5/9/2023

Distribution Reliability
Part 4 (/6)
Angel A. Recalde
Ph.D. (U. Nottingham), M. Sc. (U. Queensland), B. Eng. (ESPOL)
Master in Electrical Engineering Program FIEC 2022
ESPOL – FIEC
GYE, ECU
May – June 2023
INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL 1

Warning
• All material in this course is copyrighted and
must never be used outside the Master of
Electrical Engineering program of the FIEC-ESPOL
• This material must not be shared to other
universities, public institutions, private
courses, or undergrad programs of any public or
private universities (except FIEC-ESPOL Master’s
program).
• This material must not be used for any purpose
other than academic and learning
• This material is appropriately referenced to
their corresponding authors
• The use of all material is uniquely intended for
academic illustration and learning purposes only

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

2
2

1
 5/9/2023

Layout

• Introduction

• Reliability Assessment
• Reliability Indices
• Computational Analysis of typical distribution system
• Probability distribution of reliability indices

• Conclusions

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

3
3

Introduction
https://www.azcentral.com/story
/news/local/phoenix/2017/05/2
3/mylar-balloons-can-cause-
power-failure/338000001/

Reliability is the most important performance index in any system

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL 4

2
 5/9/2023

Introduction
• Power system is extremely complex due to:
• Physical size
• Widely dispersed geography
• Interconnections
• Operator’s own restrictions (jurisdiction, regulations, organization)
• Vertically integrated utility
• Liberalized market
• Electric energy cannot be stored efficiently and effectively in large quantities
• Unpredictable system behaviour
• Climate change (not considered few decades ago)
• Huge fires in rural areas
• Extreme cold weathers
• Floods, river overflows

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

5
5

Introduction
• Reliability related to the ability of a system to deliver electricity to all
points of utilization within the accepted standard and desired amount

• Reliability assessment includes

• Adequacy assessment
• Security assessment

• Distribution System can be designed to have maximum reliability

• Reliability can be maximized in an existing distribution network

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

6
6

3
 5/9/2023

Introduction
• Adequacy assessment
• Studies the existence of sufficient facilities within the system to meet the
system demand. It is mainly performed in the design phase
• Security assessment
• Studies the ability of a system to respond to disturbances and this is
associated with the dynamic responses. Mainly performed during the
operational phases.

• Distribution System can be designed to have maximum reliability

• Reliability can be maximized in an existing distribution network

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

7
7

Introduction
• Reliability at what cost?
• Probability of outage is reduced by increasing investment during the
planning phases
• But overinvestment leads to inefficiency and higher tariff
• In a market environment, cost of supply and reliability are to be
traded off during planning and operational phases.

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

8
8

4
 5/9/2023

Reliability assessment
Techniques

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

9
9

Reliability assessment
• Deterministic techniques
• Traditional techniques
• Do not address the probabilistic/stochastic nature of power system

• Probabilistic techniques
• Analytical and simulation approaches
• Can best address the stochastic nature of power system
• Different computational tools and techniques are in the market
• Legacy software
• Novel scientific approaches (academic research)
• Others for specific applications (cases)

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

10
10

5
 5/9/2023

Probabilistic approach
• Nature of power system is stochastic
• Distribution failure is a function of size (length of lines), design
(types of components, topology), location (influence of
surroundings), and environment (weather, human behavior)
• Uncertainties in system load and load forecasting
• Changing scenario of markets
• Deregulation and privatization
• Presence of market forces
• Problems like lack of data, computational limitations, and assessment
techniques are no more barrier

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

11
11

Probabilistic reliability assessment

• Analytical Technique
• Based on the state enumeration principle
• Represent the actual system by mathematical model
• Difficult to implement in large power systems

• Simulation technique
• Monte Carlo simulation (for stochastic system)
• Evaluates the system indices by simulating the actual processes and random
behavior of the system’s components
• Handles large physical system
• Requires large computing time and storage capacity

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

12
12

6
 5/9/2023

Reliability assessment indices

• Reliability indices are statistical aggregations of reliability
data for a well-defined set of loads, components, or
customers.
• Most reliability indices are average values of a particular
reliability characteristic for an entire system, operating
region, substation service territory, or feeder.
• Reliability index definitions follow the recently adopted IEEE
standard 1366-2003.

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

13
13

Reliability assessment indices

• Basic Reliability Indices
System performance indices

• Customer-based reliability indices (SAIFI, SAIDI, etc.)

• Load/energy-oriented reliability indices.

