Mathwords logoMathwords

Failure Rate — Definition, Formula & Example

Failure rate is how quickly items that are still working tend to fail, measured in failures per unit of time, such as failures per hour. A constant failure rate means an old part is no more likely to fail in the next hour than a new one.

For a random lifetime TT with density f(t)f(t) and cumulative distribution F(t)F(t), the failure rate (hazard rate) is h(t)=lim⁡Δt→0+P(t≤T<t+Δt∣T≥t)Δt=f(t)1−F(t)h(t) = \lim_{\Delta t \to 0^+} \frac{P(t \le T < t + \Delta t \mid T \ge t)}{\Delta t} = \frac{f(t)}{1 - F(t)}, defined where F(t)<1F(t) < 1. When h(t)=λh(t) = \lambda is constant, TT follows an exponential distribution, and the mean time between failures is 1/λ1/\lambda.

Key Formula

h(t)=f(t)R(t)=f(t)1−F(t),constant case: λ=number of failurestotal operating timeh(t) = \frac{f(t)}{R(t)} = \frac{f(t)}{1 - F(t)}, \qquad \text{constant case: } \lambda = \frac{\text{number of failures}}{\text{total operating time}}
Where:
  • h(t)h(t) = Failure rate (hazard rate) at time t, in failures per unit time
  • f(t)f(t) = Probability density function of the lifetime
  • F(t)F(t) = Cumulative distribution function: probability of failing by time t
  • R(t)R(t) = Reliability (survival) function, the probability of still working at time t
  • λ\lambda = Constant failure rate, estimated from test data

Worked Example

Problem: A lab runs 25 hard drives for 2,000 hours each. During the test, 5 drives fail and are immediately replaced, so all 25 positions run for the full 2,000 hours. Assume a constant failure rate. Find the failure rate λ\lambda, the mean time between failures (MTBF), and the probability that a drive keeps working for 1,000 hours.
Step 1: Find the total operating time across all positions. Replacing failed drives keeps every position running for all 2,000 hours, so the total is simply 25 times 2,000.
25×2,000=50,000 hours25 \times 2{,}000 = 50{,}000 \text{ hours}
Step 2: Divide failures by total operating time to get λ\lambda. The MTBF is its reciprocal.
λ=550,000=0.0001 failures per hour,MTBF=1λ=10,000 hours\lambda = \frac{5}{50{,}000} = 0.0001 \text{ failures per hour}, \qquad \text{MTBF} = \frac{1}{\lambda} = 10{,}000 \text{ hours}
Step 3: With a constant failure rate, reliability is R(t)=e−λtR(t) = e^{-\lambda t}. Substitute t=1,000t = 1{,}000.
R(1000)=e−0.0001⋅1000=e−0.1≈0.9048R(1000) = e^{-0.0001 \cdot 1000} = e^{-0.1} \approx 0.9048
Answer: The failure rate is 0.0001 failures per hour, the MTBF is 10,000 hours, and about 90.48% of drives survive 1,000 hours.

Why It Matters

Reliability engineers use failure rates to set warranty periods, plan maintenance schedules, and size spare-part inventories for aircraft, power grids, and data centers. Many products follow a bathtub curve: the failure rate starts high (early defects), stays roughly flat during useful life, and rises again as parts wear out. Only the flat middle section fits the constant-rate exponential model.

Common Mistakes

Mistake: Treating λt\lambda t as the probability of failure, for example claiming a 0.1 chance of failure in 1,000 hours when λ=0.0001\lambda = 0.0001 per hour.
Correction: A failure rate is not a probability. The probability of failure by time tt is 1−e−λt1 - e^{-\lambda t}, which here is 1−e−0.1≈0.09521 - e^{-0.1} \approx 0.0952. The shortcut λt\lambda t is only a close approximation when λt\lambda t is small.

Related Terms