MICHAEL TODINOV has a background in mechanical engineering, applied mathematics and computer science. He received his PhD and his higher doctorate (DEng) from the University of Birmingham and is currently a professor in mechanical engineering in Oxford Brookes University, UK. Professor Todinov is an internationally acclaimed expert in reliability and risk. In 2017, he received the prestige IMechE award in the area of risk reduction in mechanical engineering. He has published four research monographs and a large number of research papers in the area of reliability and risk. His name is associated with developing theoretical and computational frameworks for analysis and optimisation of repairable flow networks and for reliability analysis based on the cost of failure.

Preface xv

1 Domain-Independent Methods for Reliability Improvement and Risk Reduction 1

1.1 The Domain-Specific Methods for Risk Reduction 1

1.2 The Statistical, Data-Driven Approach 3

1.3 The Physics-of-Failure Approach 4

1.4 Reliability Improvement and TRIZ 6

1.5 The Domain-Independent Methods for Reliability Improvement and Risk Reduction 6

2 Basic Concepts 9

2.1 Likelihood of Failure, Consequences from Failure, Potential Loss, and Risk of Failure 9

2.2 Drawbacks of the Expected Loss as a Measure of the Potential Loss from Failure 14

2.3 Potential Loss, Conditional Loss, and Risk of Failure 15

2.4 Improving Reliability and Reducing Risk 19

2.5 Resilience 21

3 Overview of Methods and Principles for Improving Reliability and Reducing Risk That Can Be Classified as Domain-Independent 23

3.1 Improving Reliability and Reducing Risk by Preventing Failure Modes 23

3.1.1 Techniques for Identifying and Assessing Failure Modes 23

3.1.2 Effective Risk Reduction Procedure Related to Preventing Failure Modes from Occurring 27

3.1.3 Reliability Improvement and Risk Reduction by Root Cause Analysis 28

3.1.3.1 Case Study: Improving the Reliability of Automotive Suspension Springs by Root Cause Analysis 28

3.1.4 Preventing Failure Modes by Removing Latent Faults 29

3.2 Improving Reliability and Reducing Risk by a Fault-Tolerant System Design and Fail-Safe Design 31

3.2.1 Building in Redundancy 31

3.2.1.1 Case Study: Improving Reliability by k-out-of-n redundancy 34

3.2.2 Fault-Tolerant Design 34

3.2.3 Fail-Safe Principle and Fail-Safe Design 35

3.2.4 Reducing Risk by Eliminating Vulnerabilities 36

3.2.4.1 Eliminating Design Vulnerabilities 36

3.2.4.2 Reducing the Negative Impact of Weak Links 37

3.2.4.3 Reducing the Likelihood of Unfavourable Combinations of Risk-Critical Random Factors 38

3.2.4.4 Reducing the Vulnerability of Computational Models 39

3.3 Improving Reliability and Reducing Risk by Protecting Against Common Cause 40

3.4 Improving Reliability and Reducing Risk by Simplifying at a System and Component Level 42

3.5 Improving Reliability and Reducing Risk by Reducing the Variability of Risk-Critical Parameters 44

3.5.1 Case Study: Interaction Between the Upper Tail of the Load Distribution and the Lower Tail of the Strength Distribution 46

