I hate to date myself, but I remember the first color monitors coming into process automation operator workstations in the 1980s when I was a young systems engineer. Having a palette of colors was really nice when putting together displays for our operators on offshore platforms. Unfortunately, we didn’t have a clue about color science, nor did the industry in subsequent decades.
I bring this up because great strides have been made in recent years combining the expertise of color science with automation systems, particularly the human machine interfaces into these systems. There is a great whitepaper, Applying Color Science to Design Effective Human-Machine Interfaces (login required), written by Dirk Beer, Harvey Smallman, along with Emerson’s Cindy Scott and Mark Nixon.
Here’s the abstract:
Human operators are a key part of any process control system. As such, they constitute part of a complex, causal chain of overall system processing. Human machine interfaces (HMIs) form a key link in that chain by bridging the physical world where processes reside with the perceptual reconstruction and representation of those processes in the heads of human operators and supervisors.
If an HMI design gives rise to a flawed or inaccurate representation of a process, then error and suboptimal task performance may result. HMIs have become increasingly important links in this chain for two reasons. First, the arrival of distributed control systems (DCS) in the 1970s distanced operators from the physical entities they controlled, requiring all interaction be mediated by HMIs. Second, the ongoing introduction of complex automation into process control is increasingly changing human operators into supervisors. Supervision has complex decision-making requirements that must all be conveyed via HMIs.
The authors note that typically people don’t think about the operators as being part of the control system, but in fact through the HMI:
…form a key link in that chain by bridging the physical world where processes reside with the perceptual reconstruction and representation of those processes in the heads of human operators and supervisors.
Like my attempts to make bright color displays for the operators, the downside of indiscriminate color use misaligns to the tasks required, hides important alerts, and may mean different things in different contexts.
There are three specific functional roles to the use of color to support tasks—directing attention, aiding identification, and segmenting the visual scene. The authors emphasize:
…color should be used if, and only if, task performance can be improved by one of these functional roles.
Dirk Beer and Harvey Smallman worked with the U.S. Navy to develop and deploy a “Universal Color Requirements” for display interfaces. These guidelines consisted of 6 requirements:
- Colors must be identifiable and discriminable for different users, including colorblind users
- Colors must have sufficient contrast for legibility and visibility
- Color must remain stable in appearance across HMI layouts and lighting conditions
- There should be as few colors in the HMI as possible
- A color should have the same meaning everywhere, so if red is chosen as an alarm color, it should not also be used to represent equipment that is not running
- Color mapping to data attributes should be cognitively appropriate, task-relevant, and culturally appropriate
For process automation suppliers and project engineers, a systematic, evidence-based process should be taken on the use of color in the HMI. The process includes three steps—analyze operator task requirements, model descriminability and appearance, revise HMI design.
This task-based approach is very much like the human centered design approach we’ve highlighted in many posts on this blog. I’ll leave the specifics of these three steps and several examples shared by the authors of putting this process into practice.
The whitepaper is definitely worth taking the time to download and read. It should provide you with a good understanding of the importance of applying color science to the operators’ connection to the plant’s process.