1 2

Mulugeta D. Abera Mulugeta D. Abera

7. BASIC PROBABILITY CONCEPTS 7.1 Block diagram analysis
Block diagrams are logic diagrams which show the functional relationship among the system
Reliability prediction is the process of calculating the anticipated system reliability from elements. To be able to predict reliability of a system, the system is represented as a number of
assumed component failure rates. Reliability calculation is an imprecise calculation but provides functional blocks that are interconnected according to effect. The interconnection of blocks can
a quantitative measure of how close a design comes to meeting design objectives and permits be one of the possible three types. However, there are cases when these interconnections may not
comparisons to be made between different design proposals. Reliability prediction is important represent a system.
for the following reasons.
a) It provides an early indication of a system’s potential to meet the design reliability 7.1.1 Series connected components
requirements. For such a system to work all components must work. For a system of two blocks failure of
b) It enables an assessment of life cycle costs to be carried out. either block prevents operation of the system.
c) It enables establishment of area which contribute major part system unreliability
d) It enables the achievement of a given availability.
e) It provides prior knowledge as to the magnitude of expected maintenance work load.
For the purpose of analysis a large and complex plant/facility is regarded as a hierarchy of parts Figure 7.2 Series connection
ranked according to their function and replaceability. The figure below shows how a
plant/facility can be divided to its basic components. 7.1.2 Parallel connected components
For such a system, system failure constitutes failure of all components. The system of two blocks
fails if both blocks fail.
Figure 7.3 Parallel connection
7.1.3 Composite connection
This is combination of series and parallel connections. For the system shown, the system fails if
block A fails or if both blocks B and C fail. Failure of B or C alone is not sufficient to cause
system failure.
Figure 7.1 Hierarchy of parts of a plant/facility
 
 3 4
Mulugeta D. Abera Mulugeta D. Abera
௧ ௡
ܴ௦௬ ሺ‫ݐ‬ሻ ൌ ‡š’൥െ න ෍ ߣ௜ ሺ‫ݐ‬ሻ݀‫ݐ‬൩
଴ ௜ୀଵ
For constant failure rate case, i.e. for the useful life of components where the failure rate remains
more or less constant,
ܴଵ ሺ‫ݐ‬ሻ ൌ ݁ ିఒభ ௧ ǡܴଶ ሺ‫ݐ‬ሻ ൌ ݁ ିఒభ ௧ ǡ ήήή  ܴ௡ ሺ‫ݐ‬ሻ ൌ ݁ ିఒ೙ ௧ 
And the system reliability is given by
Figure 7.4 Composite connection
ܴଵ ሺ‫ݐ‬ሻ ൌ ‡š’ሾെሺߣଵ ൅ ߣଶ ൅ήήή  ൅ߣ௡ ሻ‫ݐ‬ሿ
For identical-unit series system the reliability plot is indicated in the figure below
The approach to block diagram analysis requires establishment of failure criteria. Moreover, in
defining the blocks the following rules should be kept in mind among others.
1. Each block should represent the maximum number of components in order to simplify
the diagram,
2. The function of each block should be easily identified.
3. Blocks should be mutually independent in that failure in one block should not affect the
probability of failure in other blocks.
4. There should be only one environment within a block.
7.2 Reliability of series systems
Figure 7.5 Series system reliability for increasing n
For a series connected system the series system fails if any one of the subsystems fails. The
survival probability of a system consisting of n components in series is given by
For series connected units and constant failure rates, the system MTBF is given by
ܴ௦௬ ሺ‫ݐ‬ሻ ൌ ܴଵ ሺ‫ݐ‬ሻ ή ܴଶ ሺ‫ݐ‬ሻ ήήή ܴ௡ ሺ‫ݐ‬ሻ ൌ ෑ ܴ௧ ሺ‫ݐ‬ሻ ஶ ஶ ௡
ͳ
‫ ܨܤܶܯ‬ൌ න ܴሺ‫ݐ‬ሻ݀‫ ݐ‬ൌ න ݁‫ ݌ݔ‬൥െ ෍ ߣ௜ ‫ݐ‬൩ ݀‫ ݐ‬ൌ ௡
Where ܴ௦௬ ሺ‫ݐ‬ሻ ൌSurvival probability of system, and ଴ ଴ σ௜ୀଵ ߣ௜
௜ୀଵ
ܴ௜ ሺ‫ݐ‬ሻ ൌ Survival probability of a component 7.3 Reliability of parallel connected systems
The reliability of a component is given by Parallel-connected can be in parallel active redundancy where all systems are active in standby
௧
or partial active redundancy where system is allowed to function until a predetermined number
ܴ௜ ሺ‫ݐ‬ሻ ൌ ‡š’ሾെ න ߣ௜ ሺ‫ݐ‬ሻ݀‫ݐ‬ሿ
଴ of components are left for successful operation.
