Information-Rich Design: A Concept for Large-Screen Display Graphics: Design Principles and Graphic Elements for Real-World Complex Processes

Abstract

The objective in this thesis research is to mitigate two problems, which are typically experienced by control room operators monitoring large-scale processes in centralized control rooms: i) How to design for rapid perception of industrial-scale data sets? ii) How to avoid keyhole effects in complex processes? In this thesis, these problems are approached through research into Large- Screen Display (LSD) design; the contribution is a concept named Information-Rich Design (IRD). The concept is not domain specific, and it is useable typically for nuclear and petroleum industries. IRD can be used as a starting point for user-centred design, as opposed to approaching the problem from the technology end first. The thesis research is based on a broad perspective, through interaction design research methods: design exploration, design studies and design practice. Design exploration was done on a small-scale early in the research process, and later through three complete LSD applications. The first two LSDs were implemented on full-scale nuclear simulators, and the most recent was implemented for an operational nuclear research reactor. Crews of certified control room operators have provided feedback for design in an iterative research process. Design studies were based on findings from basic, applied and clinical research: (1) human capabilities and characteristics, (2) principles for information visualization, (3) findings from human-computer interaction and (4) research from other related display concepts. Design practice from applying IRD commercially in Norwegian petroleum industry was fed back into the concept. The thesis research suggests that LSDs should be designed from the ground-up, acting as a stable frame of reference for process monitoring, leaving details for desktop workstations. Research found that larger displays should support bottom-up data driven processes by presenting process data as simple visual patterns, suitable for rapid visual perception. Further, LSDs should support operators in top-down search for information, and aim to avoid keyhole effects through externalized graphics, which do not load limited visual memory resources. Graphics should reduce visual complexity by creating visual hierarchies, giving critical information the most prominent visual salience, while avoiding masking primary data from less important information. Based on this, the contribution for LSD designs, are design principles and accompanying graphics. The IRD concept is theoretically validated, and externally validated through industrial applications and user tests. With a few concerns for inconsistency from using mathematical normalized scales in graphics, and readability of the grey-layered colours, the concept is generally found to be a reasonable approach on LSD design and the two research problems. The research contribution is not radical or revolutionary; rather it extends what others have found for computer graphics for smaller displays. It is positioned as applied research for LSD design, as a contribution to humancomputer interaction. The innovative part of this thesis contribution is design-patented graphics for information presentation. Further work should focus on providing more quantitative performance data, and on performing comparisons with other display concepts, particularly measuring Situation Awareness levels. Secondly, one should look at the question of consistency with other displays in the same setting. A natural extension of this thesis work would be to look at direct process interaction through LSDs.