Skip to content

Hardware Reliability Analysis - FTA / FMEA / FMEDA

This document summarizes hardware reliability analysis methods in automotive functional safety, focusing on FTA (Fault Tree Analysis) and FMEA/FMEDA in hardware metric calculations.


💡 Field Insight from Michael Lin

Background: ASIL D hardware safety metrics for an ADAS domain controller

Key Challenges:

  • Supplier FMEDA data incomplete, with non-conservative failure rate assumptions
  • FTA and FMEDA results could not be linked, leaving PMHF gaps
  • Inconsistent understanding of SPFM/LFM targets within the hardware team

My Solution:

  1. Built an FMEDA data review checklist, requiring IEC 62380 failure rates from suppliers
  2. Designed an FTA–FMEDA mapping matrix, ensuring each basic event has a failure rate
  3. Delivered hardware safety metrics training, aligning the team on ISO 26262-5

Quantified Results:

Metric Target Achieved Status
SPFM ≥ 99% 99.3%
LFM ≥ 90% 94.2%
PMHF < 10⁻⁸/h 6.2×10⁻⁹/h

Methodology Overview

flowchart TB subgraph TopDown["⬇️ Top-down"] FTA["FTA<br/>Fault Tree Analysis"] end subgraph BottomUp["⬆️ Bottom-up"] FMEA["FMEA<br/>Failure Mode & Effects Analysis"] FMEDA["FMEDA<br/>Failure Mode, Effects & Diagnostic Analysis"] end subgraph Metrics["📊 Hardware Metrics"] SPFM["SPFM<br/>Single-Point Fault Metric"] LFM["LFM<br/>Latent Fault Metric"] PMHF["PMHF<br/>Probabilistic Metric for Hardware Failures"] end FTA -->|"Identify minimal cut sets"| Metrics FMEA -->|"Failure classification"| FMEDA FMEDA -->|"Diagnostic coverage"| Metrics style FTA fill:#e3f2fd,stroke:#1976d2 style FMEA fill:#fff3e0,stroke:#f57c00 style FMEDA fill:#e8f5e9,stroke:#388e3c style Metrics fill:#fce4ec,stroke:#c2185b

FTA - Fault Tree Analysis

Overview

FTA is a deductive (top-down) analysis method: - Start from a top event (undesired system failure) - Decompose causes layer by layer - Use logic gates (AND/OR) to model causal relationships

FTA Symbols

flowchart TB subgraph Gates["Logic Gates"] OR_Gate["OR Gate<br/>Any input occurs<br/>→ output occurs"] AND_Gate["AND Gate<br/>All inputs occur<br/>→ output occurs"] end subgraph Events["Event Types"] TOP["Top Event<br/>(system-level failure)"] INT["Intermediate Event<br/>(subsystem failure)"] BASE["Basic Event<br/>(component failure)"] UNDEV["Undeveloped Event<br/>(to be analyzed)"] end

Example FTA - AEB System Failure

flowchart TB TOP["🔴 AEB Function Failure<br/>(Top Event)"] TOP --> OR1{OR} OR1 --> INT1["Perception failure"] OR1 --> INT2["Decision failure"] OR1 --> INT3["Actuation failure"] INT1 --> OR2{OR} OR2 --> B1["Camera failure"] OR2 --> B2["Radar failure"] OR2 --> B3["Fusion algorithm failure"] INT2 --> OR3{OR} OR3 --> B4["Main ECU failure"] OR3 --> B5["Software anomaly"] INT3 --> OR4{OR} OR4 --> B6["Brake actuator failure"] OR4 --> B7["Communication link failure"] style TOP fill:#ffcdd2,stroke:#c62828 style INT1 fill:#fff9c4,stroke:#f9a825 style INT2 fill:#fff9c4,stroke:#f9a825 style INT3 fill:#fff9c4,stroke:#f9a825

Core FTA Concepts

Concept Definition Meaning
Minimal Cut Set (MCS) Minimal combination of basic events leading to top event Identify critical failure paths
Single-Point Fault (SPF) Minimal cut set with one element System weak point
Common Cause Failure (CCF) Multiple components fail due to a common cause Redundancy considerations

FTA in ISO 26262

Use Case Purpose Output
Safety goal decomposition Validate FSR completeness Safety requirement verification
Quantitative analysis Calculate top event probability PMHF contribution
Architecture assessment Identify single-point faults SPFM input

FMEA - Failure Mode & Effects Analysis

Overview

FMEA is an inductive (bottom-up) analysis method: - Start from component failure modes - Analyze their effects on the system - Evaluate severity, occurrence, and detection

FMEA Process Flow

flowchart LR subgraph Input["Inputs"] I1["System architecture"] I2["Component list"] I3["Functional description"] end subgraph Analysis["Analysis Steps"] direction TB A1["1. Identify failure modes"] A2["2. Analyze effects"] A3["3. Identify causes"] A4["4. Evaluate S/O/D"] A5["5. Compute RPN"] A6["6. Define actions"] A1 --> A2 --> A3 --> A4 --> A5 --> A6 end subgraph Output["Outputs"] O1["FMEA worksheet"] O2["Risk priority"] O3["Improvement actions"] end Input --> Analysis --> Output

FMEA Evaluation Dimensions

Dimension Abbrev. Description Scale
Severity S Severity of impact 1-10
Occurrence O Failure frequency 1-10
Detection D Detectability of failure 1-10

RPN = S × O × D (Risk Priority Number)

