Degree project in
and Isolation in Active Distribution Networks
CHAITANYA DESHPANDE
Stockholm, Sweden 2015
XR-EE-ICS 2015:004
ICS
Masterthesis,
Liberalized electricity markets, increased awareness of clean energy resources and their decreasing costs have resulted in large numbers of distributed power generators being installed on distribution network. Installation of distributed generation has altered the passive nature of distribution grid. A concept of Active Distribution Network is proposed which will enable present day infrastructure to host renewable energy resources reliably.
Fault management that includes fault localization, isolation and service restoration is part of active management of distribution networks.
This thesis aims to introduce a distributed protection methodology for fault localization and isolation. The objective is to enhance reliability of the network. Faults are identi- fied based on root mean square values of current measurements and by comparing these values with preset thresholds. The method based on multi-agent concept can be used to locate the faulty section of a distribution network and for selection of faulty phases.
The nodal Bus Agent controls breakers that are associated with it. Based on indication of fault, adjacent bus Agents communicate with each other to identify location of fault.
A trip signal is then issued to corresponding Breakers in adjacent Bus Agents, isolating the faulty section of line. A case study was carried out to verify suitability of the pro- posed method. A meshed network model and multi-agent based protection scheme was simulated in Simulink SimPowerSystems. Considering nature of Distribution Network, separate breakers for each phase are considered. The distribution network protection system identified fault introduced in the network correctly along with interrupting the fault current.
Keywords
Distributed Energy Resources, Active Distribution Network, Fault Localization and Iso-
lation, Multi-Agent, MATLAB-SimPowerSystems, Distribution Network Protection.
Acknowledgements
I would like express my deep and sincere gratitude to my supervisors PhD students Yiming Wu and Mikel Armendariz Perez, their vast knowledge and guidance helped me get through toughest of situations. I would also like to thank teachers at ICS; Arshad, Nicholas and Davood, for their guidance during courses and in completion of this thesis.
I would like express gratitude towards Prof Lars Nordstrom for agreeing to be examiner to this thesis.
I would also like to thank fellow students Farhad, Buland, Clipton for their support.
Last but not least, I thank Annica for helping around with bureaucratic routine.
Chaitanya Deshpande
Abstract
1
Acknowledgements
2
Contents 3
List of Figures 6
List of Tables 10
1 Introduction 11
1.1 Background . . . 11
1.2 Thesis Statement . . . 12
1.3 Contribution of Thesis . . . 12
1.4 Organization of Thesis . . . 12
2 Related Work 13 2.1 Distributed Generation . . . 13
2.1.1 Wind Farms . . . 14
2.1.2 Solar Photovoltaic . . . 15
2.1.3 Micro Turbine . . . 15
2.1.4 Fuel Cells . . . 15
2.2 Impact of DG on Distribution Network . . . 15
3
2.3 Impact on Distribution Network Protection . . . 17
2.4 Active Distribution Networks . . . 18
2.4.1 ICT Infrastructure . . . 18
2.4.2 Active Resources and Aggregators . . . 18
2.4.3 Active Network Management . . . 19
2.5 Faults in Distribution Network . . . 19
2.5.1 Fault . . . 20
2.5.2 Single Line to Ground Fault . . . 21
2.5.3 Line to Line Fault . . . 23
2.5.4 Three Phase to Ground Fault . . . 24
2.5.5 Consequences of Faults . . . 25
2.5.6 Fault Clearing . . . 26
2.5.7 Current Limiters . . . 27
2.5.8 Fault Clearing System . . . 27
2.5.8.1 Relay Protection System . . . 27
2.5.8.2 Circuit Breaker . . . 28
2.5.9 Protection Principles . . . 28
2.5.10 Fault Identification . . . 28