Hence, for the series system of n components, the system reliability is given by 7.3.1 All systems activated
௡
௧ In this type of parallel connection all components are functioning simultaneously. Consider the
ܴ௦௬ ሺ‫ݐ‬ሻ ൌ ‡š’൥െ ෍ න ߣ௜ ሺ‫ݐ‬ሻ݀‫ݐ‬൩
௜ୀଵ ଴
system shown in figure below
Or
 
 5 6
Mulugeta D. Abera Mulugeta D. Abera
Figure 7.6 Parallel connections of two components
There are three possible ways in which the system is able to operate
- Subsystem A and B are working Figure 7.7, Parallel system reliability for increasing n
- Subsystem A is working A has failed or
- Subsystem A has failed B is working For parallel connected system where all components are activated the MTBF is given by
The system fails if only both subsystems fail. ௡
ஶ ஶ
ͳ ͳ
Assuming the failure probabilities of components A and B to be ‫ܨ‬஺ ሺ‫ݐ‬ሻ and ‫ܨ‬஻ ሺ‫ݐ‬ሻ ‫ ܨܤܶܯ‬ൌ න ܴሺ‫ݐ‬ሻ݀‫ ݐ‬ൌ න ൣͳ െ ሺͳ െ ݁ ିఒ௧ ሻ௡ ൧ ݀‫ ݐ‬ൌ ෍
଴ ଴ ߣ ݅
௜ୀଵ
respectively, the probability that the system fails is given by
‫ܨ‬ሺ‫ݐ‬ሻ ൌ ‫ܨ‬஺ ሺ‫ݐ‬ሻ ή ‫ܨ‬஻ ሺ‫ݐ‬ሻ
7.3.2 Partial Active redundancy
Therefore the survival probability of the system is
In many cases of active redundancy the number of units permitted to fail before system
ܴሺ‫ݐ‬ሻ ൌ ͳ െ ‫ܨ‬ሺ‫ݐ‬ሻ
failure occurs is less than full redundancy. For example in a system of three units full
If the survival probabilities of the components are ܴ஺ ሺ‫ݐ‬ሻ and ܴ஻ ሺ‫ݐ‬ሻ
redundancy requires only one unit to function, where as partial redundancy would exist if
‫ܨ‬஺ ሺ‫ݐ‬ሻ ൌ ͳ െ ܴ஺ ሺ‫ݐ‬ሻ and ‫ܨ‬஺ ሺ‫ݐ‬ሻ ൌ ͳ െ ܴ஺ ሺ‫ݐ‬ሻ
two units were required with only one allowed to fail.
Therefore,
For three identical units, each with reliabilityܴ, and failure probability‫ ܨ‬ൌ ͳ െ ܴ, the
ܴሺ‫ݐ‬ሻ ൌ ͳ െ ൫ͳ െ ܴ஺ ሺ‫ݐ‬ሻ൯ ή ሺͳ െ ܴ஻ ሺ‫ݐ‬ሻሻ
binomial exapnasion of ሺܴ ൅ ‫ܨ‬ሻଷ yields the various probabilities of failure.
