Which HMI Permissions Should Operators Have?
Give HMI operators the permissions needed to run their assigned equipment, including approved setpoint changes within agreed limits. Give maintenance staff access to defined diagnostic and service functions, and reserve account management and access configuration for administrators. Decide each permission by the work required and the consequence of a mistake; no role table or session timeout is universally secure.
How should you decide who can do what?
Start with actions, not job titles. “Operator access” is too broad to specify or test. List individual actions such as starting a conveyor, selecting a recipe, changing a temperature target, resetting a fault and editing an alarm limit.
For each action, agree:
- Who needs it during normal work?
- Which equipment and operating states does it apply to?
- What could happen if it is used incorrectly?
- What limits, authorisation and records should accompany it?
Separate viewing from changing. Someone who needs to see a drive parameter does not automatically need permission to edit it. Separate equipment scope too: permission to operate one production area should not automatically cover every connected line.
Record these decisions in the functional design specification, alongside operating modes, limits and acceptance tests.
What should operator, maintenance and administrator roles include?
Use the following as an illustrative starting point, then review it against actual duties and equipment risks. “Restricted” means a separately defined permission with conditions, not a hidden button that everybody knows how to reveal.
| Action | Operator | Maintenance | Administrator |
|---|---|---|---|
| View assigned equipment status and alarms | Allow | Allow for assigned work | Only where duties require |
| Start and stop normal production | Allow within assigned scope | Only where service work requires | No automatic entitlement |
| Change production setpoints | Allow within approved limits | Only for assigned tasks | No automatic entitlement |
| Acknowledge alarms | Allow for assigned area | Allow for assigned work | Only where duties require |
| Run manual movements or service tests | Restricted by task | Restricted by task and equipment state | No automatic entitlement |
| Edit alarm limits or equipment parameters | Restricted | Restricted to authorised parameters | No automatic entitlement |
| Create accounts and assign roles | Deny | Deny unless separately appointed | Allow within an agreed approval process |
| Change access rules or log settings | Deny | Deny unless separately appointed | Restricted and recorded |
Treat administration as a separate responsibility. An administrator who creates accounts need not have authority to alter a production recipe.
Where one person performs several duties, assign the necessary permissions explicitly. Consider a separate privileged account for administration so that routine operation does not require elevated access.
Which setpoint changes should operators be allowed to make?
Distinguish a production target from a limit that defines the permitted operating range.
An operator may need to adjust a target during normal production. Changing the range itself is a different decision. Specify both permissions separately, including whether they depend on the selected product, recipe or operating mode.
For every editable setpoint, define:
- The parameter name, unit and equipment affected.
- The permitted range and who approves it.
- Whether changes are allowed while equipment is running.
- Whether the value applies immediately or at a defined transition.
- Whether a reason or additional authorisation is required.
- What information is recorded.
Illustrative example: an operator can adjust a conveyor speed target within the approved range for the selected product. Changing the drive's maximum speed requires a separate engineering permission and an agreed maintenance procedure.
Do not copy that arrangement to another machine without reviewing its consequences. Specify where the limits must be enforced and test attempted changes through every available write path, including connected systems.
For safety-related settings or functions, involve a qualified assessor in defining the design and validation. Do not use the HMI role matrix as a substitute for that assessment.
Which maintenance functions need separate permission?
Avoid one broad “maintenance mode” permission that unlocks every service action.
Break maintenance access into functions: viewing diagnostics, jogging equipment, testing an output, changing a calibration value, editing a drive parameter and restoring a configuration. Decide the conditions for each.
For manual movement, specify the required equipment state, local controls and handover procedure. For parameter changes, identify the authorised parameters, approved values and restoration process.
Treat overrides and bypasses as explicit exceptions. Define who may authorise them, how their active state is shown, what record is required and how normal operation is restored. Review safety-related exceptions with a qualified assessor.
Also distinguish acknowledging an alarm from resetting a fault or changing the alarm configuration. Agree the meaning of each action rather than placing them behind one permission. Our guide to why operators ignore alarms provides a starting point for reviewing the wider alarm workflow.
How should access work on shared terminals?
Design the terminal around shared equipment while keeping privileged actions attributable to an individual.
Prefer individual sign-in for actions that require accountability. Make the active user's identity and role visible, provide an obvious sign-out control and include session handover in the operating procedure.
Avoid a shared maintenance password as the normal route to service functions. If a shared account is unavoidable, document which actions it permits and how individual authorisation will be recorded. Treat that arrangement as a limitation requiring review.
Check practical use with gloves, protective equipment and shift changes. A sign-in process that is awkward at the terminal needs redesign, not an informal instruction to leave a privileged session open.
Agree what an unattended screen should display. Where appropriate, retain a status view while requiring sign-in before a protected action.
How should session expiry work?
Choose session expiry around operating duties, unattended access and the consequence of losing a session mid-task. Do not adopt a timeout simply because it is the default.
Specify separate behaviour for ordinary operation and elevated maintenance access where appropriate. Require fresh authentication for selected high-consequence actions if the task assessment supports it.
Define what happens when a session expires:
- Does the screen return to a status view?
- What happens to unsaved edits and pending confirmations?
- Does an active maintenance operation continue, stop or require a controlled handover?
- How does the next user establish their own session?
Do not make logout an improvised equipment-stop function. Define machine behaviour separately and test expiry during representative tasks, including a setpoint edit and a service operation.
What should traceable access record?
For significant changes, request a record that answers who acted, what changed, where it changed and when.
For a setpoint edit, specify the user identity, equipment, parameter, previous value, new value, unit and timestamp. Add a reason where the consequence justifies it. Distinguish attempted, rejected and completed actions.
Include account creation, role changes, privileged access and relevant configuration changes in the review scope. Assign responsibility for reviewing records and define retention, retrieval and protection requirements.
Agree what should happen if records cannot be written or identity services become unavailable. Make any permitted fallback explicit, including its limits and recovery procedure. Test the actual records produced rather than accepting that a logging option is enabled.
What common mistakes should you avoid?
Granting access by seniority. Assign permissions to duties. A senior manager may need visibility without production write access.
Giving maintenance unrestricted control. Authorise the service functions required for the task and define their operating conditions.
Treating a hidden control as the permission check. Include direct navigation and alternative write paths in acceptance testing.
Leaving elevated sessions open for convenience. Review sign-in and handover usability before accepting persistent privileged access.
Approving roles without testing denied actions. Verify both what each role can do and what it must be prevented from doing.
Before handover, test with representative users and record the results. Review permissions again when duties, equipment or integrations change.
Frequently asked questions
Should operators have administrator access?
Do not grant it simply to resolve everyday operating problems. Identify the blocked task and assign its specific permission where justified. Keep account administration separately authorised.
Can maintenance staff change every setpoint?
Only where their assigned work requires it. Define the parameters, equipment scope, operating conditions and approval needed. Maintenance responsibility alone is not a complete access specification.
Is there one correct session timeout?
No. Choose and test expiry behaviour against the tasks, terminal location and consequences of an unattended session. Document the reasoning and any exceptions.
Where should a plant start?
List the actions available on one representative HMI, map them to duties and review the consequences with operations and maintenance. Our automation services cover control-system project work; contact us to discuss an access review.