2.6 Multi-Agent Concept . . . 30
3 Methodology Employed 32 3.1 Distribution Network . . . 32
3.2 Simulink Modelling of Distribution Network . . . 35
3.2.1 Three Phase Programmable Voltage Source . . . 35
3.2.2 Three Phase Two Winding Transformer . . . 36
3.2.3 Three Phase Stub-Lines . . . 37
3.2.4 Three Phase Dynamic Load . . . 39
3.2.5 Three Phase VI Measurement Block . . . 40
3.2.6 Three Phase Fault . . . 40
3.2.7 Three Phase Breakers . . . 42
3.3 Fault Identification and Localization . . . 43
3.4 Implementation of Fault Identification and Localization in Simulink . . . . 44
3.4.1 Fault Identification . . . 44
3.4.2 Fault Localization and Phase Selection . . . 46
3.4.2.1 Single Phase Fault . . . 47
3.4.2.2 Line to Line Fault and Three Phase to Ground Fault . . . 48
4 Results 50 4.1 Scenario 1- Fault at Bus 10 . . . 50
4.1.1 Three Phase fault . . . 50
4.1.2 Line to Line Fault . . . 53
4.1.3 Single Phase Fault . . . 56
4.2 Fault on Line 4 . . . 59
4.2.1 Three Phase fault . . . 60
4.2.2 Line to Line fault . . . 63
4.2.3 Line to Ground Fault . . . 65
5 Discussion and Conclusion 69
6 Future Work 70
Bibliography 71
A Network, Line and Transformer Parameters 74
2.1 Wind Farm [9] . . . 14
2.2 Solar Photovoltaic [10] . . . 15
2.3 Representation of a Faulted Line . . . 20
2.4 Thevenin Equivalent Voltage Representation . . . 21
2.5 Single Line to Ground Fault . . . 21
2.6 Single Line to Ground Fault- Fault Current . . . 22
2.7 Single Line to Ground Fault- Fault Voltage . . . 22
2.8 Line to Line Fault . . . 23
2.9 Line to Line Fault- Fault Current . . . 23
2.10 Line to Line Fault- Fault Voltage . . . 24
2.11 Three Phase to Ground Fault . . . 24
2.12 Three Phase to Ground Fault-Fault Current . . . 25
2.13 Three Phase to Ground Fault . . . 26
2.14 Single Phase Fault Identification . . . 29
2.15 Negative sequence quantities during two phase fault . . . 29
2.16 Negative sequence quantities during two phase fault . . . 30
3.1 Distribution Network Diagram . . . 33
3.2 Network Model in Simulink . . . 34
3.3 Grid Model in Simulink . . . 34
3.4 Three Phase Programmable Voltage Source . . . 35
6
3.5 Generator Type . . . 36
3.6 Three Phase Two Winding Transformer . . . 36
3.7 Transformer Parameters . . . 37
3.8 Three Phase StubLine . . . 38
3.9 Three Phase Load . . . 39
3.10 Sample Load Profile . . . 39
3.11 Three Phase VI Measurement Block . . . 40
3.12 Three Phase Fault Settings . . . 41
3.13 Three Phase Fault Block . . . 41
3.14 Three Phase Breaker Block . . . 42
3.15 Three Phase Breaker Block Parameters . . . 42
3.16 Fault Identification and Localization Algorithm . . . 43
3.17 Agent Architecture . . . 44
3.18 Fault Identification . . . 45
3.19 Three Phase Fault Identification . . . 45
3.20 Two Phase Fault Identification . . . 46
3.21 Single Phase to Ground Fault . . . 46
3.22 Multi-Agent based controller for Breaker46 . . . 47
3.23 Multi-Agent based controller for Breaker42 . . . 47
3.24 Localization of Single Phase to Ground Fault . . . 48
3.25 Single Phase Fault- Phase Selection . . . 48
3.26 Two Phase Fault Localization . . . 49
3.27 Fault Localization Algorithm for Phase A . . . 49
4.1 Three Phase Fault Identification on Line 6 . . . 51
4.2 Trip Signal for Fault on Line 6 . . . 51
4.3 Fault current after issuing trip signal . . . 52
4.4 Voltage, Current and PQ at Bus 6 . . . 52
4.5 Voltage, Current and PQ at Bus 10 . . . 53
4.6 Fault Identification for Line to Line Fault at Bus 6 . . . 53
4.7 Trip Signal issued for Line to Line Fault at Bus 6 . . . 54
4.8 Fault Current for Line to Line Fault at Bus 6 . . . 54
4.9 Voltage, Current and PQ at Bus 6 for Line to Line fault . . . 55
4.10 Voltage, Current and PQ at Bus 10 for Line to Line fault . . . 55
4.11 Voltage, Current and PQ at Bus 4 for Line to Line fault . . . 56
4.12 Fault Identification for Line to Ground Fault at Bus 6 . . . 56
