Monte Carlo Tools for Assessing See Data Quality for Different Types of Analyses

Abstract

After 43 years of Single-Event Effects (SEE) heavy-ion testing, there exists a large body of archival SEE results, almost all of which are intended for estimating on-orbit SEE rates. However, SEE data vary dramatically over a broad range of data-quality metrics. While variable SEE data quality may not be a major source of errors for SEE rates, other uses of the data—e.g., parametric estimation—are less forgiving of low-quality data. We apply a Monte Carlo (MC) tool using a simplified Generalized Linear Model (GLM) SEE data-fitting routine to assess how data quality affects SEE rate bounding, estimation and determination of fit parameters for SEE cross section (σ) vs. LET. The MC tool can generate 1000 MC events in about 1 minute on a standard laptop, making it suitable for on-the-fly testing decisions as well as evaluation of data quality.