• Can be used for effectively predicting the severity of

system in the future
• Are used as the means of assessing the past performance
of the system
INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL
14
14

7
 5/9/2023

Reliability Indices
Indices for the system’s assessment of reliability

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL 15

15

Reliability Indices

• Basic reliability indices

• Customer-based reliability indices
• Load/energy-oriented reliability indices

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

16
16

8
 5/9/2023

Basic reliability indices

• Average failure rate (failure per year)
• These indices do not have
𝜆 = 𝜆 deterministic values, they
are only the long term
• Average annual outage time (hr per year) expected or average values
with the underlying
𝑈 = 𝜆𝑟 probability distribution
• Fail to represent the
complete system behavior
• Average outage time (hr/failure) and response (No. of
customers affected, size of
𝑈 ∑ 𝜆𝑟 load/MW curtailment
𝑟 = =
𝜆 ∑ 𝜆
INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL
17
17

Customer–based reliability indices

• System Average Interruption Frequency Index
𝑇𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟 𝑖𝑛𝑡𝑒𝑟𝑟𝑢𝑝𝑡𝑖𝑜𝑛𝑠 ∑ 𝜆 𝑁
𝑆𝐴𝐼𝐹𝐼 = = ≤1
𝑇𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠 𝑠𝑒𝑟𝑣𝑒𝑑 ∑ 𝑁

𝜆 𝑁 : total number of customers interrupted/curtailed of load point i

𝑁 : total number of customers of load point i

𝑇𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠 𝑑𝑖𝑠𝑐𝑜𝑛𝑛𝑒𝑐𝑡𝑒𝑑 ∑ 𝑁

𝑆𝐴𝐼𝐹𝐼 = =
𝑇𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠 𝑠𝑒𝑟𝑣𝑒𝑑 ∑ 𝑁

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

18
18

9
 5/9/2023

Customer–based reliability indices

• Customer Average Interruption Frequency Index
𝑡𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟 𝑖𝑛𝑡𝑒𝑟𝑟𝑢𝑝𝑡𝑖𝑜𝑛𝑠 ∑ 𝜆𝑁
𝐶𝐴𝐼𝐹𝐼 = = ≥1
𝑡𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟 𝑎𝑓𝑓𝑒𝑐𝑡𝑒𝑑 ∑ 𝑁
∑ 𝑁
𝐶𝐴𝐼𝐹𝐼 ≥ 𝑆𝐴𝐼𝐹𝐼 𝐶𝐴𝐼𝐹𝐼 =
𝑁
• CAIFI is useful to compare performance for one calendar year with another
year, as the denominator (customer affected number) can be different for
different years
• One affected customer is counted once only (irrespective of interruption
numbers)
• CAIFI is useful to study the chronological trends of distribution reliability
INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL
19
19

Customer–based reliability indices

• System Average Interruption Duration Index
𝑠𝑢𝑚 𝑜𝑓𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟 𝑖𝑛𝑡𝑒𝑟𝑟𝑢𝑝𝑡𝑖𝑜𝑛 𝑑𝑢𝑟𝑎𝑡𝑖𝑜𝑛𝑠 ∑ 𝑈 𝑁
𝑆𝐴𝐼𝐷𝐼 = =
𝑡𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠 ∑ 𝑁
∑ 𝑁 𝑑
𝑁 : total number of customers of load point i 𝑆𝐴𝐼𝐷𝐼 =
∑ 𝑁
• Customer Average Interruption Duration Index
𝑠𝑢𝑚 𝑜𝑓𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟 𝑖𝑛𝑡𝑒𝑟𝑟𝑢𝑝𝑡𝑖𝑜𝑛 𝑑𝑢𝑟𝑎𝑡𝑖𝑜𝑛𝑠 ∑ 𝑈𝑁
𝐶𝐴𝐼𝐷𝐼 = =
𝑡𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠 𝑖𝑛𝑡𝑒𝑟𝑟𝑢𝑝𝑡𝑖𝑜𝑛𝑠 ∑ 𝜆 𝑁
∑ 𝑁 𝑑
𝐶𝐴𝐼𝐷𝐼 =
∑ 𝑁
INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL
20
20

10
 5/9/2023

Customer–based reliability indices

• Average Service Availability /Unavailability Index

𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟 ℎ𝑜𝑢𝑟𝑠 𝑜𝑓 𝑎𝑣𝑎𝑖𝑙𝑎𝑏𝑙𝑒 𝑠𝑒𝑟𝑣𝑖𝑐𝑒 ∑ 𝑁 × 8760 − ∑ 𝑈 𝑁