4.13 Trip Signal issued for Line to Ground Fault at Bus 6 . . . 57
4.14 Fault Current for Line to Ground Fault at Bus 6 . . . 57
4.15 Voltage, Current and PQ at Bus 6 for Line to Ground fault . . . 58
4.16 Voltage, Current and PQ at Bus 10 for Line to Ground fault . . . 58
4.17 Voltage, Current and PQ at Bus 4 for Line to Ground fault . . . 59
4.18 Network Simulation for fault on Line 4 . . . 59
4.19 Three Phase Fault Identification on Line 4 . . . 60
4.20 Trip Signal for Fault on Line 4 . . . 60
4.21 Fault current after issuing trip signal . . . 61
4.22 Voltage, Current and PQ at Bus 4 . . . 61
4.23 Voltage, Current and PQ at Bus 6 . . . 62
4.24 Line to Line Fault Identification on Line 4 . . . 63
4.25 Trip Signal for Line to Line Fault on Line 4 . . . 63
4.26 Fault current after issuing trip signal . . . 64
4.27 Voltage, Current and PQ at Bus 4 . . . 64
4.28 Voltage, Current and PQ at Bus 6 . . . 65
4.29 Line to Line Fault Identification on Line 4 . . . 65
4.30 Trip Signal for Line to Ground Fault on Line 4 . . . 66
4.31 Fault current after issuing trip signal . . . 67
4.32 Voltage, Current and PQ at Bus 4 . . . 67
4.33 Voltage, Current and PQ at Bus 6 . . . 68
4.34 Voltage, Current and PQ at Bus 2 . . . 68
2.1 Distributed Generation Categories . . . 14
A.1 Network Parameters . . . 74
A.2 Line Parameters . . . 74
A.3 Line Resistance, Inductance and Capacitance . . . 75
A.4 Transformer Parameters . . . 75
A.5 External Grid . . . 75
10
Introduction
This chapter presents a brief introduction to the thesis topic, the objective, the method- ology employed in the presented project and organization of thesis report.
1.1 Background
Modern society has become increasingly dependent on electricity to meet its energy demands. The advantage of electricity being easily transported at high efficiency and reasonable cost has made it a popular form of energy. There has been an ever increasing demand for electricity all across the globe. World Energy Outlook 2012 published by International Energy Agency projects over 70Traditional power plants have long been installed near the sources of energy and many a times, are away from centres of consump- tion. They are connected to the transmission network. The high voltage transmission grid though reliable suffers from possibility of cascading failures. The centrally concen- trated generation is also not efficient as it converts only about 35
To realise energy efficient solution to rising demand of electricity is having distributed generation near the centres of consumption, to maintain system stability, provide the spinning reserve and reduce the transmission and distribution with distributed optimizing control. The role of distribution network has been limited to servicing loads. This made the design of distribution network simpler as the power flow was unidirectional. With increased integration of DG the simplicity of network is set to disappear, which presents a challenge for distribution system in the form of new protection requirements and voltage stability issues.
To maximize use of renewable energy generated and to avoid long and unnecessary out- ages, a distributed approach to design distribution network protection has been consid- ered. Multi-Agent System is made of dynamic and autonomous agents working coopera- tively to achieve the overall aim, while deployed in a distributed fashion and working with
11
local constraints. Application of multi-agent system in power system has been possible with advancement in ICT, making it feasible to communication faster with equipments.
Agent based protection systems consist of distributed equipment combined with a com- munication network to cooperate in the realization of protection functions. The agents can also be provided with the capability of acting autonomously when the communication channels fail.
