News & Insights

Most major surprises in a manufacturing facility leave a trail. Often, they stem from capacity constraints that appeared in the data weeks before they became a crisis, and from infrastructure limitations that were visible in the numbers long before they blew a project budget. The warning signs often exist somewhere—in operating data, maintenance records, equipment condition, or the knowledge of experienced employees—but no one has a complete view.

That gap shows up in common stories. A manufacturer brings in an engineering team for what appears to be a straightforward project: add a production line, connect the ancillary systems, and get it running. But as the team pulls on that thread, they discover rooftop air handlers rusting out and operating unreliably.

Further investigation reveals an even larger issue: the chiller plant lacks the capacity to support the new line in the first place. What started as a production line installation has become a utility infrastructure project nobody anticipated until the work was already underway.

Why infrastructure problems remain hidden

The pattern shows up differently from plant to plant, but the underlying cause is almost always the same.

Two systems, one facility, no shared view

Many manufacturing facilities have two data worlds running in parallel. Process equipment communicates with programmable logic controllers (PLCs) located on the equipment or managed by an overarching process control system. But HVAC, compressed air, chilled water, and other utilities report to a building management system, or BMS.

In a scenario like this, both systems continuously generate data. The problem is that they aren’t always sharing that data with each other because that’s not how they were designed.

The BMS doesn’t know what’s happening on the production floor. The process PLC doesn’t know what the chiller plant is doing. Even when both systems serve the same room—a production line below, HVAC units on the roof above—that data is collected in separate places, by separate logic, with no common layer to connect them.

So, when a team evaluates a proposed expansion, they look at each system’s capacity in isolation. What they can’t easily see is how those systems interact under load, how constraints compound, or what breaks first as production scales up.

Institutional knowledge is disappearing

This gap is widening at the very moment when the cost of finding it is rising. Experienced operators and maintenance engineers are retiring at a steady rate, taking with them the institutional knowledge of how a facility behaves. This includes the informal understanding of which systems are sensitive to one another, which pieces of equipment are operating near their limits, and which corners of the plant tend to cause trouble.

When that knowledge leaves, it is often gone for good. In many cases, teams do not realize how much they rely on a particular person’s information until that person is no longer there. Meanwhile, the teams left behind are being asked to make capital decisions faster, manage more complex production schedules, and do it all with a thinner institutional context than their predecessors had.

Capital and operating decisions are made separately

Capital projects are often defined around an immediate operational need: add a line, replace a piece of equipment, increase throughput, or address a maintenance problem. The scope is built to solve that specific issue, and the budget and schedule follow.

But the capital investment and its long-term operating impact are not always evaluated together. A new production line may fit within the available floor space and meet its process requirements, while also increasing demand on chilled water, compressed air, electrical service, ventilation, controls, and wastewater systems.

Because the project may be funded through a capital budget while the resulting utility, maintenance, and staffing costs fall to operating budgets, those downstream effects can be easy to overlook. Each supporting system may appear to have sufficient capacity on its own. The problem emerges when they are all asked to respond at the same time—or when the facility absorbs years of higher operating costs that were never part of the original project decision.

How manufacturers can surface problems earlier

The path forward doesn’t require tearing out what’s already there. In most cases, the biggest gains come from changing how the next project is approached.

Connect operating data across the facility

A key way to close the gap is to treat data connectivity and future expandability as essential design requirements from the start of a project, rather than as considerations to address later. When planning an upgrade, the initial discussions should focus on how the proposed changes will interact with other system components. These will include not only the equipment being replaced but also airflow in nearby spaces, control system integration, and the impact on facility-wide utility demand once the new load is operational.

At a practical level, this means specifying instrumentation that captures real process data rather than simple on/off signals. A pressure switch indicates whether pressure is present, but a pressure transmitter shows the actual pressure, its trend, and whether it’s drifting toward a range that signals a problem.

That difference matters a lot when trying to understand how much of a system’s capacity is being used, or whether a piece of equipment is beginning to show signs of wear. Sensors that track vibration, flow, temperature, and pressure across both process and utility systems give a much fuller picture of facility behavior. That picture is what makes proactive maintenance and confident capacity planning possible.

Connecting systems more deliberately raises a reasonable question about cyber exposure. The answer is that intentional architecture provides a stronger security posture than the alternative.

