Reliability analytics

WeibullFlowBeta

Enter failure times and optional right-censored suspension times, or upload a CSV. We fit a two-parameter Weibull by maximum likelihood (MLE), with an optional three-parameter model (threshold γ) when a minimum-life offset may matter. Charts include the probability plot (asymptotic pointwise 95% band around the MLE line for 2P fits, plus an optional median-rank regression line when there are no suspensions), the fitted PDF, and a failure-time histogram, with alternative distributions ranked for at-a-glance comparison. Goodness-of-fit reports Anderson–Darling A² and Kolmogorov–Smirnov D vs the fitted Weibull CDF on complete failure data with (k, λ) [and γ when 3P] treated as known constants (not a Stephens/Lilliefors adjusted test — no p-value from those lines alone). With suspensions, Cox–Snell residuals from the MLE are compared to exponential(1) on the residual scale (including an AD-style integral that is not the standard uncensored A² table statistic). For 2P fits, parameter, B-life, and R(t) 95% intervals are asymptotic (Fisher information plus the delta method); 3P fits report point estimates and AIC/AICc without those Wald intervals. Optional AI commentary turns the fit into a structured reliability brief—plot diagnostics, density and histogram context, B-life takeaways, caveats, and practical next steps for engineers and stakeholders.