Evaluation of two crew module boilerplate tests using newly developed calibration metrics

摘要

The paper discusses an application of multi-dimensional calibration metrics to evaluate pressure data from water drop tests of the Max Launch Abort System crew module boilerplate. Specifically, three metrics are discussed: (1) a metric to assess the probability of enveloping the measured data with the model, (2) a multi-dimensional orthogonality metric to assess model adequacy between test and analysis, and (3) a prediction error metric to conduct sensor placement to minimise pressure prediction errors. Data from similar (nearly repeated) capsule drop tests show significant variability in the measured pressure responses. When compared to expected variability using model predictions, it is demonstrated that the measured variability cannot be explained by the model under the current uncertainty assumptions.