Most manufacturing facilities already operate with accumulated connectivity, including sensors, remote access pathways, and IIoT devices added gradually over time. Often, these additions occur without updates to network documentation or security measures. This undocumented and unmonitored connectivity is the primary source of cyber risk.

Designing integration with a well-defined architecture and an up-to-date map of connections replaces assumptions with evidence. This creates a solid base for operational visibility and a secure, defensible security posture.

Avoid paying twice for the same infrastructure

Infrastructure planning deserves the same forward-looking treatment. Designing for future expansion doesn’t mean spending more than a project requires today. It means thinking beyond the current scope to ask where the next system will go, what the facility will need in two years, and what it would cost to tear something out later versus accounting for it now.

For example, pipes can be sized for future capacity even if every branch is not installed today. Headers can be stubbed out with valves and flanges, making future connections straightforward. Simple provisions for future implementation, such as a spool of pipe designed to accept a sensor that isn’t needed yet, can cost very little upfront and avoid a significant retrofit later. The goal is to reduce the cost of tomorrow’s work without overbuilding today.

When this kind of thinking is built into the front end of a project, upgrades stop being isolated interventions and become part of a coherent, long-term plan for the facility. The team doing the work today is also setting up the team that will do the work in five years.

Model the facility before committing capital

Even well-designed facilities run into the limits of what engineering teams can see at the planning stage. When a capital project is large enough—a major line expansion, a new facility wing, a significant change to production capacity—the interactions between systems become too complex to reason through on paper alone. This is where performance-based digital modeling earns its keep.

A performance twin is a working model of a facility that takes real inputs and simulates real outputs. These models move far beyond a static drawing or a spreadsheet into a live environment where teams can test scenarios, trace interactions across systems, and see what the infrastructure needs to support a higher production rate. A model that connects the full picture changes what knowledge is available before committing capital.

The value of that knowledge becomes concrete quickly. One large food and beverage manufacturer was preparing to expand a facility and needed confidence in its capital estimates before presenting them to its board for approval. A performance twin was built from their existing operations, with the proposed expansion layered in, and multiple production scenarios were run against it.

The model revealed that their initial estimates had significantly underestimated the stainless steel infrastructure they’d need—the tanks, piping, valving, and pumps required to run at the production rate they were targeting. The gap was nearly $20 million. Finding that before breaking ground, rather than during construction, gave the team the ability to rescope the project and present a number they could stand behind.

The same type of modeling has proven its value in operational settings, not just capital planning. In one case, a manufacturer with five packaging lines was struggling to meet demand despite audits showing sufficient packaging capacity. When a simulation was built around the full facility, the bottleneck turned out to be upstream because liquid manufacturing wasn’t delivering product to the packaging lines reliably enough to keep them running at rate.

Without this simulation, the team was stuck in meetings, trading opinions with no shared data to resolve the disagreement. The model gave them a common view of system behavior and a clear place to focus.

Beyond resolving immediate problems, a well-built performance twin doesn’t expire at the end of a project. Teams use these models as living assets. For example, a team may run new scenarios as market conditions change, compare capacity across multiple facilities, and stress-test assumptions before the next capital decision.

The ability to move from idea to informed decision in days rather than months is a genuine shift in how quickly an organization can act. It compounds over time as the model is refined and expanded.

How can Salas O’Brien help?

If your facility has been encountering surprises during upgrades and expansions, the issue is probably not the individual systems. It’s the gaps between them.

Salas O’Brien brings the breadth of expertise to see across those boundaries and help you understand your facility as a whole system before problems surface on their own. Disciplines include:

• process engineering
• mechanical systems
• controls
• digital modeling

Our engineers approach every project with an eye toward how today’s work connects to tomorrow’s capacity. Whether that means specifying instrumentation that captures the data your future planning will depend on, designing infrastructure that accommodates growth without overbuilding now, or building a performance model that surfaces constraints before they become budget problems, we work to ensure the full picture is visible when decisions are made.

If you’re planning an expansion, working through a retrofit, or starting to think about where your facility needs to go in the next five years, reach out to one of our experts to start a conversation or email us at [email protected]

For media inquiries on this article, reach out to [email protected].