Data in US-MIL-Handbook 217 from RADC

Maintenance Management (Rome Air Development Centre - US-DoD)

Concepts and Practices ! Microelectronics

! Discrete semi-conductors
! Tubes
Failure Data and Risk ! Lasers
! Resistors and capacitors
! Inductors
! Connections and connectors
Dr. Chakib KARA-ZAITRI ! Meters
! Crystals
! Lamps, fuses and other miscellaneous items 5

Calculation of the failure

Failure Data using US-MIL-Handbook 217
! US-MIL-Handbook 217 p=Q(C1TV+C2E)L Failures / 106 hours
! CNET (French PTT) Data ! p= Device failure rate in F/106 hors
! HRD (British Telecom) ! Q= Quality Factor
! RADC Non-Electric Parts Handbook NPRD ! T= Temperature acceleration factor
! OREDA (Off-shore Data) ! V=Voltage stress derating factor
! SRD (Systems Reliability Department: Fee ! E=Application environment factor
required)
! C1= Circuit complexity factor
! ESReDA (European Safety and Reliability
! C2= Package complexity failure rate
Data: Fee required)
2
! L= Device learning factor 6

Regression equation for failure

data models (Micro-electronics) Field data in NPRD95
! Complexity (number of gates, pins and ! Electromechanical parts
equivalent number of transistors) ! Mechanical parts
! Junction temperature ! Hydraulic parts
! Package (ceramic and plastic package) ! Pneumatic parts
! Technology (CMOS, NMOS, Bipolar, etc.)
! Type (Memory, Random LSI, analogue,
etc.)
! Voltage
! Quality level (for screening and burn-in)
! Environment 3 7

Data in Faradip.Three Data in HRD5

! Microelectronics ! Integrated circuits
! Discrete semi-conductors, tubes and lamps ! Discrete semi-conductors
! Passive electrical components ! Capacitors
! Instruments and analysers ! Resistors
! Connectors, switches, PCBs and cables ! Electromechanical and wound components
! Electromechanical and rotating machinery ! Opto-electronics
! Power supplies, sources and transformers ! Surge protection
! Mechanical items (pumps and valves) ! Switches
! Computers ! Visual devices
! Alarms, fire protection, arrestors and fuses ! Miscellaneous (including microwave)
4 8

Failure data and Risk.doc 11

 Definition of a disaster
Data in OREDA from
(Oxford Dictionnary)
! British Petroleum Norway
! Elf Aquitaine Norge
! Norsk Agip A/S "Anything that befalls of ruinous
! A/S Norske Shell or distressing nature; a sudden
! Norsk Hydro a.s. or great misfortune, mishap or
! Statoil
misadventure; a calamity.
! Saga Petroleum a.s.
! Total Oil Marine plc.

9 13

Data in OREDA
! Fire and gas detection systems Another definition of a disaster
! Process alarm systems
! Fire fighting systems "An event which afflicts a
! Emergency shut down systems
! Pressure relieving systems community, the consequences of
! General alarm and communication systems which are beyond the immediate
! Evacuation systems financial, material or emotional
! Vessels
! Valves resources of the community".
! Pumps
! Heat Exchangers
! Compressors
! Power generation and switching
! Cranes
! Drilling systems and equipment 10 14

Guided Weapon Systems

(DX/99/013-100) BDS definition of a disaster
! Discrete electronic and electromesh An event where ten or more
! Connectors fatalities result from one event
! Micro-electronics which occurs over a relatively short
! One-shot devices
period of time."
! Mechanical items

11 15

Failure mechanisms The Bradford Disaster Scale

! Alloy formation ! Friction Number of fatalities Magnitude
! Biological effects ! Ionisation of gases
! Chemical and ! Ion migration 10 1
electrolytic changes ! Magnetic degradation 100 2
! Contamination ! Mechanical stresses
! Depolymerisation ! Metallic effects
1000 3
! Electrical contact failure ! Moisture gain or loss 10000 4
! Evaporation ! Molecular migration
! Fatigue !
100000 5
Stress relaxation
! Film deposition ! Temperature cycling 1000000 6
12 16

Failure data and Risk.doc 12

 Absolute limits of the scale Frequency of disasters p.a.

Losses Type Magnitude BDS Europe U.K

Frequency p.a. Frequency p.a.
----- ---------------------------------------- ---------
------ -------------------- --------------------
1010 Destruction of the entire planete 10
1 9 2
1020 Destruction of the entire galaxy 20 2 2 0.33
10 30 Destruction of the entire universe 30 3 0.2 0.04
4 0.05 0.007
5 0.01 0.001

