Weibull Reliability Calculator
What Is a Weibull Reliability Table?
A Weibull Reliability Table is a precomputed reference used by reliability engineers to read off reliability, failure probability, and hazard rate values without computing the Weibull exponential formula each time. The table organizes values by shape parameter β and normalized time t/η — the ratio of actual time to characteristic life.
The underlying formula is the Weibull reliability function:
Where R(t) is the probability that a component survives to time t, β is the shape parameter, and η is the scale parameter (characteristic life). The table lets engineers look up this value for any combination of β and t/η without a calculator.
Key fact: At t = η (normalized time = 1.0), R(t) = exp(−1) ≈ 0.368 regardless of β. This means 63.2% of units fail by time η — which is why η is called the characteristic life. This fixed point makes η a reliable anchor for comparing different Weibull models.
Weibull Reliability Table R(t): Core Reference
Values below are R(t) — the probability of survival to normalized time t/η for each shape parameter β. Select the tab for the quantity you need. Click any cell to load values into the calculator.
R(t) = exp(−(t/η)β). F(t) = 1 − R(t). h(t)·η = β(t/η)β−1. f(t)·η = β(t/η)β−1·exp(−(t/η)β). Values computed to 4 decimal places. t/η = normalized time (mission time ÷ characteristic life η). Click any cell to highlight it.
Shape Parameter β: The Most Important Number in Weibull Analysis
β tells you which phase of the failure lifecycle your component is in. Every maintenance decision — whether to replace proactively or run to failure — depends on correctly identifying β from failure data.
Infant Mortality / Early Failures
Failure rate decreases over time. Defective units fail quickly; survivors become more reliable. Run-in testing and burn-in screening address this phase. Examples: manufacturing defects, assembly errors, poor materials.
Random Failures (Exponential)
Constant failure rate. Failures occur independently of age — the component is equally likely to fail at any time. Preventive replacement provides no benefit. Examples: electronic components under steady load, random overload events.
Wear-Out Failures
Failure rate increases with age. Scheduled maintenance and replacement before the wear-out threshold makes engineering sense. β ≈ 2 is common for bearings and seals. β ≈ 3.5 approximates a normal distribution.
Rapid Wear-Out / Fatigue
Very steep wear-out curve — most units fail in a narrow time window. Tight maintenance intervals are justified. Common in metal fatigue, corrosion, and high-cycle mechanical components.
Common mistake: Using a fixed maintenance interval when β < 1 actually increases total failures, because replacing components that have survived their burn-in phase removes the most reliable units still in service. Scheduled replacement only makes sense when β > 1.
How to Use the Weibull Reliability Table: Step by Step
Using the table requires only three inputs: your shape parameter β, your scale parameter η, and the mission time t. The procedure below walks through both the table lookup and the formula verification.
Step 1 — Estimate Weibull Parameters
Fit β and η to your failure data using a Weibull probability plot, maximum likelihood estimation (MLE), or least squares regression. Statistical software such as Minitab, ReliaSoft Weibull++, or Python's scipy.stats.weibull_min can automate this step. The Weibull probability plot gives β as the slope of the fit line.
Step 2 — Compute Normalized Time t/η
Divide your mission time t by η. This normalizes time so the table is unit-independent — it does not matter whether you measure time in hours, cycles, or miles.
Example: t = 3000 hours, η = 8000 hours → t/η = 0.375. Use 0.40 (nearest tabulated value).
Step 3 — Locate Your Row and Column
Find the row matching t/η and the column matching β. The cell gives R(t). If your β falls between table columns, linearly interpolate: R ≈ R₁ + (β − β₁)/(β₂ − β₁) × (R₂ − R₁).
Step 4 — Calculate F(t) and h(t)
Failure Probability: F(t) = 1 − R(t)
Hazard Rate: h(t) = (β/η)(t/η)β−1
Step 5 — Engineering Interpretation
R(t) = 0.90 means 90% of units survive to time t. F(t) = 0.10 means 10% fail by that time. If your required reliability is R(t) ≥ 0.90, set F(t) = 0.10 and solve for t: t = η × (−ln(0.90))^(1/β). This is the B10 life — the age by which 10% of units fail — a standard warranty and maintenance design parameter.
