Systems Integration / HMI / Optimization

Human-Centered Designs: The Secret to Reducing Human Errors

Learn About Different Techniques That Have Avanced to Evaluate and Quantify the Probability of Human Errors Occurring in Operational Situations

By Gus Kellogg, Rahul Bobbili, Dianna Overmyer

Following major accidents such as the toxic release from the Union Carbide plant in Bhopal and the Piper Alpha oil platform fire and explosion, detailed investigations revealed that up to 80% of accidents may be attributed, at least in part, to the actions or omissions of people. During the past several decades a number of techniques have been advanced to evaluate and quantify the probability of human errors occurring in a host of operational situations.

(What Caused the Three Mile Island Accident?)

In 1986, J.C. Williams introduced the "human error assessment and reduction technique" (HEART). Among Williams' findings were that error rates among users:

  • Unfamiliar with a situation are 17x greater than for users familiar (trained) with the situation.
  • Required to navigation multiple screens to complete a task are 13x greater than completing that task on a single screen.
  • Using cluttered graphic displays with unnecessary colors, images and data-fields can be as much as 10 times higher.

One of the outgrowths of these many studies is human-centered design" (HCD) for process automation systems. Rooted in military and international standards, HCD is far more than making products easier to use; it's about making operators jobs easier to do, and that can only be achieved by fully understanding the environment in which operators perform their duties.

When evaluating an operator's environment, key HCD considerations include:

  • Graphic color themes - in general lighter display backgrounds are better than dark (unless the control room is dim, in that case, dark is better) and that the use of bright colors should be minimized.
  • Pattern recognition - using shapes or patterns to graphically view the data enables operators to easily scan process values without having to read and analyze each value. Using bar graphs to clearly define the relationship between measured values and setpoints, alarm limits, etc. simplifies analysis for operators, as does using obvious and consistent means of presenting device status (i.e., on/off, open/closed, abnormal, fault, etc.) information.
  • Contextual alarm information - permitting authorized/experienced personnel to embed knowledge (i.e., functional classification, recommended action, time to respond, probable cause, and consequence of inaction) into alarms in a consistent, easy to access manner.

Unlike other design philosophies that often force users to change their behavior to accommodate the implemented solution, human-centered designs strive to optimize the solution around how users can, want or need to work. The result is a more comprehensive understanding of the process resulting in less human errors.

Want to learn more? Read "Biodiesel Production Can Be Safe and Efficient."