ܴሺ‫ݐ‬ሻ ൌ ܴ஺ ሺ‫ݐ‬ሻ ൅ ܴ஻ ሺ‫ݐ‬ሻ െ ܴ஺ ሺ‫ݐ‬ሻ ή ܴ஻ ሺ‫ݐ‬ሻ - ܴଷ gives the probability of zero failure
For n redundant components in parallel the survival probability of the system is given by - ͵ܴଶ Ǥ ሺͳ െ ܴሻ gives the probability of 1 failure
ܴ௦௬ ሺ‫ݐ‬ሻ ൌ ͳ െ ൫ͳ െ ܴଵ ሺ‫ݐ‬ሻ൯ ή ሺͳ െ ܴଶ ሺ‫ݐ‬ሻሻ ήήή ൫ͳ െ ܴ௡ ሺ‫ݐ‬ሻ൯ - ͵ܴሺͳ െ ܴሻଶ gives the probability of 2 failures
If ܴଵ ሺ‫ݐ‬ሻ ൌ ܴଶ ሺ‫ݐ‬ሻ ൌήήή ܴ௡ ሺ‫ݐ‬ሻ the system reliability is given by - ሺͳ െ ܴሻଷ gives the probability of 3 failures
ܴ௦௬ ሺ‫ݐ‬ሻ ൌ ͳ െ ܳሺ‫ݐ‬ሻ௡ Partial active redundancy is also known as r out of m unit network. At least r units out of m
Where: ܳሺ‫ݐ‬ሻ ൌ ͳ െ ܴሺ‫ݐ‬ሻ is the unreliability of each of the component in the parallel system. active units must function successfully for the system success. The system reliability is given
The reliability plot of identical-unit parallel system in which all units are activated is by
indicated in the figure below.
 
 7 8
Mulugeta D. Abera Mulugeta D. Abera
௠
݉ ௜ ௠ି௜
ii) The standby subsystems are assumed to have identical, constant failure rates to the
ܴ௦௬ ൌ ෍ ቀ ቁ ൫݁ ିఒ௧ ൯ ൫ͳ െ ݁ ିఒ௧ ൯
‫ݎ‬ main subsystem.
௜ୀ௥
݉ ௠Ǩ iii) The standby subsystems are assumed not to fail while in the idle state.
Whereቀ ቁ ൌ
‫ݎ‬ ௜Ǩሺ௠ି௜ሻǨ
iv) Defective subsystems are assumed to remain so. No repair is affected until the system
The MTBF of partially active redundancy systems is gives by
has failed.
ஶ ஶ ௠
݉ ௜ ௠ି௜ The Poisson distribution is used to determine the reliability of standby systems since they
‫ ܨܤܶܯ‬ൌ න ܴሺ‫ݐ‬ሻ ݀‫ ݐ‬ൌ න ൥෍ ቀ ቁ ൫݁ ିఒ௧ ൯ ൫ͳ െ ݁ ିఒ௧ ൯ ൩ ݀‫ݐ‬
଴ ଴
‫ݎ‬
௜ୀ௥ display the constant Ȝt characteristic of the Poisson process.
௠
ͳ ͳ If the probability of no failure is݁ ିఒ௧ , the probability of one failure is given by ߣ‫ݐ‬Ǥ ݁ ିఒ௧ , the
‫ ܨܤܶܯ‬ൌ ෍
ߣ ݅ ሺఒ௧ሻమ
௜ୀ௥ probability of two failures is Ǥ ݁ ିఒ௧ and so on. These terms may be arranged to represent
ଶ
the reliability of various standby configurations.