1.2 Thesis Statement
With increased integration of DER the distribution network can no longer be assumed to be radial networks. This requires changing the simplistic protection scheme that has been traditionally employed to protect network from faults. The power flow no longer being unidirectional presents new challenges to be addressed before integration of more DG.
This thesis tries to address the issue of fault localization on distribution network based on distributed protection through multi-agent system. Protection malfunctions on one hand can raise safety issues while they can also be a reason for underutilization of DG.
The aim of the thesis is to minimize outage in case of fault while servicing maximum load and making sure optimum utilization of infrastructure.
1.3 Contribution of Thesis
The thesis tries to address following issues-
1. Survey effects of integration of DG on distribution network 2. Proposes a fault localization algorithm based on distributed control applying multi-agent concept. 3.
Models a distribution network and multi-agent system in Simulink- SimPowerSystems 4.
Undertakes a case study to verify the proposed algorithm.
1.4 Organization of Thesis
The thesis is organized as follows- Chapter 2 summarizes Literature Review and Previous
Work. Chapter 3 discusses the Methodology Employed in accordance with the thesis
statement. Chapter 4 presents the results obtained for the cases considered. Chapter 5
includes conclusion and future work. The Simulink model of network and implementation
of Multi-Agent concept is included in the Appendices at the end of the report.
Related Work
This chapter presents the theoretical background related to the thesis topic. The chapter presents different types of Distributed Generation and their impact on DN, Active Dis- tribution Networks, Fault characterization and Protection Schemes.
The chapter further discusses Multi-Agent concept and its application for protection in Distribution Network.
Early power systems were designed to cater local loads through a simple infrastructure connecting generator to loads. The advent of AC system and efficiency constraints made it necessary to have centralized generation. These centralized generators, of higher ca- pacity were all connected to transmission network. The role left to distribution network was to service loads, by transferring power from transmission networks to loads.
The deregulation of energy market in the early 90s, has led to changing the centralized approach of generation, enabling power plants to produce electricity according to its marginal cost. This model resulted in a more complex power system with greater uncer- tainties. In addition, the advent of renewable energy further complicated the ability of utilities to adhere to the traditional system. Renewables have âœgrid priorityâ and the lowest marginal cost, meaning the grid is obligated to purchase their electricity first [5].
In recent developments, model of personal power plants [6] is being discussed where a building will have its own self-sufficient grid. Prosumer, a new term has been coined for power system participants who consume, produce as well as store energy while optimizing its utilization.
2.1 Distributed Generation
Since there are many definitions available, the paper "Distributed generation: a defi- nition" [7] considered different issues that may be used to define DG precisely. These
13
included, purpose, location, rating, power delivery area, technology, mode of operation, ownership and penetration of DG. Much of these issues were found to be irrelevant in defining the term. An attempt to present an inclusive definition from known international and regional definitions was made in "Review of the Distributed Generation Concept:
Attempt of Unification" [8]. The general definition arrived at is, distributed generation is an electric power source that is directly connected to distribution network or on the con- sumer side of the network, which is smaller in size compared to a centralized generator.
Different DG technologies available are- Wind Turbines, Solar Photovoltaic and Sterling Engine, Micro-Turbines, Combined Heat and Power Plants and Fuel Cells. Electricity is generated by employing suitable technology depending on available local resources. The categories suggested for types of DG depending on their capacities are as follows-
Table 2.1: Distributed Generation Categories
Category Capacity
Micro 1 Watt<5 kWatt Small 5 kWatt<5 MWatt Medium 5 MWatt<50 MWatt
Large 50 MWatt<300 MWatt
Some of the available DG technologies are-
2.1.1 Wind Farms
Wind power generation harnesses energy of wind to generate electricity. The turbines blades rotate as wind blows, generating lift. The hub of the turbine rotates in turn, which is connected to a drive train and coupled with a synchronous generator. Power electronic converters are connected to shape output of turbine and also to offer support to grid reactive power support in case of voltage sag.
Figure 2.1: Wind Farm [9]
2.1.2 Solar Photovoltaic
Series connected arrays of solar cells convert solar irradiance into usable electrical energy.
