Lecture 4 - Reliability Basics

Question 1

What does reliability do?

Answer 1

• Failures happen, we need to understand why and how.
• We want to PREVENT failures and CONTROL their impact(s).
• Prevention of failures requires an understanding of the failure mechanism(s) and the physics of failure.
• Most failure mechanisms, their interactions and processes of degradation are not fully understood. Hence PREDICTION of failures involved uncertainty and is inherently probabilistic.
• Predicting failure occurrence requires application of stochastic processes and the random nature of failures.

Question 2

What are the 3 focus areas for reliability engineers?

Answer 2

• Hardware
• Software
• Human

Question 3

What is a system?

Answer 3

• A system is a collection of components that provides a FUNCTION

Question 4

What is the relationship between a system and a component?

Answer 4

• The line between component and system is arbitrary and depends on objectives, scope, state of the art and conventions.

Question 5

Whats in the hardware domain?

Answer 5

Component reliability
* Failure modes
* Failure mechanisms
* Failure probability model

System reliability
* Functional model
* Failure/ success logic model
* Dependencies between components
* Failure mechanisms
* Probability models

Question 6

What is in the software domain?

Answer 6

• Software is increasingly used to perform functions that may previously have been done by hardware and/ or humans.
• Software does not fail on the ways that hardware does.
• Every copy of a computer program is identical so failures due to variability cannot occur. When a bug does exist, it exists in all copies of the program.
• Software does not decade, except in a few special senses (e.g if the media degrades).
• Errors can be difficult to find in software.
• Software failures can also occur as a function of the machine environment. Identical copies can behave differently depending on factors such as ‘age’ since reboot.

Question 7

What is in the human factors domain?

Answer 7

Skill-based: Inattention and Over-attention
Rule-based: Misapplication of good rules and Application of Bad Rules
Knowledge-based: Selectivity, Out of sight out of mind, Confirmation Bias, Over-confidence, Illusory correlation, Halo effects, Problems with causality, Problems with complexity.

Question 8

What are conceptual failure models?

Answer 8

Failures are due to a complex set of interactions between:
* The stresses that act on and within the system, and
* The materials/ elements of the system.

There are 4 simple conceptual models (Dasgupta, 1991)
* Stress-strength: the item fails if and only if the stress exceeds the strength. This depends on the occurrence of events.
* Damage-endurance: a stress causes damage that accumulates irreversibly (corrosion, wear, fatigue). The item fails when damage exceeds the endurance.
* Challenge-response: an element is bad but only when the element is challenged (needed) does it reveal itself to be bad.
* Tolerance-requirement: performance is satisfactory if its tolerance remains within requirement.