Supporting data for A Novel Approach to Explore Detrimental Weak Emergent Behavior Considering Noise Parameters: a Case Study to Detect a Vessel in Distress
This dataset contains the input and output data from an industrial case study aiming to detect undesired non-intuitive behavior in an engineered complex system (an Autonomous Surface Vessel (ASV) on a Search and Rescue (SAR) mission). We used the Taguchi method to set up experiments, conducted the experiments in a case company specific test arena, and performed different forms of regression analysis.
The speadsheet consists of 10 sheets:
- Additional theory for DoE and regression
- Screening experiments
- Transition rationale from screening to investigation
- Investigation experiments for system setting 1
- Investigation experiments for system setting 2
- Plots for system setting 1
- Plots for system setting 2
- Fractional Factorial vs Taguchi study
- Additional trial and error experiments
- Dimensional analysis experiments
Article Abstract - Related Publication
This paper presents a comprehensive research on the application of the Design of Experiments approach to explore engineered complex systems to reveal potential detrimental weak emergent behavior. The proposed methodology utilizes orthogonal arrays in combination with regression analysis to systematically explore the parameter space of a specific system function. By screening and investigating system boundaries and potential pain points, this research introduces a novel use of orthogonal arrays to effectively detect and map areas within a system's parameter space that does not comply with defined functional acceptance criteria. The findings demonstrate that this approach enables a systematic exploration of engineered complex systems to reveal inherent detrimental weak emergent behavior, thereby enhancing test coverage, expanding system knowledge, and facilitating mitigation efforts.