Maximum Power Point Tracking (MPPT) method is used to extract maximum power from solar panels. A battery bank and a charge controller are associated with the solar photovoltaic plant to store energy.
Figure 2.2: Solar Photovoltaic [10]
2.1.3 Micro Turbine
Micro-gas turbines are a type of internal combustion engine where chemical energy of fuels is converted into electricity. Gas turbines accept most commercial fuels, such as petrol, natural gas, propane, diesel, and kerosene as well as renewable fuels such as E85, biodiesel and biogas [10].
2.1.4 Fuel Cells
Fuel cell generates DC voltage based on interaction of hydrogen and oxygen in presence of a conducting electrolyte. The DC voltage thus generated is converted to AC by means of an inverter. Advantage of fuel cells is absence of any moving component. Higher cost of operation is the disadvantage.
2.2 Impact of DG on Distribution Network
Addition of DG has positive and negative impacts on various aspects of distribution net-
work operation. These include impact on voltage regulation, losses in feeder, harmonics
in the network and on level of short circuit currents. Distribution networks are often built
in mesh structures but operated in open loop configuration. This results in the network
being a radial one with unidirectional power flow. This assumption has traditionally
influenced operation of distribution network. In radial distribution systems voltage is regulated with the help of load tap changing transformers (LTC) at substations, by line regulators on distribution feeders and shunt capacitors on feeders or along the line. The connection of DG may result in changes in voltage profile along a feeder by changing the direction and magnitude of real and reactive power flows. The impact can be either pos- itive or negative depending on nature and location of DG. Medium-sized and especially small DG technologies often use asynchronous generators (also known as induction gen- erators), as they are significantly cheaper than synchronous generators. Grid-connected asynchronous generator is not capable of providing reactive power. It actually requires reactive power from the grid during the start-up process and at operation [10] thus af- fecting voltage profile. DG can provide voltage support to raise the low voltage at the end of the feeder. However, the voltage where DG is injected may be raised over the upper limit [? ]. Optimally located DG units can help in minimising feeder losses by supplying active and reactive power to the network. However, most DGs are customer owned and hence grid operator canâ
TMt decide location of DGs. Normally it is assumed that, DGs located away from Distribution Substation would help in reducing losses over distribution network. DGs may introduce harmonics to the network depending on type of generator and interconnection configuration. The harmonics could be from generator itself, for example from synchronous generator or from power electronic converter. The older SCR based inverter that are line commutated and produce high levels of harmonic current. Most new inverter designs are based on IGBTs that use pulse width modulation to generate the injected "sine" wave. These newer inverters are capable of generating a very clean output and they should normally satisfy the IEEE 1547 requirements [12].
For radial feeders the symmetrical short circuit current is calculated as the ratio between voltage and the total impedance from the source to the fault location. This implies that the short circuit current decreases as the fault location moves away from the source. In absence of DG connected down-stream to the fault, there is no contribution to the fault other than the one coming from the source. Under these premises the coordination of protective devices is necessary only for upstream protective devices neglecting any other down-stream devices. When DG is connected to the system a secondary short circuit current contribution coming from the DG to the fault takes place. This contribution is similarly calculated as the ratio between the voltage and the total impedance from the DG to the fault. This short circuit current contribution coming from the DG adds to the main source contribution current at the fault location increasing the total short circuit level of the fault [13]. Impact of DG on distribution network costs was assessed in [14];
it presents quantification of how DG affects distribution network costs in presence of
distributed generation. The results show that total distribution network costs increase
with DG penetration.
2.3 Impact on Distribution Network Protection
Reliability of power system depends closely on its protection system. The addition of distributed generation (DG) to the power distribution system imposes a challenging and unconsidered fault condition to the traditional distribution overcurrent protection scheme. In the event of fault, DG short circuit contribution may cause erroneous opera- tion of protective devices located downstream to the fault. The fundamental principle of the vulnerability assessment is the comparison of maximum and minimum short circuit levels before and after the distributed generation is added. Those protective devices installed at buses where the DG short-circuit contribution exceed the pre-DG minimum short circuit level are the most likely to miss-operate under fault conditions.