Mean Time to Failure (MTTF) from Weibull Parameters
MTTF is the average time until failure for a population of non-repairable components. For the Weibull distribution, it depends on both β and η:
Where Γ is the gamma function. The table below gives the Γ(1 + 1/β) multiplier for common β values so you only need to multiply by η.
| β | Γ(1 + 1/β) | MTTF / η | Engineering context |
|---|---|---|---|
| 0.5 | 2.0000 | 2.000 | Infant mortality; MTTF exceeds η |
| 1.0 | 1.0000 | 1.000 | Exponential; MTTF = η |
| 1.5 | 0.9027 | 0.9027 | Mild wear-out |
| 2.0 | 0.8862 | 0.8862 | Bearings, seals (Rayleigh) |
| 2.5 | 0.8873 | 0.8873 | Gears, structural fatigue |
| 3.0 | 0.8930 | 0.8930 | Similar to normal distribution |
| 3.5 | 0.8990 | 0.8990 | ≈ Normal distribution |
| 5.0 | 0.9182 | 0.9182 | Rapid wear-out, fatigue failures |
Γ values computed from the standard gamma function. MTTF = η × Γ(1 + 1/β). For repairable systems, MTBF (Mean Time Between Failures) is used instead of MTTF.
Worked Example: Bearing Reliability Analysis
Scenario: A maintenance team analyzes failure data from 50 industrial pump bearings. Weibull analysis gives β = 2.1 and η = 8,500 hours. The maintenance schedule is every 4,000 hours. The team needs to verify: (a) reliability at the maintenance interval, and (b) the B10 life.
Solution — Step by Step
| Step | Calculation | Result |
|---|---|---|
| 1 | Parameters from Weibull analysis | β = 2.1, η = 8,500 hrs |
| 2 | Normalized time at t = 4,000 hrs | t/η = 4000/8500 = 0.471 |
| 3 | Reliability at t = 4,000 hrs | R = exp(−0.4712.1) ≈ exp(−0.209) ≈ 0.811 |
| 4 | Failure probability at maintenance interval | F(4000) = 1 − 0.811 = 0.189 (18.9%) |
| 5 | B10 life (10% failure probability) | tB10 = η×(−ln(0.90))1/β = 8500×0.10540.476 ≈ 2,870 hrs |
| 6 | MTTF | MTTF = 8500 × Γ(1.476) ≈ 8500 × 0.887 ≈ 7,540 hrs |
Engineering Recommendation
At the 4,000-hour maintenance interval, 18.9% of bearings are already failing. Since β = 2.1 > 1, the failure rate increases with age, making scheduled replacement sound. However, the B10 life is approximately 2,870 hours. To guarantee less than 10% failures before replacement, the maintenance interval should be reduced to roughly 2,800 hours. This analysis follows the standard methodology described by O'Connor and Kleyner in Practical Reliability Engineering (Wiley, 2012) and is consistent with ISO 281 bearing life calculation standards.
B-Life Values: Maintenance Planning Reference
B-life (Bearing life notation, also Lx life) is the time at which x% of units will have failed. B10 means 10% fail by that time; B50 is the median life; B90 means 90% have failed. These are computed as tBx = η × (−ln(1 − x/100))^(1/β) and are the standard basis for maintenance interval design.
| B-Life | Failure% | β = 1.0 (t/η) | β = 2.0 (t/η) | β = 3.5 (t/η) | Use case |
|---|
tBx/η = (−ln(1 − x/100))1/β. Multiply by η to get actual time. B10 is the standard for rolling element bearing replacement intervals per ISO 281.
The Bathtub Curve and Weibull Analysis
The bathtub curve describes the three phases of product failure over a lifecycle. Weibull analysis models each phase with a distinct β value. Understanding which phase a component is in determines the correct maintenance strategy.
Phase 1: Infant Mortality
β < 1. Failure rate is high at first, then drops as weak units fail out. Burn-in testing, quality screening, and extended warranty inspection address this region. Maintenance replacement makes failures worse in this phase.
Phase 2: Useful Life
β ≈ 1. Constant, low failure rate. Failures are random and independent of component age. Corrective maintenance (repair after failure) is more cost-effective than preventive replacement during this phase.
Phase 3: Wear-Out
β > 1. Failure rate climbs steeply with age. Scheduled preventive maintenance before the B10 or B20 life is justified and reduces total maintenance costs. Most mechanical components eventually reach this phase.