7.3.3 Parallel redundancy system in standby
For the arrangement where one subsystem operates and another is in a standby the reliability
In the standby configuration one or more subsystems are standing by to take over operation
of the system is given by:
upon failure of the basic unit. The standby units are not operative until a failure-sensing
ܴ௦௬ ൌ ݁ ିఒ௧ ൅ ߣ‫ି ݁ݐ‬ఒ௧
device senses the failure of the basic unit and turns on and connects the standby unit. Should
For a system with two standby units
failure take place in subsystem A, B is switched in its place and should B fail C is switched
ሺߣ‫ݐ‬ሻଶ ିఒ௧
on. ܴ௦௬ ൌ ݁ ିఒ௧ ൅ ߣ‫ݐ‬Ǥ ݁ ିఒ௧ ൅ ݁
ʹ
For a system with n standby units
ሺߣ‫ݐ‬ሻ௡ିଵ ିఒ௧
ܴ௦௬ ൌ ݁ ିఒ௧ ൅ ߣ‫ݐ‬Ǥ ݁ ିఒ௧ ൅ ‫ ڮ‬൅ Ǥ݁
ሺ݊ െ ͳሻǨ
When it cannot be assumed that the fault-sensing and switching device is 100% reliable, the
reliability of the system is affected by the switch reliability. Assuming that the switching
process doesn’t affect the basic unit for one standby unit, the system reliability is given by
ܴ௦௬ ൌ ݁ ିఒ௧ ൅ ܴ௦௪ Ǥ ߣ‫ݐ‬Ǥ ݁ ିఒ௧
Where ܴ௦௪ ൌReliability of fault-sensing and switching device
MTBF of standby systems
Figure 7.8 Redundant systems in standby ሺߣ‫ݐ‬ሻଶ ሺߣ‫ݐ‬ሻ௡ିଵ
ܴሺ‫ݐ‬ሻ ൌ ݁ ିఒ௧ ቈͳ ൅ ߣ‫ ݐ‬൅ ൅ ‫ڮ‬൅ ቉
ʹ ሺ݊ െ ͳሻǨ
ஶ
In parallel systems in standby, the following assumptions are initially made.
‫ ܨܤܶܯ‬ൌ න ܴሺ‫ݐ‬ሻ ݀‫ݐ‬
i) The means of sensing that failure has occurred and for switching from the defective to ଴
ஶ
the standby is assumed to be failure free. ሺߣ‫ݐ‬ሻଶ ሺߣ‫ݐ‬ሻ௡ିଵ
‫ ܨܤܶܯ‬ൌ න ݁ ିఒ௧ ቈͳ ൅ ߣ‫ ݐ‬൅ ൅ ‫ڮ‬൅ ቉ ݀‫ݐ‬
଴ ʹ ሺ݊ െ ͳሻǨ
 
 9 10
Mulugeta D. Abera Mulugeta D. Abera
݊
‫ ܨܶܶܯ‬ൌ ‫ ܨܤܶܯ‬ൌ
ߣ
To ensure the reliability of the system, i.e. for the system to operate successfully at least one of
The MTBF of a system consisting of the basic unit and one standby unit is
the paths AA’ or BB’ should function. However,, since neither A nor B is fully reliable, a third
ʹ
‫ ܨܤܶܯ‬ൌ element C is added to supply either A’ or B’. Hence four paths are possible: AA’, BB’, CA’ or
ߣ
The ‫ ܨܤܶܯ‬of a system with two standby units is CB’.
͵ System reliability for this arrangement can be calculated from Baye’s theorem.
‫ ܨܤܶܯ‬ൌ
ߣ ܴ௦௬ ൌ ͳ െ ܴ௖ ሺͳ െ ܴ஺ᇲ ሻሺͳ െ ܴ஻ᇲ ሻ െ ሺͳ െ ܴ௖ ሻሺͳ െ ܴ஺ ܴ஺ᇲ ሻሺͳ െ ܴ஻ ܴ஻ᇲ ሻ
In general for n-1 standby units ready to take over if the basic should fail, the MTBF of the This system reliability is arrived at the following procedure. The system success requires at least
system is one of the paths AA’, BB’, CA’ or CB’ to be successful.
