The short answer
Trends on the HMI
A trend on an HMI shows how a value has changed over time, and it is the tool operators use to see where the process is heading rather than where it is. Good trend design puts small embedded trends on the process displays for the values that matter, provides pre-built trend groups that put related values together on sensible axes, and offers an ad hoc trend tool for investigation. Time spans, scales, and colors are chosen for the question each trend answers, and every trend is fed by the historian, not by the display.
Key points
- An embedded trend on the display shows direction. A number shows only position.
- Group related values: the level with the pump speed, the pressure with the flow, the residual with the dose.
- Fixed scales on operating trends, autoscale on investigation trends. Say which on the trend.
- Time span follows the process: minutes for pressure, hours for a wet well, days for a tank.
- Trends read from the historian. A display that trends from its own buffer loses the past when it closes.
Two kinds of trend
An operating trend lives on a process display or a faceplate and answers one question continuously: is this value where it should be, and which way is it moving. It is small, its scale is fixed to the normal operating range, and its time span is short enough that the recent movement fills it. An investigation trend is opened to find out what happened: several values, a longer span, adjustable scales, cursors, and the ability to scroll back through history. Both are needed, and confusing them produces either an operating display cluttered with full trend windows, or a trend tool that shows so much that nothing can be read.
Embedded trends
ISA-101 and the high-performance HMI practice put small trends directly on the process display for the handful of values that decide how the process is doing: the wet well level, the clearwell level, the discharge pressure, the chlorine residual. A number tells the operator where the value is; the sparkline beside it tells them where it has been for the last hour and whether it is climbing. That is the difference between seeing a problem forming and seeing an alarm.
- Fixed vertical scale set to the normal operating range, with the alarm limits shown as lines, so that the shape of the trace means the same thing every time.
- A fixed time span, typically 30 minutes to a few hours depending on the process, that fills the width with recent history.
- One value per embedded trend, or two closely related ones in distinct line weights. Not four.
- Gray or muted trace colors consistent with the display palette; the alarm limit lines in the alarm colors only.
- The current value as a number beside the trend, with units.
- A click that opens the full investigation trend for that value and its group.
Trend groups
Values are understood in relation to each other. A level trend with the pump run status and speed overlaid shows why the level did what it did. A pressure trend alone shows a dip; with flow, it shows whether the dip was demand or a pump. Pre-configured trend groups, built by someone who knows the process and opened with one click from the display or the faceplate, are the most used trend feature in a plant, and they are worth the time to design.
| Group | Values | Typical span |
|---|---|---|
| Lift station | Wet well level; pump 1 and pump 2 run status as bands; pump speed; motor current | 4 hours |
| Distribution pressure | Zone pressure; setpoint; pump speed; station flow; tank level | 8 hours |
| Disinfection | Chlorine residual at the point of entry; dose setpoint; feeder output; plant flow; pH | 24 hours |
| Filter | Effluent turbidity; head loss; flow; run time since backwash | 72 hours |
| Tank | Level; inflow; outflow; time of day | 7 days |
| Communications | Poll success rate per site; latency; retries | 24 hours |
Axes, scales, and spans
- Vertical scale
- Fixed for operating trends so the shape is comparable; autoscale for investigation with the scale visible. Values with different units get their own axes, and no more than two axes on one chart before it becomes unreadable.
- Digital states
- Run status, valve position, mode: shown as bands or a stepped trace at the bottom, not as a line that pretends to be analog.
- Time span
- Chosen from how fast the value changes and how far back the operator needs to see. Pressure and flow in minutes to hours; wet well level in hours; tank level and residual in a day; filter runs in days.
- Sampling
- The trend reads the historian at whatever the historian stored. A trend that appears smooth at 5-second data and stepped at 1-minute data is showing the historian configuration, not the process.
- Time axis
- Absolute time with the date where the span crosses midnight; a live mode that scrolls and a frozen mode that does not.
- Cursors
- Two cursors on the investigation trend, reading the value of every trace at each and the difference between them. This is how a drawdown rate or a dose response time is read.
Trends and the historian
Every trend should read from the historian, so that the same data is available on every client, so that history is available back as far as the historian keeps it, and so that a trend opened on a client that was just started shows the past. A display that trends from its own in-memory buffer shows a blank chart on open and forgets everything on close. The historian configuration, its collection rates and its compression, decides what a trend can show; the trending design and the historian design are the same exercise.
Common mistakes
- Autoscaled operating trends, so that a tiny fluctuation fills the chart and looks like an event.
- Rainbow traces. Five bright colors on one chart, none of them meaning anything.
- Trend spans set for commissioning and never changed: a 5-minute window on a value that moves over hours.
- No trend groups, so every investigation starts with building a trend from scratch and the operator does not bother.
- Digital states plotted as analog lines from 0 to 1 on the same axis as a level in feet.
- A historian collecting at one minute for a pressure loop that oscillates at 20 seconds, so the trend shows aliasing instead of the oscillation.
Frequently asked questions
- How many embedded trends should a display have?
- The high-performance HMI practice suggests a handful on an overview display, for the values that define the state of the process. More than six and the display is a trend page; fewer and the operator has no sense of direction. Area displays can carry more, one per key loop.
- Should trend colors match the alarm colors?
- No. Traces are in neutral colors; the alarm colors are reserved for alarm limit lines and for alarm indication. A red trace on a normal value teaches operators that red does not mean alarm.
- What sampling rate does the historian need for trends?
- Fast enough to show the fastest behavior the operator needs to see. Pressure and flow loops with a drive commonly need one to five second data; levels and residuals are fine at 10 to 60 seconds. Exception or deadband collection stores the changes rather than every sample, and the trend interpolates between them.
- Can operators build their own trends?
- They should be able to, in an ad hoc trend tool, and the useful ones should be promoted to trend groups by whoever maintains the HMI. The ad hoc tool is where investigation happens; the groups are the ones everyone uses.
Related topics
- High Performance HMI DesignGrey backgrounds, reserved color, and displays designed to make abnormal conditions obvious rather than to look impressive in a demo.
- ISA-101: Human Machine InterfacesWhat the ISA-101 standard for human machine interfaces covers, its lifecycle from philosophy to maintenance, the display hierarchy, and what adopting it changes on a project.
- HistoriansWhat a process historian does that a database does not, how data gets into it, how compression and retrieval work, how to size and keep it, and the settings that quietly throw data away.
- HMI FaceplatesThe pop-up that controls one device: what a pump, valve, or loop faceplate must show, the mode and command layout, interlock and permissive display, security by role, and building faceplates from a standard object so every device behaves the same.
- Values Frozen On ScreenThe HMI shows numbers that are not changing while the plant runs. Whether the controller, the communication driver, the tag server, or the display stopped, and the order to check them in.
- HMI NavigationHow operators move through the displays: the display hierarchy from overview to detail, a fixed navigation bar with area alarm indication, one-click paths from an alarm to the device, the rules that keep any display within two clicks, and what to leave out.
Direct contact
Have a controls question?
Reach Eric Sullivan directly about anything on this site, a controls or automation topic, or one of his personal projects.