3.6 Reliability & Maintainability Concepts

Section 3.6: Reliability & Maintainability Concepts

Reliability is a quantitative measure of performance over time. Formally, reliability $R(t)$ is defined as the probability that a product, component, or system will perform its intended function satisfactorily for a specified operating period ($t$) under stated environmental and operational conditions.

Mathematically, if $T$ is the continuous random variable representing time-to-failure, the reliability function $R(t)$ is the complement of the cumulative distribution function (unreliability $F(t)$):

$$R(t) = P(T > t) = 1 - F(t) = 1 - \int_{0}^{t} f(\tau) d\tau$$

Where $f(t)$ is the failure probability density function (PDF).

---

Key Reliability Metrics

Quality and reliability engineers utilize specific metrics depending on whether an item is repairable or non-repairable:

Note on FITs: One FIT (Failure In Time) is defined as $1 \text{ failure per } 10^9 \text{ operating hours}$.

---

Constant Failure Rate & The Exponential Distribution Model

When a component operates during its useful life phase, its failure rate $\lambda(t)$ remains constant over time (i.e., $\lambda(t) = \lambda$). Under the constant failure rate assumption, time-to-failure follows the exponential distribution.

Reliability & Hazard Formulas

$$\text{Failure Rate } \lambda = \frac{1}{\text{MTBF}}$$ $$R(t) = e^{-\lambda t} = e^{-\frac{t}{\text{MTBF}}}$$ $$\text{Unreliability } F(t) = 1 - e^{-\lambda t}$$ $$\text{Failure PDF } f(t) = \lambda e^{-\lambda t}$$

Critical Statistical Property: The Benchmark at $t = \text{MTBF}$

Setting operating time $t = \text{MTBF}$ yields:

$$R(\text{MTBF}) = e^{-1} = \frac{1}{e} \approx 0.3679878 \quad (36.8\%)$$

Exam Tip: Operating a system for a duration equal to its MTBF does not guarantee $100\%$ or $50\%$ survival. In a constant failure rate model, there is only a $36.8\%$ probability that a component will survive to its MTBF; $63.2\%$ of units will fail prior to reaching MTBF.

Practical Calculation Example: Industrial Compressor

An industrial plant operates a cooling compressor with a documented MTBF of $4,000\text{ operating hours}$.

1. Calculate the failure rate $\lambda$:

   $$\lambda = \frac{1}{4,000\text{ hours}} = 0.00025\text{ failures per hour}$$

2. Calculate the probability of surviving a 1,000-hour mission ($R(1000)$):

   $$R(1000) = e^{-(0.00025 \times 1000)} = e^{-0.25} \approx 0.77880 \quad (77.88\%)$$

3. Calculate expected field survivals: If 500 compressors are deployed across various plant facilities for 1,000 operating hours, the expected number of surviving units is $500 \times 0.7788 = 389.4 \approx 389\text{ units}$.

---

The Bathtub Curve Hazard Function

The lifetime hazard rate $\lambda(t)$ of a population of electromechanical hardware follows the classic Bathtub Curve, which consists of three distinct chronological phases:

Hazard Rate lambda(t)
    ^
    |  Infant Mortality   |   Useful Life (Random Failures)   |   Wear-Out Phase
    |  (Burn-in Phase)    |   Constant Hazard Rate lambda     |   Increasing Hazard
    |\                    |                                   |                    /
    | \                   |                                   |                   /
    |  \__________________|___________________________________|__________________/
    +----------------------------------------------------------------------------------> Time (t)
       Phase 1: dlambda/dt < 0        Phase 2: dlambda/dt = 0       Phase 3: dlambda/dt > 0

1. Phase 1: Infant Mortality / Burn-in Phase ($d\lambda/dt < 0$)

• Characteristics: High initial failure rate that rapidly decreases over early operating time.
• Root Causes: Manufacturing defects, assembly errors, flawed soldering, substandard materials, or contaminated semiconductor wafers.
• Engineering Mitigation: Burn-in testing or Environmental Stress Screening (ESS) (e.g., thermal cycling, random vibration) prior to customer shipment to trigger and purge early failures in the factory.

2. Phase 2: Useful Life Phase ($d\lambda/dt = 0$)

• Characteristics: Low, constant failure rate ($\lambda(t) = \lambda$). Failures occur randomly due to unpredictable external stress spikes.
• Root Causes: Lightning strikes, unexpected operator abuse, extreme pressure surges, random environmental shock.
• Mathematical Model: Exponential distribution ($R(t) = e^{-\lambda t}$).

3. Phase 3: Wear-out Phase ($d\lambda/dt > 0$)

• Characteristics: Failure rate increases rapidly with age.
• Root Causes: Mechanical friction, fatigue cracking, insulation breakdown, chemical oxidation, corrosion, polymer degradation.
• Engineering Mitigation: Implement Preventive Maintenance (PM) and scheduled replacement before components enter the steep wear-out region.

---

Maintainability & Operational Availability

While reliability measures the probability of avoiding failure, maintainability $M(t)$ measures the ease and speed with which a failed system can be restored to operational status.

$$\text{Maintainability } M(t) = 1 - e^{-\mu t} = 1 - e^{-\frac{t}{\text{MTTR}}}$$

Where $\mu = \frac{1}{\text{MTTR}}$ is the repair rate.

Operational Availability ($A$)

Availability ($A$) combines reliability and maintainability into a unified performance metric representing the probability that a system is operating satisfactorily at any random point in time.

$$\text{Operational Availability } A = \frac{\text{Uptime}}{\text{Uptime} + \text{Downtime}} = \frac{\text{MTBF}}{\text{MTBF} + \text{MTTR}}$$

Comprehensive Calculation Example

A automated assembly line operates for $2,400\text{ hours}$ in a quarter. During this period, the line experiences $6\text{ unscheduled breakdown events}$. Total downtime accumulated during repair, part fetching, and re-commissioning is $48\text{ hours}$.

1. Calculate MTBF:

   $$\text{MTBF} = \frac{2,400\text{ operating hours}}{6\text{ failures}} = 400\text{ hours per failure}$$

2. Calculate Failure Rate $\lambda$:

   $$\lambda = \frac{1}{400} = 0.0025\text{ failures per hour}$$

3. Calculate MTTR:

   $$\text{MTTR} = \frac{48\text{ downtime hours}}{6\text{ repairs}} = 8\text{ hours per repair}$$

4. Calculate Operational Availability $A$:

   $$A = \frac{\text{MTBF}}{\text{MTBF} + \text{MTTR}} = \frac{400}{400 + 8} = \frac{400}{408} \approx 0.98039 \quad (98.04\%)$$

Improving availability can be accomplished either by increasing MTBF (design for reliability) or by reducing MTTR (design for maintainability, quick-change modules, standardized diagnostic tools).