݊
‫ ܨܤܶܯ‬ൌ ܴ௦௬௦Ȁ஼௦ ൌ ሾͳ െ ሺͳ െ ܴ஺ᇲ ሻሺͳ െ ܴ஻ᇲ ሻሿܴ௖
ߣ
With C unsuccessful
7.4 Neither series nor parallel connections ܴ௦௬௦Ȁ஼௨௦ ൌ ሾͳ െ ሺͳ െ ܴ஺ ܴ஺ᇲ ሻሺͳ െ ܴ஻ ܴ஻ᇲ ሻሿሺͳ െ ܴ௖ ሻ
In calculating system reliability, there are cases in which the system configuration cannot be Hence, system reliability is given by the sum of the above reliability terms
reduced to simple series and parallel models or any combination of these. For such situations ܴ௦௬௦ ൌ ܴ௦௬௦Ȁ஼௦ ൅ ܴ௦௬௦Ȁ஼௨௦
where uncertainty exists to the configuration of the system, Baye’s theorem can be helpful to Substituting and simplifying yields
determine the system reliability. The use of Baye’s theorem is demonstrated as follows. ܴ௦௬௦Ȁ௨௦ ൌ ͳ െ ܴ௖ ሺͳ െ ܴ஺ᇲ ሻሺͳ െ ܴ஻ᇲ ሻ െ ሺͳ െ ܴ௖ ሻሺͳ െ ܴ஺ ܴ஺ᇲ ሻሺͳ െ ܴ஻ ܴ஻ᇲ ሻ
Consider the block diagram shown in figure below. The connection of the blocks in this case is
neither parallel nor series. 7.5 Maintainability prediction
In assigning any maintenance workload to a mechanic or a technician it is not important to assign
the job that has to be done, it is also important to specify the time during which the job has to be
accomplished. Doing this will help optimize the resource allocation and will help eliminating
idle man-hour, idle machine time and so on. Hence maintenance engineers should engage
themselves in determining the time which a specific maintenance job is going to take. This is
done by determining the maximum repair time, the median repair time and the maintainability
probability. To this end determining the maintainability function plays an important role.
7.6 Maximum Repair time
Maximum repair time is the repair maintenance (restoration, replacement) time ‫ݐ‬ெ೘ೌೣ which
should not be exceeded or exceeded only with a small probability.
For example,
Figure 7.9 Neither series nor parallel conncetion
 
 11 12
Mulugeta D. Abera Mulugeta D. Abera
If ‫ݐ‬ெ೘ೌೣ ൌ ʹǤ͵Ͳʹ͸‫ܴܶܶܯ‬ǡ then Given‫ܯ‬ሺ‫ݐ‬ሻ and ݃ሺ‫ݐ‬ሻ, to determine the non-constant failure rate ߤሺ‫ݐ‬ሻ multiply the equation for
