Read remaining life as a range

Distinguish a Weibull RUL point estimate from its 90% interval, operating-condition factor, and hazard rate.

Question for this chapter

Is EQ-005's RUL of 1.9813 years a fixed replacement date?

Why this matters now

Remaining Useful Life (RUL) applies a lifetime pattern learned from failure history to a current asset. New data can change the result, so read a representative estimate together with its plausible range.

Try it

Open the reliability_analysis pipeline and select Run. It fits a Weibull lifetime distribution from failure history and then adjusts each asset with its load and temperature conditions. The run writes ten rows to rul_estimates.

Sort by rul_estimate_years in ascending order and inspect the first three rows.

Successful result

Asset	RUL estimate	90% interval	Hazard rate	Condition factor
`EQ-008`	1.9643 years	0–3.9116 years	0.086639	0.7756
`EQ-005`	1.9813 years	0–3.6664 years	0.082668	0.7045
`EQ-003`	2.8840 years	0.2915–4.8540 years	0.072016	0.7984

Two assets, EQ-008 and EQ-005, have RUL below two years.

Read the result

Point estimate of 1.9813 years: a representative value under the current model and inputs, not a countdown to a guaranteed failure.
90% interval of 0–3.6664 years: uncertainty created by resampling the observed failure times. The lower bound reaching zero is a reason for early verification, not proof of immediate failure.
Condition factor of 0.7045: a multiplier derived from average load and temperature relative to reference conditions. Its allowed range is 0.5–1.5; below one reduces the baseline estimate here.
Hazard rate of 0.082668: the instantaneous failure tendency at the current age. It does not mean an 8.2668% probability of failure within a specific period. This scenario uses it for relative ranking.

Deep dive — what do Weibull and bootstrap contribute?

The Weibull model fits a lifetime curve using both observed failures and the operating periods of assets that had not failed. Bootstrap repeatedly resamples the observed failure times and reruns the fit to show how much the estimate moves. This pipeline uses 100 repetitions, a 90% confidence level, and random seed 42.

Next decision

EQ-005 has grade B HI, yet its RUL range calls for early verification. Next, inspect the subsystems, sensors, and inspection records that provide its field context.