Reliability Engineering and Quality Assurance Flashcards
Reliability definition
Defined as the ability of a system to perform its intended mission when operating for a designated period of time, or through a planned mission scenario (or series of scenarios), in a realistic operational environment.
– A system with a 90% reliability has a 90% probability that the system will operate the mission duration without a critical failure
Inherent within this definition are the elements of probability, satisfactory performance, time or mission-related cycle, and specified operating conditions.
– The tools of reliability engineers include heavy doses of probability and statistics and specialized tools like fault trees and reliability block diagrams, as well as traditional engineering tools of modeling and simulation
Reliability, Availability, Maintainability, & Supportability (RAMS)
Reliability Block Diagram (BRD) as a concept
Represents the system’s functioning state (success or failure) in terms of the function states of it’s components
The RBD demonstrates the effect of a failure of a component in the success or failure of the system
RBDs for a system
All real systems are built up of some system elements that are arranged reliability-wise in series and some system elements that are arranged relaibility-wise in parallel
“Reliability-wise in series”
If even one component fails, the whole system fails (like the links of a chain… none of them can fail or the whole thing fails)
Indicates an ‘product’ relationship in terms of Boolean algebra
- That same equation also applied when calculating the probability of system failure (for each factor, you just calculate the probability that that component will fail in the given time period and
The least reliable component
“Reliability-wise in parallel”
Indicates Redundancy, extremely common in automotive and aerospace
Quantifying this is a little trickier than it seems:
The reliability of a system build up from components reliability-wise in parallel is 1-ΠF
ΠF is the probability that ALL the components will fail within the mission duration (product of CDF for each component failure)
Quantifying Reliability
The exact number depends on the time period
A component’s reliability is represented by the probability that it will fail in the given time period (solved using an integral)
System reliability is represented by the probability that the entire system will fail in the given time period (mission duration)
- For components ‘reliability wise in series’: It is the product of those integrals
F-function of a component
Represents the probability that the component will fail.
It is equivalent to one minus the reliability function F=1-R
Reliability function of a component
Represents the probability that the component will NOT fail.
It is equivalent to one minus the reliability function R=1-F
Mean time to failure (MTTF)
Represents how long the component or system will be expected to last, on average
This is particularly important is exponential failure distributions
Failure distributions
Calculating MTTF for a system
Reduction of complex RBDs
“Fault” Definition
A latent defect condition-incompatibility, degradation, or deterioration; OR a hazard that has the potential to materialize into a failure
Fault Tree Analysis
Represents an option when it comes to understanding how a system may fail
- Uses Boolean algebra symbols to represent sequencing and dependencies
Cut Set in Fault Tree Analysis
Set of basic events whose simultaneous occurrence would ensure system failure if they occur
Described using set theory
MINIMAL cut sets represent the smallest set of events that could cause a system failure (meaning that each of the events in the cut set actually contribute to the root cause and there are no extraneous events in the set)
Each cut set has a probability
Root Cause
The actual
Basic Assumption of a Cut Set
You are assuming that the basic events are independent
NOT a Markov Chain
Probabilistic Risk Assessment
ISO 9001
Failure Rate
Failure Modes and Effects Analysis (FMEA)
Common Continuous Reliability Distributions
The two most commonly used reliability distributions are the exponential and the Weibull. Both are among the most mathematically tractable to work with; the exponential is a constant failure rate model, and the Weibull is adaptable to a wide variety of failure rate scenarios.
1) Exponential
2) Weibull
3) Normal
4) Lognormal
5) Beta
6) Gamma
7) Rayleigh
8) Uniform
9) Extreme Value
10) Logistic
11) Log logistic
12) Pareto
13) Inverse Gaussain
14) Makeham
15) Hyperexponential
16) Muth
Exponential Failure Model
Weibull Failure Model
Part Stress Analysis (PSA)
– assessment of a part’s reliability due to construction and application
– utilizes specific attribute data such as component technology, package type, complexity and quality, as well as application-specific data such as electrical and environmental stresses
– applicable when most of the design is completed and a detailed parts list including part stresses is available
Part Count Analysis (PCA)
– a less-refined estimator relying on default values of most of the part and application-specific parameters
– applicable during bid proposal and early design phases when insufficient information is available to use the part stress analysis models
MIL HDBK 217
Bellcore (Telecommunications Industry standard)
Accelerated Life Testing
Highly Accelerated Life Test v. Highly Accelerated Stress Screen
Bathtub Curve