FMEA Worksheet Template

Component Function Failure Mode Effect S Cause O Current Controls D RPN
Temp sensor Measure temperature Output drift Control deviation 6 Aging 3 Range check 4 72
Temp sensor Measure temperature No output Loss of function 8 Open circuit 2 Signal monitoring 2 32

FMEDA - Failure Mode, Effects & Diagnostic Analysis

Overview

FMEDA is an extension of FMEA, focused on: - Quantifying hardware failure rates - Assessing diagnostic coverage - Calculating ISO 26262 hardware metrics

FMEDA vs FMEA

Feature FMEA FMEDA
Failure rate Qualitative / semi-quantitative Quantitative (FIT)
Diagnostic coverage Not included Core element
Failure classification Risk level Safe / SPF / RF / MPF
Standard General quality ISO 26262-specific

FMEDA Failure Classification

flowchart TD FM["Component Failure Mode"] FM --> Q1{Violates safety goals?} Q1 -->|No| SAFE["✅ Safe Fault"] Q1 -->|Yes| Q2{Diagnosed?} Q2 -->|Fully covered| DETECTED["🔍 Detected Fault"] Q2 -->|Partially covered| RESIDUAL["⚠️ Residual Fault"] Q2 -->|Not covered| Q3{Independent redundancy?} Q3 -->|Yes| MPF["🔸 Multi-Point Fault"] Q3 -->|No| SPF["🔴 Single-Point Fault"] style SAFE fill:#c8e6c9,stroke:#388e3c style DETECTED fill:#fff9c4,stroke:#f9a825 style RESIDUAL fill:#ffe0b2,stroke:#f57c00 style MPF fill:#ffccbc,stroke:#e64a19 style SPF fill:#ffcdd2,stroke:#c62828

Failure Rate Allocation

Failure Category Symbol Definition Metric Impact
Safe fault λ_S Does not violate safety goals Not counted in risk
Single-point fault λ_SPF Directly violates safety goals Impacts SPFM
Residual fault λ_RF Not fully covered by diagnostics Impacts SPFM
Multi-point (detected) λ_MPF_D Detected multi-point fault -
Multi-point (latent) λ_MPF_L Latent multi-point fault Impacts LFM

ISO 26262 Hardware Metrics

Three Key Metrics

flowchart LR subgraph SPFM_Calc["SPFM Metric"] SPFM_F["SPFM = 1 - (λ_SPF + λ_RF) / λ_total"] end subgraph LFM_Calc["LFM Metric"] LFM_F["LFM = 1 - λ_MPF_L / (λ_MPF_L + λ_MPF_D)"] end subgraph PMHF_Calc["PMHF Metric"] PMHF_F["PMHF = Σ(failure probability contribution)"] end

Metric Targets (ISO 26262-5)

ASIL SPFM Target LFM Target PMHF Target
ASIL B ≥ 90% ≥ 60% < 10⁻⁷/h
ASIL C ≥ 97% ≥ 80% < 10⁻⁷/h
ASIL D ≥ 99% ≥ 90% < 10⁻⁸/h

Diagnostic Coverage Levels

Level DC Range Typical Measures
None < 60% No diagnostics
Low 60-90% Range check, plausibility check
Medium 90-99% Redundant comparison, CRC
High ≥ 99% HW redundancy, diversity design

FTA + FMEA/FMEDA Integration

flowchart TB subgraph FTA_Role["FTA Role"] F1["Identify system-level failure paths"] F2["Determine minimal cut sets"] F3["Validate architecture integrity"] F4["Quantify system failure probability"] end subgraph FMEDA_Role["FMEDA Role"] M1["Quantify component failure rates"] M2["Classify failure modes"] M3["Assess diagnostic coverage"] M4["Calculate hardware metrics"] end subgraph Integration["Integrated Analysis"] I1["FMEDA provides basic event rates"] I2["FTA computes system probability"] I3["Together support PMHF"] end FTA_Role --> Integration FMEDA_Role --> Integration Integration --> Result["Hardware Safety Integrity Evaluation"]
Phase FTA Tasks FMEA/FMEDA Tasks
Concept design Initial FTA to identify critical paths Initial FMEA to identify failure modes
System design Refine FTA to validate architecture System-level FMEA
Hardware design Hardware FTA Hardware FMEDA for metrics
Detailed design Quantitative FTA (PMHF) Component-level FMEDA
Verification Validate FTA results Validate FMEDA assumptions

Implementation Notes

Data Sources

Data Type Source Notes
Failure rates SN 29500, IEC 62380 Industry standard databases
Supplier data Component supplier FMEDA More accurate real data
Field data After-sales return analysis Most real but delayed

Common Challenges

Challenge Mitigation
Insufficient failure rate data Conservative estimates + sensitivity analysis
Diagnostic coverage assessment difficulty Fault injection testing
Common cause failure modeling Beta-factor method
High analysis effort Tool support + layered analysis

Last updated: 2026-01-25

ML
Verified

Michael Lin

MAGNA Quality Manager
Architect of Compliance-Wächter
AI for ISO 21434 & UN R155

作为 MAGNA 质量经理,我在德国汽车电子行业深耕 10+ 年。每天我都见证着一个新的瓶颈: 网络安全合规 (UN R155 / ISO 21434)

我正在构建 Compliance-Wächter — 一个 AI 驱动的 Copilot, 自动化 TARA 分析,将合规文书工作减少 70%