‫ܯ‬൫‫ݐ‬ெ೘ೌೣ ൯ ൌ ͲǤͻͲ ‫ܯ‬ሺ‫ݐ‬ሻ by ߤሺ‫ݐ‬ሻ to get
೟
Or, ߤሺ‫ݐ‬ሻ݁ ି ‫׬‬బ ఓሺ௧ሻௗ௧ ൌ ሾͳ െ ‫ܯ‬ሺ‫ݐ‬ሻሿߤሺ‫ݐ‬ሻ
‫ܯ‬൫‫ݐ‬ெ೘ೌೣ ൯ ൌ ͳ െ ߙ ൌ ͲǤͻͲ ݃ሺ‫ݐ‬ሻ ൌ ሾͳ െ ‫ܯ‬ሺ‫ݐ‬ሻሿߤሺ‫ݐ‬ሻ
In general, from which the repair rate ߤሺ‫ݐ‬ሻ is evaluated to be
௧ಾ೘ೌೣ
ି൬ ெ்்ோ ൰
݃ሺ‫ݐ‬ሻ
‫ܯ‬൫‫ݐ‬ெ೘ೌೣ ൯ ൌ ͳ െ ݁ ൌͳെߙ ߤሺ‫ݐ‬ሻ ൌ
ͳ െ ‫ܯ‬ሺ‫ݐ‬ሻ
From which we obtain
‫ݐ‬ெ೘ೌೣ
݈݊ሺߙሻ ൌ െ 7.8 Maintainability for a given restoration time with a weibull time to restore distribution
‫ܴܶܶܯ‬
In the maintainability function
Or,
௧భ
‫ݐ‬ெ೘ೌೣ ‫ܯ‬ሺ‫ݐ‬ଵ ሻ ൌ ܲሺ‫ ݐ‬൑ ‫ݐ‬ଵ ሻ ൌ න ݃ሺ‫ݐ‬ሻ݀‫ݐ‬
‫ ܴܶܶܯ‬ൌ െ  ଴
݈݊ሺߙሻ
6.7 The case of ߤሺ‫ݐ‬ሻnot constant for ߤሺ‫ݐ‬ሻ not constant, the ‫݂݀݌‬is given by The weibull times to restore probability density function is
ഁ
೟ ߚ ‫ ݐ‬െ ߛ ఉିଵ ିቀ௧ିఊ ቁ
݃ሺ‫ݐ‬ሻ ൌ ߤሺ‫ݐ‬ሻ݁ ି ‫׬‬బ ఓሺ௧ሻௗ௧ Ǥ ݀‫ݐ‬ ݃ሺ‫ݐ‬ሻ ൌ ൬ ൰ Ǥ݁ ఎ
ߟ ߟ
and the maintainability probability function is
From which the maintainability function is obtained to be
௧
‫ܯ‬ሺ‫ݐ‬ሻ ൌ െ න ݃ሺ‫ݐ‬ሻ݀‫ݐ‬ ௧ ିఊ ഁ
ିቀ భఎ ቁ
଴ ‫ܯ‬ሺ‫ݐ‬ଵ ሻ ൌ ͳ െ ݁
௧ ೟ In weibull times to restore distribution
‫ܯ‬ሺ‫ݐ‬ሻ ൌ െ න ߤሺ‫ݐ‬ሻ݁ ି ‫׬‬బ ఓሺ௧ሻௗ௧ Ǥ ݀‫ݐ‬
଴ - Ș is the characteristic time to restore
Let - ߚ is a measure of the consistence of restoration
௧
െ න ߤሺ‫ݐ‬ሻǤ ݀‫ ݐ‬ൌ ߥ ՜ െ ߤሺ‫ݐ‬ሻǤ ݀‫ ݐ‬ൌ ݀ߥ - ߛ is the minimum time known as the threshold time
଴
For the weibull distribution, the time to restore ‫ݐ‬ଵ for a desired maintainability probability ‫ܯ‬ሺ‫ݐ‬ଵ ሻ
then the maintainability probability function is given by
is obtained by solving the above equation for‫ݐ‬ଵ .
௧
‫ܯ‬ሺ‫ݐ‬ሻ ൌ െ න ݁ ఔ Ǥ ݀‫ݐ‬ ‫ݐ‬ଵିఊ ఉ
଴ െ݈݊ሾͳ െ ‫ܯ‬ሺ‫ݐ‬ሻሿ ൌ ൬ ൰
ߟ
‫ݐ‬
‫ܯ‬ሺ‫ݐ‬ሻ ൌ െ݁ ఔ ȁ From which we obtain
Ͳ
೟ ‫ݐ‬ ଵൗ
‫ܯ‬ሺ‫ݐ‬ሻ ൌ ݁ ି ‫׬‬బ ఓሺ௧ሻௗ௧ ȁ ‫ݐ‬ଵ ൌ ߛ ൅ ߟሼെ݈݊ሾͳ െ ‫ܯ‬ሺ‫ݐ‬ሻሿሽ ఉ
Ͳ
Upon simplification, the maintainability function is obtained to be As an example consider the case for which‫ܯ‬ሺ‫ݐ‬ଵ ሻ ൌ ͻͷΨ. Then,
೟
ଵൗ
‫ܯ‬ሺ‫ݐ‬ሻ ൌ ͳ െ ݁ ି ‫׬‬బ ఓሺ௧ሻௗ௧ ‫ݐ‬ଵ ൌ ߛ ൅ ߟሼെ݈݊ሾͳ െ ͲǤͻͷሿሽ ఉ
Or,
 
 13
Mulugeta D. Abera
ଵൗ
‫ݐ‬ଵ ൌ ߛ ൅ ߟሺʹǤͻͻͷ͹͵ሻ ఉ
The median time to restore ‫ݐ‬ଵ is obtained for ‫ܯ‬ሺ‫ݐ‬ଵ ሻ ൌ ͷͲΨ and is obtained from
ଵൗ
‫ݐ‬ଵ ൌ ߛ ൅ ߟሺͲǤ͸ͻ͵ͳͷሻ ఉ