A validation metric is the means by which&nbsp; the discrepancy between simulation results and experimental results is quantified.&nbsp;&nbsp; Ideally, this measurement is done between stochastic distributions that represent the uncertainty in the simulation results and the experimental results.&nbsp;&nbsp; The area metric (also known as the Minkowski L1 metric) is one example of a validation metric that has certain desirable features, such as the ability to function in the presence of epistemic uncertainty.&nbsp;&nbsp; Like all metrics, it also has limitations.&nbsp;&nbsp; One example is the possibility that the metric can decrease as measurement system uncertainty increases in some cases.&nbsp;&nbsp; This presentation focuses on a small improvement to the classic method in order to address some of its limitations.In the classic area metric approach, the Cumulative Density Function (CDF) of the experimental data represents the uncertainty due to both the measurement system, as well as that due to specimen variability.&nbsp;&nbsp; Similarly, the simulation CDF includes both numerical uncertainty as well as that due to model input uncertainty.&nbsp;&nbsp; We will discuss that measurement system and numerical uncertainty are of a different class (or type) of uncertainty compared to specimen variability and input uncertainty.&nbsp;&nbsp; The former two are always detrimental to a validation effort while the latter are better thought of as a characteristic of the system being studied.&nbsp;&nbsp;&nbsp; We will show that by treating numerical uncertainty and measurement system uncertainty as epistemic intervals, the behavior of the area metric is improved.&nbsp; For example, it eliminates the possibility that the area metric result can be improved by using a less precise measurement system.&nbsp;In our improved scheme, the measurement of the distance between the experimental and simulation results becomes a matter of calculating the distance between two probability boxes.&nbsp;&nbsp; This leads to two values for the area metric, which can be interpreted as an interval that is believed to contain the true model form uncertainty.&nbsp;&nbsp; We believe that this small adjustment to the classic framework is particularly important for models of populations that have large specimen-specimen variability, which is frequently the case in models of biomedical systems.&nbsp;&nbsp;

Not All Uncertainties Are Created Equal:  Re-Visiting the Area Metric

Paper Proceedings: https://asmedigitalcollection.asme.org/VVS/VVS2021/volume/84782

ASME 2021 Verification and Validation (VVS2021)

Supporting the need among researchers worldwide for additional online and multimedia content, ASME now offers the ASME Conference Video Collection. This platform will collect and provide long-term access to conference presentation videos, encompassing all technology areas within ASME's technical conference and events program. About 30,000-40,000 engineers across all mechanical engineering and related disciplines participate in ASME conferences each year. This includes authors of approximately 6,000 to 8,000 technical papers published by ASME, the majority of whom also submit video presentations for our virtual conferences. The video collection will be available as an add-on to existing product packages or as a separate subscription.