The HMI is where every dollar of your control system meets reality. Bad HMIs cause alarm floods, missed setpoints, and operator errors that don”t show up in any audit until something explodes. Good HMIs are nearly invisible — operators just know the plant is fine. ISA-101 and the High Performance HMI movement are how we get there.

The five principles that matter

1. Operator”s eye, not the engineer”s

The first commandment of HPHMI: design from the operator”s perspective. Not from your P&ID. Operators do not care about pump P-101A”s data sheet — they care about “is this pump on, and is it doing its job?” Every screen should answer that question in under five seconds.

2. Greyscale baseline, color only for exceptions

The biggest visual mistake in legacy HMIs is over-coloring. Bright reds and greens for normal valves, glowing yellows for normal flows. The result: when a real alarm fires, the operator”s eye has nowhere new to land. HPHMI baselines everything in greyscale. Color is reserved exclusively for abnormal conditions.

  • Process variable in normal range: dark grey on light grey background
  • Process variable trending toward limit: dark amber
  • Process variable in alarm: red
  • Equipment failure: red with red border

3. Information density follows the eye

Use the layout to guide the operator. Critical PV values bigger than supporting metrics. Setpoints smaller than process values. Alarms on the right edge of the screen where operators have peripheral awareness. Trends below related controls so the relationship is spatial, not cognitive.

4. The seven-second rule

An operator should be able to identify the state of a unit operation within seven seconds of glancing at the screen. If they need to read text, the screen has failed. Use shapes, sizes, and positions to convey state — not labels.

5. Alarms tell you what to do, not what happened

“FT-101 HIGH” tells the operator nothing actionable. “Tank 4 overflow risk — stop pump 7” tells them what to do. Alarm text should be imperative and short. ISA-18.2 alarm rationalization is not optional — every alarm needs a defined response, a priority, and a measurable consequence of inaction.

Common HMI mistakes I see in the field

  • Schematic overload — using full P&ID symbols on operating screens. P&IDs are reference documents, not control surfaces. Strip them down.
  • Decorator lighting — bevels, gradients, drop shadows on equipment icons. They add visual noise without information. Flat is faster to read.
  • Hidden setpoints — operator has to right-click to see the SP for a loop. Bad. Setpoint should be visible at all times next to the PV.
  • Unnamed faceplates — operator clicks on a pump and gets a popup with nothing identifying which pump it controls. Always include the tag in the title.
  • Audible alarm tones for every priority — train the operator to ignore the alarm system. Use audio only for critical alarms; visual for everything below.

The roadmap to a high-performance HMI

  1. Audit your alarms using ISA-18.2. Most plants find 30–60% of their alarms have no defined response. Eliminate or downgrade them first.
  2. Design Level 1 overview screens showing each unit operation as a single-glance summary. No process detail, no faceplates — just status.
  3. Convert one operating screen as a pilot. Run it for a month with one operator team and capture feedback. Do not deploy plant-wide on day one.
  4. Standardize. Build an HMI style guide (colors, fonts, sizes, navigation) and enforce it with templates in your DCS or SCADA platform.
  5. Train operators on the new philosophy before deployment. Otherwise greyscale screens look “broken” to people raised on color-everywhere HMIs.

The payoff is significant. Plants that have rationalized alarms and migrated to HPHMI typically see 30–50% reduction in operator workload during normal operation, and a measurable improvement in time-to-detect during abnormal events.

Hierarchy of Display Levels

ISA-101 defines four levels of display hierarchy. Understanding the hierarchy is the foundation of any HPHMI project:

  • Level 1 — Overview: One screen covers the entire plant or a major area. Shows only the highest-level process status indicators. No faceplates, no trends. An operator glancing at Level 1 should know within seconds whether the plant is normal or in an abnormal state. Typically one screen per control area or production unit.
  • Level 2 — Unit Overview: Shows one process unit (a distillation column, a reactor train, a compressor unit) in enough detail to manage it. Key PVs, control loop status, major equipment status. This is the primary working screen for the operator responsible for that unit.
  • Level 3 — Loop/Equipment Detail: A specific control loop, equipment item, or interlock group. Contains setpoints, tuning parameters, and the trend necessary to understand recent history. Accessed when an operator needs to diagnose or adjust a specific element.
  • Level 4 — Diagnostic: Detailed diagnostic information — loop diagnostics, valve diagnostics, network status, device health. Typically accessed by engineers and maintenance technicians rather than operators.

A common mistake is trying to show Level 2 detail on a Level 1 screen. The screen becomes cluttered, the key status information is buried, and operators ignore Level 1 in favour of jumping straight to Level 2. Keep each level focused on its purpose.

Standardising Navigation

Navigation consistency is as important as visual design. Operators who have worked on the same HMI for years develop muscle memory for where things are. A migration project that changes the navigation model without operator input creates serious error risk during the transition period.

  • Place the navigation bar in the same location on every screen — typically left side or top
  • Use breadcrumbs (Plant > Area 3 > Unit 7 > Reactor Feed) so operators always know where they are in the hierarchy
  • Make it possible to go from any Level 3 screen to any other Level 3 screen in the same unit in two clicks or fewer
  • Always show the alarm summary as a persistent element on every screen — never make the operator navigate away from their working screen to see the alarm list

Measuring HMI Effectiveness

Once deployed, measure whether the HPHMI is actually improving operations:

  • Alarm rate per operator position per shift (target: fewer than 1 alarm per 10 minutes)
  • Time-to-acknowledge for critical alarms (are operators responding faster?)
  • Operator error rate (incorrect setpoint entries, valve operations on wrong equipment)
  • Abnormal situation response time (how long from first alarm to controlled situation)

Survey operators directly and anonymously six months after deployment. The engineers who built the HMI are not the right judges of whether it works. The operators who use it at 3 AM during an upset are.