications, leading to the complete interconnection of all enterprise systems. As with controller hardware and programming, there has been a shift away from proprietary implementations to more open offerings.

Traditional industrial buses like DeviceNet have long been available to users as proven, reliable devices. But wired and even wireless Ethernet options are now prevalent, with several leading industrial communications protocols available. Improvements in physical form factors, including weather-resistant housings with standard connectors and PoE power, now make Ethernet devices suitable for industrial environments.

Some protocols, such as EtherNet/IP, PROFINET, and Modbus-TCP, are tied to brands and models of end devices, while others are optimized for types of automation tasks (such as EtherCAT for motion control). While EtherCAT is not new, the inclusion of this protocol in more capable PLCs now means that low- to medium-complexity motion control applications can be integrated into an automation platform without the need for separate motion controllers.

Ethernet-APL is an operational technology (OT)-optimized environment that simplifies the deployment of wired Ethernet in field devices. IO-Link is evolving as an optimized fieldbus for basic discrete automation devices with the corresponding communication capabilities and intelligence.

Connecting OT to IT to securely enable Industrial Internet of Things (IIoT) applications and transfer data to support remote visualization and analytics requires a different class of communication protocols. OPC UA and MQTT dominate this role. While some of their capabilities overlap, there are optimal use cases for both protocols, and users can implement them simultaneously. Other supporting tools, such as Node-RED, have become the preferred method for graphical processing and transferring data to the cloud for use by other applications.

From sensor to controller, from local server to cloud and browser – what does it all mean? In the “old days,” smaller controllers had a limited feature set, so achieving full connectivity required larger devices or multiple layers of integration. Today and in the future, users will want these options available on even the most basic and inexpensive automation platforms (Figure 2).

The Role of Integrated Robotics

For many years, robotics has largely existed as a specialized form of automation, requiring custom integration into upstream and downstream systems. This is changing as robotics in general and collaborative robots (cobots) in particular are set to become one of the largest growth areas in all of industrial automation over the next 5 to 10 years (Figure 3). In related developments, machine vision systems have advanced significantly over the past decade, and many have become compatible with robots, allowing them to be easily integrated into a variety of applications.

Modern automation platforms must be ready to keep up with the times by providing the necessary processing power, programming instructions, and technologies to seamlessly integrate with robotics and machine vision. A modern PLC with such capabilities placed next to robotics as an automation platform will have a distinct advantage.

The Role of Artificial Intelligence in the Future of PLCs

No forward-looking article on industrial automation written in 2024 can ignore the potential impact of artificial intelligence (AI) and machine learning (ML) for real-time situational analysis and response. However, there is a lot of noise around this topic because programmable logic controllers are currently not ideally suited for this task as an automation platform. Although in the future, some advanced versions of PLCs will be able to run AI/ML algorithms in real time.

Instead, PLCs are well positioned to act as a field interface for higher-level AI and machine learning resources, providing users with real-time, comprehensive, and context-specific data.

On the other hand, generative artificial intelligence (Gen-AI) will play a more important role in PLCs in terms of code generation in the coming years. Development environments with integrated AI-supporting tools can help users, perhaps even entry-level specialists, develop automation logic based on libraries and verified code. AI used as a development tool can help speed up development, increase code reliability, and minimize unnecessary and routine labor.

The future PLC is part of the automation platform

Over the next decade, programmable logic controllers as we know them will certainly not go away, whether they are called PACs, edge controllers, automation platforms, or something else. But there will also be no single controller technology that can do everything at all price points.

Instead, PLCs will continue to evolve based on available technology and user needs, as they have for the past five decades. Providing real-time control and reliable monitoring will be a priority, but they will add even more advanced programming and connectivity features to improve the user experience and speed of project implementation.

So, PLCs are not going away anytime soon, and new technologies, supported by user demands, will help them evolve as a core automation platform.