2.1 SCADA, PLC, HMI, and Data Links
Key Takeaways
- A field instrument measures the process, a PLC or RTU handles inputs and configured control logic, SCADA supervises and stores data, and the HMI presents information and authorized controls to the operator.
- A displayed value is trustworthy only when its tag, unit, timestamp, quality or communication status, and field condition agree.
- A frozen value, lost communication, or mismatched analog span can make a healthy process look abnormal—or conceal a real upset.
- Class I operators should verify conditions, follow approved SOPs, document actions, and escalate control-logic work to authorized instrumentation or controls personnel.
Quick answer: A Supervisory Control and Data Acquisition (SCADA) system is the plant-wide layer that gathers, displays, records, and may supervise process data. A programmable logic controller (PLC) or remote terminal unit (RTU) receives field inputs and carries out configured logic. A human-machine interface (HMI) is the operator's screen. On the exam, treat every screen value as one link in a measurement chain—not as automatic proof of the field condition.
Why the data path matters
The 2025 WPI Class I outline expects application, not just acronym recall: operators interact with SCADA, understand data-communication integration, recognize PLC programming and maintenance concepts, and operate an HMI.
| Layer | Main job | What the operator asks |
|---|---|---|
| Field device | Senses or acts on the process | Is the sensor, valve, or pump actually in the stated condition? |
| Signal and network | Carries a value or status | Is communication healthy, current, and correctly scaled? |
| PLC or RTU | Reads inputs and executes approved logic | Is the controller running its configured sequence and receiving valid inputs? |
| SCADA server or historian | Collects, time-stamps, alarms, and stores | Is the tag current, complete, and available as a trend? |
| HMI | Presents data and authorized commands | Does the display show the correct tag, unit, mode, and operating state? |
A sensor value travels toward the HMI; an authorized operator command can travel back through the supervisory layer to a controller and final device. The HMI is not the PLC, and clicking a screen object does not itself prove that a valve moved or a pump started.
Read a SCADA value as a packet of evidence
Before using a number, inspect more than its digits:
- Identity: Confirm the tag and process location. Filter 2 effluent turbidity is not combined filter effluent turbidity.
- Units and scale: Confirm gallons per minute versus million gallons per day, or milligrams per liter versus another unit.
- Time: Look for a fresh timestamp and a changing value where normal process variation is expected.
- Quality and mode: Note communication alarms, out-of-service flags, manual or automatic mode, maintenance status, and whether the equipment is actually online.
- Corroboration: Compare the screen with a local gauge, equipment feedback, another related tag, or an approved independent measurement.
EPA technical guidance describes the common 4–20 milliampere signal: 4 mA represents the low end and 20 mA the high end of the configured span. The transmitting instrument and receiving controller must use the same span. If one side is rescaled and the other is not, a perfectly valid current becomes an incorrect displayed process value. The safe operator response is to flag the mismatch and involve authorized instrumentation staff—not to improvise a live logic change.
Separate process failure from data-path failure
A process upset usually has physically consistent companions. Rising raw-water turbidity may be followed, after realistic travel time, by changes in coagulant demand, settled-water performance, and filter behavior. A data-path problem often violates those relationships: one value freezes exactly while related variables move, a command status changes without physical feedback, or the HMI disagrees with a local gauge.
Use the sequence observe → verify → act → document:
- Observe: Read the alarm text, timestamp, mode, trend, and associated tags. Acknowledging an alarm means it was seen; it does not correct the cause.
- Verify: Follow the plant standard operating procedure (SOP). Check a safe local indication or independent measurement and contact the control room or field operator as appropriate.
- Act: Take only authorized process actions. If water quality may be affected, protect the process first under the emergency or upset procedure and promptly escalate.
- Document: Record the original indication, verification, action, notifications, and outcome. This preserves the timeline for operations and maintenance.
Pattern-to-response guide
| Pattern | Likely concern | First safe response |
|---|---|---|
| Value flatlines while related tags vary | Frozen input, lost signal, or stale display | Check timestamp/status and verify locally |
| HMI says “open,” flow remains zero | Failed device, false feedback, isolation, or bad meter | Confirm permissives and field condition under SOP |
| Several unrelated tags disappear together | Network, controller, or power problem | Notify control staff and use approved fallback monitoring |
| One tag steps after maintenance | Scaling, configuration, or instrument problem | Compare before/after records and request instrumentation review |
Operator boundary and cyber awareness
A Class I operator may select approved modes, enter permitted setpoints, review trends, acknowledge alarms, and follow established manual-operation procedures. PLC program edits, alarm suppression, interlock bypasses, firmware work, and network configuration belong to specifically authorized personnel under change control. EPA and CISA warn that exposed HMIs can permit unauthorized viewing or process changes; unexpected screens, settings, users, or commands are reasons to preserve evidence and report immediately. Never defeat a safeguard merely to make an alarm disappear.
Worked decision scenario
The HMI shows finished-water flow suddenly fixed at zero, yet the clearwell level is falling slowly, pump current is normal, and a local flow indicator is nonzero. Those independent conditions do not support zero flow. The best Class I conclusion is unverified data-path failure, not immediate plant shutdown and not a blind chemical-feed adjustment. Check tag status and time, obtain the approved field reading, notify the appropriate control or instrumentation role, operate under the documented fallback, and record the discrepancy until the trustworthy signal is restored.
Official source trail
Which statement correctly distinguishes an HMI from a PLC in a treatment plant?
An HMI shows zero flow, but pump current is normal and a safe local indicator shows flow. What is the best first response?
What can happen if an analog instrument is rescaled but the receiving PLC input retains the old span?