Operator interfaces for plants where every decision counts.

• Alarm flood instead of alarm management

  During a fault, dozens of messages arrive at once and the originating alarm cannot be told from consequential ones. The operator learns to acknowledge alarms away — and one day misses the one that matters.

• Displays that show everything and highlight nothing

  The mimic is a full-colour copy of the P&ID with every value the same size. Whether the plant is running at setpoint or drifting is only visible to someone who carries the numbers in their head.

• Unclear states, unclear options

  Is the machine stopped, waiting, or starting up? Is the recipe step finished or merely triggered? If the state display cannot answer that within a second, the operator reaches for the phone, a sticky note or a second screen.

• Knowledge lives with the veterans

  The shift works because two colleagues know every quirk of the plant. Onboarding takes months and consists of spoken rules: “Ignore the value at the bottom, it is never current.” Each of those rules marks a spot where the interface does not create trust by itself.

• Rollout without acceptance

  The MES or line-control software is live and machine acceptance has been signed — and six months later shift handover runs on Excel and shouting again. Training alone does not fix this when the software does not reflect how shop-floor workers actually work, or when key users were never involved.

• Control room / control station
• Process control system (PCS / DCS)
• HMI / plant visualisation
• SCADA
• MES
• Operator display / mimic
• Alarm management
• Alarm flood
• State indication
• Situation awareness
• Operator concept
• Shift operation
• Machine acceptance
• Key user

Control room / control station

The workplace from which operators monitor and run a plant or line. Long phases of observation alternate with short phases of high density in which they must recognise, decide and act.

HMI (Human Machine Interface)

The operator interface between human and machine — from the panel on the machine to the plant visualisation in the control room. ANSI/ISA-101.01-2015 describes the HMI lifecycle from philosophy and style guide through design to operation.

MES (Manufacturing Execution System)

The production-level system between ERP and machine control: orders, worker guidance, confirmations, quality. MES rollouts rarely fail on interfaces; they often fail on terminals that do not support the worker within the cycle time.

Alarm management and alarm rationalisation

The systematic handling of alarms across their lifecycle: philosophy, rationalisation (does this alarm require an operator action?), design, monitoring of alarm rates. Described in ANSI/ISA-18.2-2016, IEC 62682:2022 and EEMUA 191 (Ed. 4, 2024).

Alarm flood

More alarms than an operator can process — typical during start-up, shutdown or an upset. EEMUA 191 and ISA-18.2 give benchmark values for manageable alarm rates.

Situation awareness

Following Endsley (1995): perceiving the relevant elements, comprehending their meaning and projecting their development. A good operator display shows not only values but deviation and trend.

High-performance HMI

The design approach set out by Hollifield et al. (2008): grey backgrounds, colour only for deviations and alarms, analogue representation of target ranges, a clear display hierarchy — so that an abnormal state is recognisable at a glance.

1. 01

   The signal list is not an operator concept

   Many HMI projects start with the question of how to fit every signal on the screen. The better question is which decision the operator has to make in which situation. Thinking from the signal yields a complete display that highlights nothing; thinking from the task yields a hierarchy.

2. 02

   What veterans explain to newcomers is the requirements list

   The most valuable source in a control room is the induction: “Watch out here”, “ignore that value”, “with more than three alarms, look at the trend display”. Every sentence marks a spot where the interface does not create trust by itself. Fix those spots first and you win the operators; modernise only the looks and you lose them.

3. 03

   Automation shifts the work, it does not remove it

   Bainbridge described it in 1983: the more is automated, the less often the human intervenes — and the harder and more important that intervention becomes. Normal operation may be quiet, but the transition into an upset must orient the operator immediately. Our tests therefore always include the scenario in which the automation is not helping.

4. 04

   Observe, then build — for internal tools too

   In steam-turbine manufacturing we worked with industrial engineers for months, created manufacturing routings in SAP ourselves and only then named the core problems: paper parts lists, highlighters, copy-paste from old projects. A searchable interface, a template library and Excel-to-SAP automation cut time on task by 80 to 90 percent.

5. 05

   Modernising means structure first, screens second

   In the legacy software of Ebner Industrieofenbau the decisive step was not the new design but the product and information architecture: which modules, which objects, which terms? Only then followed user flows, high-fidelity concepts and the design-system direction — and existing customers still recognised their software.

Frequently asked questions

Do you know our domain, or do we have to explain everything?

We bring experience from industrial engineering in steam-turbine manufacturing and from modernising machine software for industrial furnaces, and we know ISA-101, ISA-18.2, EEMUA 191 and VDI/VDE 3699. Even so, only your team knows your plant. That is why we start with observation in the control room or at the operator station, not with a questionnaire.

Do you work with our engineering team — WinCC, TIA Portal, Qt?

Yes, that is the normal case. We design in Figma and hand over display hierarchy, components, states, alarm classes and style guide documented so that your team can build in WinCC Unified, TIA Portal, Qt/QML or web technologies. We do not program the visualisation ourselves, but we review interim builds.

Does this work for existing systems, without rebuilding everything?

Yes. Three levers are usually enough: clean up the alarm structure, clarify the display hierarchy, and make the spots unambiguous where operators hesitate today. For grown machine software we work with a migration path — structure first, then screens — instead of a rebuild that scares existing customers.

How does a usability test with operators in shift operation work?

We plan around the shift schedule and usually test with 5–8 operators per round, 45–60 minutes each, at the operator station, on a simulator or remotely. Scenarios come from normal operation, start-up, upset and shift handover. Recordings only by agreement with the works council; we evaluate time to recognise and operating errors, not people.

What does “informed by ISA-101 or VDI/VDE 3699” mean in practice?

That we apply the principles of these standards and guidelines — display hierarchy, state coding, alarm rationalisation, legibility under control-room conditions — and relate our decisions to them. It does not mean that we hold a certification or attest conformity; ISA-101, ISA-18.2 and EEMUA 191 have no certification scheme for that anyway.

Do you also support the rollout — training, key users, deployment?

Yes. We capture the workflows on the shop floor, develop a key-user concept, build training material around real work tasks, prepare acceptance and UAT scenarios from the operator's point of view and support the rollout to further lines and sites including hypercare. The goal is that the software is used after go-live — not merely accepted.

Can you join our project team as an external role — HMI designer, UX/UI designer for machine builders, usability engineer?

Yes. We work as a studio or as a single role in the team of the machine builder, automation integrator or operator — also via staffing partners. Typical role titles: HMI designer, UX/UI designer for machine and plant engineering, usability engineer, UX consultant for plant visualisation. Remote with a base in Munich, on site for observation and tests.

How does an operator concept come about?

Scope and approach depend on the plant, the user groups and the depth of integration. We clarify both in an initial conversation about the starting point and the goal.