Practical note: Real systems often have a mixed-Weibull population — some units fail early (β < 1) while others show wear-out (β > 1). A Weibull probability plot that shows a curved fit rather than a straight line suggests a mixed-mode failure population requiring separate analysis for each failure mode.
Weibull Reliability Formula Glossary
Every symbol in Weibull reliability analysis, with formula, interpretation, and the most common misuse.
| Symbol | Name | Formula | Engineering meaning | Common error |
|---|---|---|---|---|
| R(t) | Reliability function | exp(−(t/η)β) | P(component survives to time t) | Confusing with MTTF |
| F(t) | Failure probability (CDF) | 1 − R(t) | P(failure by time t) | Using F when R is needed |
| f(t) | PDF (failure density) | (β/η)(t/η)β−1·R(t) | Rate of new failures per unit time | Confusing with h(t) |
| h(t) | Hazard rate (failure rate) | (β/η)(t/η)β−1 | Instantaneous failure rate for survivors | Treating as failure count |
| β | Shape parameter | — | Controls failure rate trend over time | Assuming β = 1 without data |
| η | Scale parameter (char. life) | — | Time at which 63.2% have failed | Treating η as MTTF (only equal when β=1) |
| MTTF | Mean Time to Failure | η × Γ(1 + 1/β) | Average life of non-repairable units | Confusing with MTBF for repairable |
| Bx | B-life (percentile life) | η·(−ln(1−x/100))1/β | Time at which x% have failed | Confusing B10 with MTTF |
Engineering Case Studies: Weibull Analysis in Practice
The following examples show how reliability engineers apply the Weibull table across different industries. Each uses real-world typical parameter values.
Wind Turbine Gearbox — Predictive Maintenance
Failure data from 200 offshore wind turbine gearboxes yields β = 2.4, η = 55,000 hours. At the standard inspection interval of 20,000 hours: t/η = 0.364, R(20000) ≈ 0.876. About 12.4% of gearboxes need replacement before 20,000 hours. The B10 life is roughly 14,200 hours, suggesting the inspection interval is already close to the 10% threshold. Operators use this to justify oil sampling and vibration monitoring at 12,000 hours to catch early-stage failures.
Automotive Transmission Seals — Warranty Analysis
A transmission seal population has β = 1.8 and η = 180,000 km. The standard warranty is 100,000 km. t/η = 0.556, R(100,000) = exp(−0.5561.8) ≈ 0.748. About 25.2% of seals are expected to develop leaks before the warranty ends. Because β > 1 confirms wear-out behavior, material compound changes and lip design improvements were tested to raise η and reduce warranty claims. This analysis method is standard in automotive FMEA per SAE J1739.
Semiconductor Memory — Accelerated Life Testing
Electronic components often show β < 1 during early life, reflecting manufacturing variation in oxide quality. A memory chip batch shows β = 0.6, η = 50,000 hours at 85°C. Burn-in testing at high temperature for 168 hours (one week) removes the highest-risk units before they reach customers. The Arrhenius acceleration factor translates burn-in stress conditions to field conditions. After burn-in, the surviving population has effectively higher reliability because the weakest units have been screened out.
Aircraft Hydraulic Actuators — Certification Testing
Aerospace reliability requirements often specify R ≥ 0.9999 per flight hour for critical actuators. With β = 3.2 and η = 50,000 flight cycles, R(1 cycle) ≈ 1 − (1/50000)3.2 ≈ essentially 1.0 per individual cycle. Weibull analysis over the design life of 30,000 cycles gives R(30,000) = exp(−(0.6)3.2) ≈ 0.779, which falls below the 0.999 design target, driving the redesign cycle. FAA AC 25.1309 provides the regulatory framework for this analysis.
Frequently Asked Questions
Is η the same as MTTF?
No. η (characteristic life) equals MTTF only when β = 1 (exponential distribution). For all other β values, MTTF = η × Γ(1 + 1/β). For β = 2, MTTF ≈ 0.886η. Treating η as MTTF when β ≠ 1 overstates or understates the average life, which leads to incorrect maintenance intervals and warranty cost projections.
Can the Weibull distribution fit any failure dataset?
