Retrofitting Legacy Plants with IoT: A Step-by-Step Roadmap

Most Australian plants weren’t built with data in mind. They were built to run. Reliably, safely, and for decades.

That’s why plant managers keep asking the same question. How do you bring an old facility into the digital age? And do it without tearing it apart?

The answer, in most cases, is a retrofit. An IoT retrofit adds sensors, gateways, and data links to existing equipment. This lets you monitor and analyse it digitally. You don’t need to replace the core control system.

It’s a lower-cost, lower-risk path to digitalisation. It’s also the path most brownfield sites in Australia are now taking.

The stakes are real. Unexpected production interruptions are estimated to cost industrial manufacturers about $50 billion every year. Poor maintenance alone can cut a plant’s output by 5 to 20 percent. A retrofit is one of the most direct ways to claw that back.

This guide covers six things: why brownfield sites differ from new builds, a 5-step retrofit roadmap, where IO modules fit in, what a retrofit costs, common mistakes to avoid, and a realistic ROI timeline.

Quick definitions

• Brownfield digitalisation. Adding digital monitoring to a plant that already exists.
• Industrial gateway. A device that turns data from old equipment into a format modern software can read.
• Edge computing. Processing sensor data on-site, close to the machine, instead of sending it all to the cloud first.
• IO module. Hardware that adds spare inputs and outputs to a full control panel.

Why Brownfield Sites Differ from New Builds

A greenfield plant is built with modern control systems from day one. Data and connectivity are part of the design.

A brownfield plant is different. It’s already running. It usually has old PLCs, ageing instruments, and years of undocumented tweaks.

That gap changes the whole approach. In a new build, digital systems are designed alongside the process. In a retrofit, digital systems have to work around the process. They can’t interrupt production. They can’t void warranties. And you often can’t assume there’s a full record of how the plant is wired today.

A few practical realities of brownfield sites:

• Mixed-age equipment. Some machines are five years old. Others are thirty. They often use different protocols, like Modbus, OPC UA, or industrial Ethernet.
• Limited documentation. Drawings may be outdated or missing. You’ll usually need to check the site in person before work starts.
• Production keeps running. Downtime is expensive. Retrofit work is planned around existing shutdowns, not new ones.
• Safety systems come first. Any new sensor or gateway must fit in without touching safety interlocks or certifications.

This is why a full “rip and replace” of the control system is rarely the first move. A retrofit works with what’s already there. It adds a data layer on top of proven equipment. Most projects start small, on a few critical assets, before going site-wide.

Brownfield Retrofit vs Greenfield Build

A 5-Step Retrofit Roadmap

A good retrofit follows a set order. Skipping steps is the most common reason projects stall. For example, jumping straight to predictive analytics before you have good data.

Here’s a practical, staged approach.

Step 1: Assess and Baseline the Plant

Before you install anything, audit what you have. Look at asset criticality, failure history, and existing instruments.

From this audit, find the 10 to 15 assets that drive the most downtime or cost. Set baseline numbers too. Things like OEE, MTBF (mean time between failures), and MTTR (mean time to repair).

This step matters for two reasons. It shows you where digitalisation will help most. And it gives you a “before” picture to track progress later.

Step 2: Digitise Maintenance Records

Many brownfield plants still run maintenance on paper or scattered spreadsheets. Before you add sensors, get this data into one system. Most businesses use a Computerised Maintenance Management System (CMMS) to plan, track, and manage maintenance tasks efficiently. 

This creates one source of truth. Work orders, parts used, and downtime causes all live in one place. It becomes the base that sensor data will feed into later.

Step 3: Install Sensors, Connect Gateways, Start Collecting Data

This is where the physical work begins. Bolt-on sensors, for vibration, temperature, pressure, flow, or current, go on the critical assets from Step 1.

Gateways then translate old protocols into IoT-friendly formats. They often use a lightweight messaging protocol called MQTT to move that data.

If your equipment has spare inputs, this step is simple. If it doesn’t, IO modules can bridge the gap. More on that below.

Early wins are common here. Just seeing machine temperature, vibration, or run hours often reveals issues nobody knew about. This happens even before you add any analytics.

A worked example: on a food processing line, new vibration sensors flagged a bad reading on a conveyor motor. This happened within two weeks of install. The team replaced the bearing during the next planned shutdown. No unplanned breakdown mid-shift. That one catch offset a good chunk of the sensor cost.

Step 4: Use Analytics

Once you have some data history, put it to work. Start simple. Set threshold alerts, like flagging vibration above a safe level.

From there, build dashboards. Show asset health, trends, and maintenance priorities across the plant.