3.6 Improving Reliability and Reducing Risk by Making the Design Robust 48

3.6.1 Case Study: Increasing the Robustness of a Spring Assembly with Constant Clamping Force 50

3.7 Improving Reliability and Reducing Risk by Built-in Reinforcement 51

3.7.1 Built-In Prevention Reinforcement 51

3.7.2 Built-In Protection Reinforcement 51

3.8 Improving Reliability and Reducing Risk by Condition Monitoring 52

3.9 Reducing the Risk of Failure by Improving Maintainability 56

3.10 Reducing Risk by Eliminating Factors Promoting Human Errors 57

3.11 Reducing Risk by Reducing the Hazard Potential 58

3.12 Reducing Risk by using Protective Barriers 59

3.13 Reducing Risk by Efficient Troubleshooting Procedures and Systems 60

3.14 Risk Planning and Training 60

4 Improving Reliability and Reducing Risk by Separation 61

4.1 The Method of Separation 61

4.2 Separation of Risk-Critical Factors 62

4.2.1 Time Separation by Scheduling 62

4.2.1.1 Case Study: Full Time Separation with Random Starts of the Events 62

4.2.2 Time and Space Separation by Using Interlocks 63

4.2.2.1 Case Study: A Time Separation by Using an Interlock 63

4.2.3 Time Separation in Distributed Systems by Using Logical Clocks 64

4.2.4 Space Separation of Information 65

4.2.5 Separation of Duties to Reduce the Risk of Compromised Safety, Errors, and Fraud 65

4.2.6 Logical Separation by Using a Shared Unique Key 66

4.2.6.1 Case Study: Logical Separation of X-ray Equipment by a Shared Unique Key 66

4.2.7 Separation by Providing Conditions for Independent Operation 67

4.3 Separation of Functions, Properties, or Behaviour 68

4.3.1 Separation of Functions 68

4.3.1.1 Separation of Functions to Optimise for Maximum Reliability 68

4.3.1.2 Separation of Functions to Reduce Load Magnitudes 70

4.3.1.3 Separation of a Single Function into Multiple Components to Reduce Vulnerability to a Single Failure 71

4.3.1.4 Separation of Functions to Compensate Deficiencies 71

4.3.1.5 Separation of Functions to Prevent Unwanted Interactions 71

4.3.1.6 Separation of Methods to Reduce the Risk Associated with Incorrect Mathematical Models 72

4.4 Separation of Properties to Counter Poor Performance Caused by Inhomogeneity 72

4.4.1 Separation of Strength Across Components and Zones According to the Intensity of the Stresses from Loading 72

4.4.2 Separation of Properties to Satisfy Conflicting Requirements 74

4.4.3 Separation in Geometry 75

4.4.3.1 Case Study: Separation in Geometry for a Cantilever Beam 75

4.5 Separation on a Parameter, Conditions, or Scale 76

4.5.1 Separation at Distinct Values of a Risk-Critical Parameter Through Deliberate Weaknesses and Stress Limiters 76

4.5.2 Separation by Using Phase Changes 77

4.5.3 Separation of Reliability Across Components and Assemblies According to Their Cost of Failure 77

4.5.3.1 Case Study: Separation of the Reliability of Components Based on the Cost of Failure 78

5 Reducing Risk by Deliberate Weaknesses 81

5.1 Reducing the Consequences from Failure Through Deliberate Weaknesses 81

5.2 Separation from Excessive Levels of Stress 82

5.2.1 Deliberate Weaknesses Disconnecting Excessive Load 82

5.2.2 Energy-Absorbing Deliberate Weaknesses 85

5.2.2.1 Case Study: Reducing the Maximum Stress from Dynamic Loading by Energy-Absorbing Elastic Components 85

5.2.3 Designing Frangible Objects or Weakly Fixed Objects 86

5.3 Separation from Excessive Levels of Damage 87

5.3.1 Deliberate Weaknesses Decoupling Damaged Regions and Limiting the Spread of Damage 87

5.3.2 Deliberate Weaknesses Providing Stress and Strain Relaxation 88

5.3.3 Deliberate Weaknesses Separating from Excessive Levels of Damage Accumulation 90

5.4 Deliberate Weaknesses Deflecting the Failure Location or Damage Propagation 91

5.4.1 Deflecting the Failure Location from Places Where the Cost of Failure is High 91

5.4.2 Deflecting the Failure Location from Places Where the Cost of Intervention for Repair is High 92

5.4.3 Deliberate Weaknesses Deflecting the Propagation of Damage 92

5.5 Deliberate Weaknesses Designed to Provide Warning 92

5.6 Deliberate Weaknesses Designed to Provide Quick Access or Activate Protection 94

5.7 Deliberate Weaknesses and Stress Limiters 94

6 Improving Reliability and Reducing Risk by Stochastic Separation 97

6.1 Stochastic Separation of Risk-Critical Factors 97

6.1.1 Real-Life Applications that Require Stochastic Separation 97

6.1.2 Stochastic Separation of a Fixed Number of Random Events with Different Duration Times 99

6.1.2.1 Case Study: Stochastic Separation of Consumers by Proportionally Reducing Their Demand Times 102

6.1.3 Stochastic Separation of Random Events Following a Homogeneous Poisson Process 105

6.1.3.1 Case Study: Stochastic Separation of Random Demands Following a Homogeneous Poisson Process 106

6.1.4 Stochastic Separation Based on the Probability of Overlapping of Random Events for More than a Single Source Servicing the Random Demands 106

6.1.5 Computer Simulation Algorithm Determining the Probability of Overlapping for More than a Single Source Servicing the Demands 108

6.2 Expected Time Fraction of Simultaneous Presence of Critical Events 110

6.2.1 Case Study: Expected Fraction of Unsatisfied Demand at a Constant Sum of the Time Fractions of User Demands 112

6.2.2 Case Study: Servicing Random Demands from Ten Different Users, Each Characterised by a Distinct Demand Time Fraction 114

6.3 Analytical Method for Determining the Expected Fraction of Unsatisfied Demand for Repair 114

6.3.1 Case Study: Servicing Random Repairs from a System Including Components of Three Different Types, Each Characterised by a Distinct Repair Time 115

6.4 Expected Time Fraction of Simultaneous Presence of Critical Events that have been Initiated with Specified Probabilities 116

6.4.1 Case Study: Servicing Random Demands from Patients in a Hospital 117

6.4.2 Case Study: Servicing Random Demands from Four Different Types of Users, Each Issuing a Demand with Certain Probability 118

6.5 Stochastic Separation Based on the Expected Fraction of Unsatisfied Demand 119

6.5.1 Fixed Number of Random Demands on a Time Interval 119

6.5.2 Random Demands Following a Poisson Process on a Time Interval 120

6.5.2.1 Case Study: Servicing Random Failures from Circular Knitting Machines by an Optimal Number of Repairmen 122

7 Improving Reliability and Reducing Risk by Segmentation 125

7.1 Segmentation as a Problem-Solving Strategy 125

7.2 Creating a Modular System by Segmentation 127

7.3 Preventing Damage Accumulation and Limiting Damage Propagation by Segmentation 129

7.3.1 Creating Barriers Containing Damage 129

7.3.2 Creating Weak Interfaces Dissipating or Deflecting Damage 131

7.3.3 Reducing Deformations and Stresses by Segmentation 131

7.3.4 Reducing Hazard Potential by Segmentation 131

7.3.5 Reducing the Likelihood of Errors by Segmenting...