𝐴𝑆𝐴𝐼 = =
𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠 ℎ𝑜𝑢𝑟𝑠 𝑑𝑒𝑚𝑎𝑛𝑑𝑒𝑑 ∑ 𝑁 × 8760

∑ 𝑈𝑁
𝐴𝑆𝑈𝐼 = 1 − 𝐴𝑆𝐴𝐼 =
∑ 𝑁 × 8760

∑ 𝑁 × 8760 − ∑ 𝑁 𝑑
𝐴𝑆𝐴𝐼 =
∑ 𝑁 × 8760

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

21
21

Load/Energy-oriented reliability indices

• Calculation of load/energy-oriented indices requires the
value of load curtailment LC and outage duration d for each
load point or busbars
• Probabilistic values for LC and d are the load bus indices LOLE
(Load of Load Expectation) and LOLP (Loss of Load
Probability) obtained/calculated from the composite
reliability assessment
• LOLE: Expected number of hours per year that a country’s
electricity production park cannot meet its demand
• For simplicity, deterministic values of LC is taken as the
average load at each load point/busbar La, and Ui is taken as
outage duration di.
INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL
22
22

11
 5/9/2023

Load/Energy-oriented reliability indices

• Value of La can be calculated by the following two
approaches:
𝐿 =𝐿 ×𝑓

𝐿 is the peak load demand, 𝑓 is the load factor for each bus

𝑡𝑜𝑡𝑎𝑙 𝑒𝑛𝑒𝑟𝑔𝑦 𝑑𝑒𝑚𝑎𝑛𝑑𝑒𝑑 𝑖𝑛 𝑝𝑒𝑟𝑖𝑜𝑑 𝑜𝑓 𝑖𝑛𝑡𝑒𝑟𝑒𝑠𝑡 𝐸

𝐿 = =
𝑝𝑒𝑟𝑖𝑜𝑑 𝑜𝑓 𝑖𝑛𝑡𝑒𝑟𝑒𝑠𝑡 𝑡

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

23
23

Load/Energy-oriented reliability indices

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

24
24

12
 5/9/2023

Load/Energy-oriented reliability indices

• Energy Not Supplied (ENS)
𝐸𝑁𝑆 = 𝐿 ,𝑈

𝐿 , is the average load connected at load point i

• Average Energy Not Supplied (AENS)

𝑡𝑜𝑡𝑎𝑙 𝑒𝑛𝑒𝑟𝑔𝑦 𝑛𝑜𝑡 𝑠𝑢𝑝𝑝𝑙𝑖𝑒𝑑 ∑ 𝐿 ,𝑈

𝐴𝐸𝑁𝑆 = =
𝑡𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠 𝑠𝑒𝑟𝑣𝑒𝑑 ∑ 𝑁

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

25
25

Load/Energy-oriented reliability indices

• Average Customer Curtailment Index

𝑡𝑜𝑡𝑎𝑙 𝑒𝑛𝑒𝑟𝑔𝑦 𝑛𝑜𝑡 𝑠𝑢𝑝𝑝𝑙𝑖𝑒𝑑

𝐴𝐶𝐶𝐼 =
𝑡𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠 𝑎𝑓𝑓𝑒𝑐𝑡𝑒𝑑

• ACCI is a useful index (like CAIFI) to monitor the

chronological changes in average energy not supplied

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

26
26

13
 5/9/2023

Reliability assessment
• Application of system performance Indices:
• To Assess the chronological changes in system performance and to
identify the weak area of system for reinforcement
• To Use the existing indices as a guide of acceptable values for
future reliability assessment, and
• To Compare previous predictions with the actual operating
experiences

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

27
27

Reliability assessment
• Customer Related Indices are the most widely used indices in
the industry (EPRI)
• Among the most used: SAIDI, SAIFI, ASAI

See MATLAB example:

Distribution_Reliability_basic_example_1.m

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

28
28

14
 5/9/2023

Exercise
Handwritten exercise

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL 29

29

Exercise (by hand)

• Simple 3-load point radial system. Calculate SAIDI, SAIFI, ASAI
A B C

L1 L2 L3

Load 𝒇 𝒉𝒓 𝒇 Load No. Avg. Load

𝝀𝑳 ( ) 𝒓𝑳 (𝒉𝒓) 𝑼𝑳 ( ) Line 𝝀 ( ) 𝒓 (𝒉𝒓) customers
point 𝒚𝒓 𝒚𝒓 𝒚𝒓 point demand (kW)