As more data builds up, add predictive maintenance models. According to Deloitte research, predictive maintenance can cut planning time by 20 to 50 percent. It can lift uptime by 10 to 20 percent. And it can cut maintenance costs by 5 to 10 percent. These numbers vary by industry and by how mature your programme is. Treat them as a guide, not a guarantee.

Step 5: Scale Gradually

Once your pilot works, extend it. Add more assets and more lines.

This is also the point to standardise. Use the same sensor types, gateway settings, and data formats across the site. That makes future additions cheaper and simpler.

Keep measuring ROI as you go. Track downtime reduction, cost savings, and energy gains. Use those results to justify the next phase. At this point, digitalisation stops being a project. It becomes an ongoing habit.

Where IoT IO Modules Fit In

Here’s a common bottleneck. Older PLCs and control cabinets often don’t have enough spare inputs or outputs. That means you can’t connect new sensors directly.

Replacing the whole PLC just to add a few monitoring points is usually overkill. It costs more and disrupts a system that already works.

This is where universal, programmable IO modules come in. They bridge new sensors and old control gear. They add extra inputs and outputs without a full PLC upgrade. In practice, this means:

• You can add extra IO to existing panels without rewiring the whole cabinet.
• You can collect data from sensors that would otherwise have nowhere to plug in.
• You keep flexibility, since IO modules can be reconfigured as needs change.

Sarom Global works with these challenges often, as part of its plant digitalisation and instrumentation services. It’s also the Australian partner for LUCID’s universal, programmable USB IO modules. In practice, that means when a retrofit hits a hardware wall, there’s usually a simple fix. Not a costly panel replacement.

The hardware isn’t really the point. What matters is what it enables: getting data off equipment that was never built to share it, using gear that’s already proven and already installed.

Retrofitting vs Full System Replacement

What Does an IoT Retrofit Cost?

There’s no single number here. Cost depends on a few factors:

• How many assets you monitor. A pilot on 5 to 10 machines costs far less than a site-wide rollout.
• Sensor type and count. Basic vibration or temperature sensors are cheap. Specialised sensors, like ultrasonic or oil-condition monitors, cost more.
• Spare IO capacity. Spare panel inputs mean lower wiring cost. If you need IO modules, that adds hardware cost, but far less than a PLC swap.
• Network setup. Sites with no existing network may need gateways, wireless access points, or cabling.
• Software fees. Ongoing costs for the dashboard or analytics platform. Usually charged per asset or per data point.
• Labour. Time spent setting up gateways, mapping data, and checking readings against known-good baselines.

Because costs vary so much, most plants get a clearer picture from a site assessment than from a general estimate. Starting with a small pilot, rather than a full rollout, also keeps the first bill manageable. It gives you real data to justify the next phase too.

Avoiding Common Retrofit Mistakes

Most retrofits don’t fail because of the tech. They fail because of how the project is run. Here are the mistakes worth planning around.

• Trying to do everything at once. A plant-wide rollout in one go spreads resources thin and delays results. Fix: start with a few critical assets, prove the value, then expand.
• Ignoring cybersecurity. Connecting old OT (operational technology) gear to IT networks and the cloud opens up new risk. Fix: separate OT and IT networks, encrypt your data, limit access, and follow a framework like the NIST Cybersecurity Framework.
• Vendor lock-in. Picking sensors or platforms that only work with one vendor can block future growth. Fix: choose open, well-supported protocols and modular hardware.
• Poor operator training. A great dashboard is useless if operators don’t trust it. Fix: bring operators in early, keep interfaces simple, and pick a “champion” on each shift.
• No success metrics. Without a baseline and clear KPIs, you can’t prove the retrofit is working. Fix: set metrics like OEE, MTBF, MTTR, downtime hours, and energy use before you start.
• Poor change management. Digitalisation changes how people work, not just what’s installed. Fix: treat it as an organisational change too. Communicate clearly and set realistic timelines.

Expected ROI Timeline

Results vary by plant size, asset criticality, and scope. But a general pattern holds across most brownfield projects:

Many plants see real drops in downtime and maintenance cost once predictive maintenance is fully in place. Actual results depend on plant type, asset condition, and how well the retrofit is run. Treat any specific savings figure with caution unless it comes from a credible, named source.

Conclusion

For most brownfield sites, the practical path forward isn’t a full rebuild. It’s a retrofit.

Add sensors and connectivity to equipment that already works. Start small, on the assets that matter most. Measure the results properly. Then scale gradually, standardising as you go. Replace core systems later, only if it genuinely makes sense.

This approach lowers the cost and risk of digitalisation. It still moves the plant toward better visibility, better maintenance planning, and, eventually, predictive maintenance.

If you’re weighing up where to start, Sarom Global can help. Request a plant assessment to find your highest-impact assets. Or book a consultation to talk through a staged roadmap built around your plant.

Frequently Asked Questions