17 21

Classification in terms of the Number of disasters in Europe

number of fatalities 10-year interval
250

Number of fatalities Class 200

101 - 102 1 150

Nombrede dsastres
102 - 103 2
100
103 - 104 3
104 - 105 4 50

105 - 106 5 0
1898 1908 1918 1928 1938 1948 1958 1968 1978 1988
18 22

Evaluation of disasters using EUROPEANDISASTERFATALITIES1888-1988

NATURAL/ MANMADEDISASTERS
BDS 25000

Disaster N.of fatalities Magnitude Classification

20000 MANMADEDISATERS

------------------------ ------------------ ------------- ---------------- NATURALDISASTRES

Clapham Commun 36 1.56 1

15000
No. of FATALITIES

Hillsborough 95 1.98 2
10000
Piper Alpha 166 2.22 3
Hrald Free Ent. 188 2.27 4
5000
Sisme Armnien 24000 4.38 5
Inond. Bangladesh 2000000 6.33 6
0
1898 1908 1918 1928 1938 1948 1958 1968 1978 1988
19 10YEARPERIOD 23

Frequency of disasters EUROPEANDISASTERS BDSCLASS1

1888 - 1988 PROBABILITYOFFREQUENCYPERANNUM
0,2
OBSERVEDPROBABILITY
Class Europe UK 0,15
EXPECTEDPROBABILITY
--------- --------- ------
PROBABILITYOF
FREQUENCY

1 901 237 0,1

2 217 44
0,05
3 19 1
4 4 1 0
5 1 0 1 2 3 4 5 6 7 8 9 10 11 12
20 FREQUENCYOFOCCURENCE 24

Failure data and Risk.doc 13

 EUROPEANDISASTERSBDSCLASS2
PROBABILITYOFFREQUENCYPERANNUM
Risk Management Philosophy
0.3
! Safety requirements
OBSERVEDPROBABILITY
0.25 Constraints which contour the life cycle of a
EXPECTEDPROBABILITY technological system from the view point of safety, and
0.2 with which compliance must be verified.
PROBABILITYOF
FREQUENCY

0.15 ! Risk management

Systems and procedures for reducing or eliminating risk
0.1 including human, organisational, environmental and
technological aspects.
0.05
! Residual risk
0 How it is assessed, evaluated, communicated to the
appropriate management level and how it is fed back
0 1 2 3 4 5 6 7
into the requirement compliance verification loop.
FREQUENCYOFOCCURRENCE 25 29

U.K. DISASTERSBDSCLASS1 Risk assessment model

PROBABILIYOFFREQUENCYPERANNUM
Safety optimisation
0,3
0,25 OBSERVEDPROBABILITY
EXPECTEDPROBABILITY
0,2
Requirements Compliance Verification
PROBABILITYOF
FREQUENCY

0,15
0,1
0,05
Residual risk assessment
0
0 1 2 3 4 5 6

FREQUENCYOFOCCURRENCE Risk acceptance

26 30

Definition of Hazard Risk Management Philosophy

! Qualitative safety optimisation
! Quantitative safety optimisation
Hazards are threats to humans and the
things they value.
The description of threats requires
!Subjective data
knowledge of the causal sequence of
events that link early stages, such as
human needs and wants, with eventual
experience of human harm in the form
of death or injury.

27 31

Definition of Risk Overall Risk Assessment

Risks are quantitative measures of ! Problem definition

human harm, most frequently
! System boundary
expressed as conditional
probabilities for experiencing harm. ! Probabilistic risk assessment

28 32

Failure data and Risk.doc 14

 Risk analysis stages Individual Risk Example
! There are 50,000 car fatalities per year in the USA.
! System description
! Identification of unwanted events ! The individual risk (probability) of death of any of
! Estimation of probabilities the 200 million people in the USA, per year, from
car accidents is:
! Calculation of damages
! Interpretation of the risk 50,000 deaths / year Deaths
---------------------------- = 0.00025 --------
200,000,000 persons person-year

33 37

Acceptable risk Risk definition

! Risk assumed voluntarily ! Risk borne un-voluntarily
!
!
Effect immediate
No alternative available
!
!
Effect delayed
Many alternatives available
Risk = Frequency x Consequence
! Risk known with certainty ! Risk not known Risk (consequence / unit time)
! Encountered occupationally ! Encountered non-occupationally
Frequency (events / unit time)
! Common hazard ! Dread hazard
! Affects average people ! Affects especially sensitive people Consequence (consequence / event)
! Will be used as intended ! Likely to be misused
! Consequences reversible ! Consequences irreversible

34 38

Risk Example The Fatal Accident Frequency Rate

FAFR
! E.g.:The number of people expected to be killed per
year due to a reactor explosion is 0.0001. I.e., for ! Thisis the average number of
every 10,000 hours, we predict one death. deaths by accidents in 10^8
hours of a particular activity.
! It is of societal interest to compare a number such as
1 death per 10,000 hrs with the risks involved in
everyday living in order to obtain some idea of what
might constitute a reasonable outcome for a risk
analysis.

35 39

Risk definition

! Dictionary
The possibility of loss or injury to
people and property.
! Risk Engineers
The probability of loss or injury to
people and property.

36

Failure data and Risk.doc 15

Failure data and Risk.doc 16