L1 0.20 6.0 1.2 A 0.20 6.0 L1 200 1000

L2 0.30 5.7 1.7 B 0.10 5.0 L2 150 700

L3 0.45 6.4 2.9 C 0.15 8.0 L3 100 400

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

30
30

15
 5/9/2023

Computational analysis
Analysis of the typical distribution system

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL 31

31

Typical Radial Distribution Network

• Radial system
1 2 3 4

a b c d

A B C D

Length 𝒇 Distri- Length 𝒇 Load No. Avg. load

Section 𝝀 ( ) 𝒓 (𝒉𝒓) 𝝀 ( ) 𝒓 (𝒉𝒓) customers
(km) 𝒚𝒓. 𝒌𝒎 butor (km) 𝒚𝒓. 𝒌𝒎 point demand (kW)

1 2 0.1 4 a 1 0.2 2 A 1000 5000

2 1 0.1 4 b 3 0.2 2 B 800 4000

3 3 0.1 4 c 2 0.2 2 C 700 3000

4 2 0.1 4 d 1 0.2 2 D 500 2000

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

32
32

16
 5/9/2023

Application to radial systems

• CASE 1 (Base Case)
• Failure rates are proportional to the line length
• Assumptions:
• Any fault for 1ph or 3ph will trip all 3 phases
• All component failures are short circuits, then, each failure will
cause the main breaker to operate
• There are no arrangements for isolating the failure section, then
each failure must be repaired before the main breaker can be
reclosed.

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

33
33

Application to radial systems

• CASE 1 (Base Case)

See MATLAB example:

Distribution_Reliability_example_2_base_case.m

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

34
34

17
 5/9/2023

Application to radial systems

• CASE 2 (effect of lateral distributor protection)
• Assumptions:
• Base case is modified by installing fusegear at the tree point in
each lateral branch
• A shortcircuit on any of the lateral distributor causes its
appropriate fuse to blow and disconnect (other load points are
intact)
• Still, shortcircuits in main sections causes the main breaker to
operate

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

35
35

Application to radial systems

• CASE 2 (effect of lateral distributor protection)

See MATLAB example:

Distribution_Reliability_example_2_lat_fusegear.m

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

36
36

18
 5/9/2023

Application to radial systems

• CASE 3 (effect of lateral distributor protection +
disconnecting switches –effect of disconnects-)
• Assumptions:
• Start with CASE 2
• Provision of disconnects at judicious points along the main feeder
• These switches do not open during the fault, the main breaker still
operates if such fault occurs along the lateral end parts
• Once the fault is detected, corresponding disconnect switch is
opened and the breaker is reclosed.
• Let the total isolation and switching time be 0.5 hr.

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

37
37

Application to radial systems

• CASE 3 (effect of lateral distributor protection +
disconnecting switches –effect of disconnects-)

See MATLAB example:

Distribution_Reliability_example_2_disconnects.m

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

38
38

19
 5/9/2023

Application to radial systems

• CASE 4 (effect of lateral distributor protection +
disconnecting switches –effect of disconnects- + effect of
protection failure)
• Assumptions:
• Start with CASE 3
• In real world system, primary protection may fail to operate and,
in such condition, back up protection operates.
• Study the impact that protection failure will have on the reliability
indices

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

39
39

Application to radial systems

• CASE 4
• Assume fusegear reliability is 0.9 (operates 9 times
successfully out of 10 faults)
• Failures of fusegear in distributors ‘b’, ‘c’, and ‘d’ contribute
to load point ‘A’; similarly, loads ‘B’, ‘C’, and ‘D’ are affected.
• Conditional probability theory:
𝑓 𝑟𝑎𝑡𝑒 = 𝑓 𝑟𝑎𝑡𝑒|𝑓𝑢𝑠𝑒 𝑜𝑝𝑒𝑟𝑎𝑡𝑒𝑠 × 𝑃 𝑓𝑢𝑠𝑒 𝑜𝑝𝑒𝑟𝑎𝑡𝑒𝑠 + 𝑓 𝑟𝑎𝑡𝑒|𝑓𝑢𝑠𝑒 𝑓𝑎𝑖𝑙𝑠 × 𝑃 𝑓𝑢𝑠𝑒 𝑓𝑎𝑖𝑙𝑠

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

40
40

20
 5/9/2023

Application to radial systems

• CASE 4 (effect of lateral distributor protection +
disconnecting switches –effect of disconnects- + effect of
protection failure)

See MATLAB example:

Distribution_Reliability_example_2_protectionfail.m

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

41
41

Effect of transferring loads

• Open points can be moved in order to recover load that has
been disconnected
• Many radial distribution networks have the provision of
disconnect switches which are kept normally open (N/O),
and during the event of a fault, such switch is closed, and
another switch is made N/O switch.
• This operational procedure improves the load point
reliability.

