Industrial Automation

Industrial Automation is a categorisation that implies a grade of quality, and applies across many industries. The devices and software involved are generally commonplace, though the application to each industry carries its own nuances. Historically, industrial automation made use of off-the-shelf components from major automation companies, with System Integrators or vendors piecing them together using industry-hardened devices. The purpose has always been improving repeatability and accuracy where humans were less reliable, or becoming more expensive to use for the task.

One major advantage of mature automation systems is capturing operational information, alarms, events, operator actions, and just about any status worth tracking. This “process data” can run the plant more efficiently and safely, and it's genuinely used for forensic purposes in pharmaceutical, heavy industry, and safety systems. Industrial automation has also stayed a fairly closed industry for a generation, open-source hardware and software hasn't been well accepted in heavy industry or critical infrastructure, partly because of “the hardened principle” and the decades of proven performance vendors can point to. That historical performance shapes the risk profile directly, and the risk of loss of life or environmental damage is simply too high to gamble on.

Open-source's real disadvantage is exactly that missing decades of operational history industrial vendors can claim, along with the complexity of version release management and how library updates ripple through interlinked systems and services. Cheaper “kit” hardware and software can deliver the same basic functionality, turning a switch on automatically, at a fraction of the cost, but often with far less than 100% reliability, where a hardened, end-to-end designed system delivers a much higher certainty of repeatability. The price gap up front can be small compared to the price of getting it wrong later.

Some of the largest industrial automation vendors in Australia and globally include ABB, Emerson, General Electric, Honeywell, Mitsubishi, Omron, Rockwell, Schneider, Aveva, Siemens, and Yokogawa. Alongside them sit specialist software companies whose bespoke solutions complement the core automation platforms and add real performance capability, process historians from OSIsoft, GE, and AspenTech are a good example.

Industrial automation vendors and their partners have been doing a version of IoT for decades. As internet speed, security, and connectivity have improved, many already-hardened automation solutions can now make use of public infrastructure that simply wasn't viable before, power, water, and federal agencies once had to fund their own private communications networks at high cost, or share infrastructure, or rely on a hybrid using public systems for redundancy only.

As domestic bandwidth and accountability from public telecom providers have both improved, what was once private network connectivity has effectively become a virtual private network running over public infrastructure. That's pulling big utilities toward shared, more affordable infrastructure. On the product side, most controllers, smart devices, and instruments have been IP-network-ready for some time, and the recent push toward “cloud ready” devices pushes process data straight to a cloud data warehouse, but the underlying principles of embedded controller reliability and purpose-built operating systems haven't changed. That's where industrial automation will hold its ground for a while yet, the applications and analytics around how we interface with these systems are genuinely new, even if the core tech mostly isn't.

The IoT era has brought genuinely useful analytics and interfaces that don't require a engineering degree to operate, fostering a user base more comfortable with purpose-built functionality over highly configurable platforms. Where industrial automation products once needed heavy customisation or a gateway device to share data beyond the control network, that's still true on hardened, cyber-secure networks with no reason to expose real-time data externally, but on less sensitive networks, IoT devices offer a simple, low-cost way to gather data about assets and the environment.

Most IoT devices ship with application software giving near-immediate access via a mobile or desktop app, a real head start for first-time users. Many provide read-only data without control, though others exist specifically to operate an asset, which raises a separate question of who gets what level of access. The crude simplicity of accessing some IoT devices makes critical asset owners understandably nervous, and it's the application and use of these devices, not their raw capability, that presents the real challenge for broader acceptance in industrial automation. Many asset managers keep IoT and Industrial Control System networks segregated for exactly this reason, aggregating data at the enterprise layer through a data historian, even when downstream consumers don't realise the underlying mix of secure and less secure devices.

It's easy to look past IoT's shortfalls: device, application, and operating system limitations, network vulnerabilities, and the support burden behind all of it. When someone loads a consumer app, it's by personal choice, with little pressure to learn it, if it's too hard to use, it gets deleted just as fast as it was installed. That principle doesn't hold in a business environment, where support matters and usability still drives adoption, even though user preference isn't really part of the equation.

Building a bespoke application for every business problem creates a long-term maintainability challenge, less visible today while organisations are still building out their solutions and working out what is and isn't actually a problem. Some automation solutions that skipped good engineering design principles have already created a sprawl of unique configurations and an unnecessary cost burden to maintain. IoT deployments face a similar challenge as varying device types, quality, accuracy, and support levels pile onto the same network. Astute asset managers already factor the full cost of ownership, every supporting peripheral device included, not just the initial purchase price, into evaluating IoT inside an automation solution.

There's also a “delete what doesn't work” pattern emerging in commercial and public organisations, much like consumer apps. Some first-generation IoT solutions are already being retired well inside a year because they didn't meet the business need. With far less complexity than legacy systems that ran for 10 to 20 years, it's genuinely easier to tell quickly whether a newer IoT system is fit for purpose, and the smaller cloud footprint is less intrusive than on-premise infrastructure ever was. For subscription-based IoT, that may be a real step forward in how fast organisations can evaluate new applications altogether.