The two-parameter Weibull model fits a wide range of failure patterns, but it assumes all units eventually fail due to the same failure mode. If multiple independent failure modes are present, the data may show curvature on a Weibull probability plot, suggesting a competing risks model or mixed-Weibull analysis. The three-parameter Weibull adds a location parameter γ for datasets where no failures occur before time γ, such as minimum safe life calculations in aerospace.
How many failures do you need for a reliable Weibull analysis?
There is no fixed minimum, but statistical precision improves substantially with sample size. With fewer than 5 failures, confidence intervals on β are very wide. Practical guidance from the ReliaSoft ReliWiki suggests at least 10–20 failure observations for reasonable parameter estimates. Suspended (censored) observations — units still running at the analysis date — are also informative and should be included using rank regression or MLE methods.
What is the difference between MTTF and MTBF?
MTTF (Mean Time to Failure) applies to non-repairable components — bearings, seals, bulbs. The unit fails and is replaced. MTBF (Mean Time Between Failures) applies to repairable systems — engines, pumps, aircraft — where the unit is repaired and returned to service. For non-repairable components following a Weibull distribution, MTTF = η × Γ(1 + 1/β). For a repairable system with Weibull failure modes, computing effective MTBF requires additional system-level modeling. The NIST Engineering Statistics Handbook covers both cases.
How does Weibull analysis relate to Six Sigma?
Weibull analysis is a standard tool in Design for Six Sigma (DFSS) and reliability engineering phases of Six Sigma projects. It appears in the Design and Verify phases of the DFSS DMADV methodology, where teams must demonstrate that new designs meet reliability targets. The ASQ Body of Knowledge for Certified Reliability Engineers (CRE) includes Weibull analysis as a core competency. Understanding how to read a Weibull probability plot and interpret β and η is part of the standard CRE exam curriculum.
References and Further Reading
O'Connor, P. D. T. and Kleyner, A. (2012). Practical Reliability Engineering (5th ed.). John Wiley & Sons. — The standard textbook reference for Weibull analysis in mechanical engineering contexts, covering parameter estimation, confidence bounds, and maintenance applications.
Abernethy, R. B. (2006). The New Weibull Handbook (5th ed.). Self-published. barringer1.com — The practitioner's definitive reference for Weibull data analysis, written by the engineer who led the development of many aerospace Weibull standards.
NIST/SEMATECH (2012). e-Handbook of Statistical Methods: Section 8.1.6 — Weibull Distribution. National Institute of Standards and Technology. itl.nist.gov — U.S. government statistical reference covering the Weibull distribution, parameter estimation, and reliability applications.
ReliaSoft (2015). Life Data Analysis Reference (Weibull++). ReliaSoft Corporation. reliawiki.com — Open-access technical reference covering all aspects of Weibull distribution fitting, including median ranks, MLE, and reliability plot interpretation.
ASQ Reliability Division (2020). Body of Knowledge for the Certified Reliability Engineer (CRE) Exam. American Society for Quality. asq.org — The official exam content outline covering Weibull analysis requirements for ASQ reliability engineering certification.
Related Statistical Tables and Resources
Understanding What the Weibull Table Tells You
Why R(η) = 0.368 for All β
When t = η, the exponent in R(t) = exp(−(t/η)β) becomes exp(−1β) = exp(−1) ≈ 0.368, regardless of β. This mathematical property defines η as the characteristic life — the time by which 63.2% of units have failed for any Weibull shape. Engineers use η as a standardized comparison point when evaluating different designs or materials.
Why Higher β Means Steeper Reliability Decay
As β increases, the Weibull PDF becomes taller and narrower — failures concentrate around a specific age. A β = 5 component has very high reliability for most of its life, then fails rapidly in a short window. This is why fatigue-dominated components seem to last a long time but then fail suddenly. The table reflects this: high-β columns show R close to 1.0 at low t/η, then drop sharply.
Relationship Between Weibull and Other Distributions
The Weibull distribution includes three well-known distributions as special cases. At β = 1, it reduces to the exponential distribution, used in electronic component reliability. At β = 2, it becomes the Rayleigh distribution, used in radar signal analysis. At β ≈ 3.44, it closely approximates the normal distribution, which explains why some mechanical component failures that engineers historically modeled as normally distributed can be equally well described with a Weibull model.