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

42
42

21
 5/9/2023

Effect of transferring loads

• No transfer restrictions:
• All load can be transferred to another feeder
• System in Case 3 is modified where line section 4 can be
connected or disconnected to other system via N/O switch
• CASE 5 (transfer no restriction)
See MATLAB example:
Distribution_Reliability_example_2_norestrictransf.m

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

43
43

Effect of transferring loads

• Transfer restrictions:
• There cannot be always possible of unrestricted transfer of load
from one system to another system through N.O. switches
• There can be two possible conditions:
• If load is transferable, outage time due to failure event = isolation time
• If load is not transferable, outage time due to failure event = repair time
• The average of these values can be evaluated using the concept of
expectation as following:

𝑜𝑢𝑡𝑎𝑔𝑒 𝑡𝑖𝑚𝑒 = 𝑜𝑢𝑡𝑎𝑔𝑒 𝑡𝑖𝑚𝑒|𝑡𝑟𝑎𝑛𝑠𝑓𝑒𝑟𝑠 × 𝑃 𝑡𝑟𝑎𝑛𝑠𝑓𝑒𝑟 + 𝑜𝑢𝑡𝑎𝑔𝑒 𝑡𝑖𝑚𝑒|𝑛𝑜 𝑡𝑟𝑎𝑛𝑠𝑓𝑒𝑟 × 𝑃 𝑛𝑜 𝑡𝑟𝑎𝑛𝑠𝑓𝑒𝑟

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

44
44

22
 5/9/2023

Effect of transferring loads

• Transfer restrictions
• System in Case 3 is modified where line section 4 can be
connected or disconnected to other system via N/O switch
• CASE 6 (transfer restriction)
• It may be preferable to consider the amount of load that can
be recovered based on the load that has been disconnected
and the available transfer capacity of the second system at
that particular loading level on the system.
• This requires a more exhaustive analysis.

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

45
45

Effect of transferring loads

• CASE 6 (transfer restriction)

See MATLAB example:

Distribution_Reliability_example_2_restrictransf.m

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

46
46

23
 5/9/2023

Probability distribution of
reliability indices
Non-average failure rates and restoration times

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL 47

47

Probability distribution of reliability indices

• The reliability indices calculated in the previous examples are
based on the average values of basic indices
• Due to random nature of failure and repair, these values
could be different in different years
• Such deviation are represented by the concept of probability
distribution

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

48
48

24
 5/9/2023

Probability distribution of reliability indices

• Failure Rate
• Failure times are generally assumed to be exponentially
distributed
• Components operate in their useful or operating life period
• System failure rate for radial networks depends only on the component
failure rates and not the restoration time
• Non exponential failure rate do not affect the failure rate distribution
• Based on the following assumptions, load point failure rate of a
radial network obeys a Poisson distribution.
• The probability of n failure in time is:
𝜆𝑡 𝑒
𝑃 𝑛 =
𝑛!
INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL
49
49

Probability distribution of reliability indices

• Using the above equation, P(n), i.e., the probability of any
number of failures/yr at each load point can be calculated
using the given average 𝜆

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

50
50

25
 5/9/2023

Probability distribution of reliability indices

• Restoration time
• Load point outage duration can be approximated by Gamma
distribution if the restoration times are exponentially distributed.
• However, in real system, restoration times are not exponentially
distributed.
• When non-exponential distributions are used to represent the
restoration times, the load point duration cannot be generally
represented by Gamma distribution and also becomes difficult to
fit to any defined distribution.
• In such conditions, Monte Carlo Simulation gives the correct
solution.
• Still, without knowing the correct distribution, the average values
of outage duration evaluated using above are valid, because it is
only the distribution around these average that are affected.
INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL
51
51

Conclusions
Conclusion of reliability analysis (first section)

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

52
52

26
 5/9/2023

Conclusions
• The technique described above are applicable for reliability
assessment of radial and meshed networks while operating
as a single radial system
• For more complex system like meshed and parallel systems,
the above techniques needs to be extended.
• To assess the realistic reliability of distribution system, it
requires thorough understanding of relevant operational
characteristics and policies

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

53
53

References

• Kersting, Distribution system modeling and analysis, CRC, 2018

• Short, Electric power distribution handbook, CRC, 2017
• Billinton, Reliability evaluation of power systems, Springer, 1996
• Billinton, Reliability evaluation of Engineering systems, Springer,
1992

INGP1061, Angel A. Recalde ©, May-June 2023, FIEC-ESPOL

54
54

27