Protection schemes implemented for Distribution Network protection traditionally in- volve over current protection with suitable time delays, circuit breakers, auto-reclosers and fuses. Fault protection is obtained by means of over-current relay, earth fault direc- tional relays and zero sequence voltage and current relays. Overcurrent relays are not directional. But load flow may not remain unidirectional in case of DG connected to DN and it can introduce short circuit current in the opposite direction to that of introduced by main grid. This DG contribution may prompt the uncoordinated operation of de- vices located downstream to the fault location unveiling a potential vulnerability of the protection system [13].
The presence of DG can cause various problems related to incorrect operation of system protections. Conflicts between DG and protection schemes are typically due to:
• unforeseen increase in short circuit currents;
• lack of coordination in the protection system;
• ineffectiveness of line reclosing after a fault using the ARD
• difficult lines back-feeding
• Undesired islanding and untimely tripping of generators interface protections [15].
Moreover, assuming the possibility to manage portions of distribution networks in is- landed configuration, it would be necessary to consider the potential difficulties in setting up adequate protection schemes that can operate properly when the Microgrid is both connected to and disconnected from the main network to cope with internal faults.
Thus issues associated with introduction of DGs to a distribution network include "blind- ing of protection", "false/sympathetic tripping", "recloser-fuse miscoordination", and
"failed auto-reclosing" [16].
2.4 Active Distribution Networks
Active Distribution Network is a migration from earlier "fit and forget" version of dis- tribution grid. It is proposed to make grid more flexible by providing ways to exercise control and management in an optimal way. ADN enables better utilization of infras- tructure by means of control function and ICT.
In its report for the project IDE4L on Specifications of Active Distribution Networks [17], Active distribution network is defined as follows- Active distribution network consists of infrastructure of power delivery, active resources and active network management.
It combines passive infrastructure with active resources, ANM functionalities and ICT infrastructure.
2.4.1 ICT Infrastructure
Extensive use of ICT makes it possible to collect information from various points in a network which can be utilised for
• Network Planning and operation
• Network performance optimization
• Network control and protection
For collecting data, monitoring of entire network needs to be done. Installation of Ad- vanced Metering Infrastructure, Micro-PMUs, increasing the number of IEDs on distri- bution side is a few measures to improve distribution side monitoring. A strong com- munication network is necessary to transmit the measurements collected to a centrally connected location where it could be utilized for monitoring and management of network in real time.
2.4.2 Active Resources and Aggregators
Active resources are the ones that could be monitored and controlled by ANM. Active DERs, Micro-grids, flexibility services from Aggregator, STATCOMs, could be classified amongst active resources. The Aggregator is a service provider and can supplement the production and consumption portfolios of an electricity retailer by flexible DERs [17].
Services offered by the aggregator for commercial purposes are-
• Price elasticity at day-ahead market to manage price risk
• Demand response to manage forecasting risks and to minimize balance costs
• Demand response offer for regulation power market
• Distributed generation offer to day-ahead and intraday markets
Services provided by Technical aggregator to TSOs and DSOs are-
• Reserve market (disturbance and control reserves: virtual inertia, spinning, fast and slow reserves)
• Power flow management (e.g. control offer for regulation power market)
• Reactive power support and Voltage control
• Power quality control
• Back-up power
2.4.3 Active Network Management
Active network management is based on distribution network automation and DERs.
It includes distribution control centre information systems, substation automation, sec- ondary substation automation and advanced metering infrastructure.
Active distribution network enables optimal management of the whole network with dif- ferent DG scenarios, EV and heat pump penetration levels and DER scenarios that will affect the loading of networks and transformers, the voltage profile of LV network and line overloading. The new functionalities required for the operation of the distribution net- works like distribution state estimation, automatic FLISR, coordinated voltage control, power flow control, static and dynamic distribution model order reduction to coordinate with TSOs, and the availability of stability indicators [17].
2.5 Faults in Distribution Network
Faults that occur in Distribution network include Symmetrical 3 phase faults and asym-
metrical Line- Ground fault, Line â“ Line and Double Line to Ground faults. This section
has taken inspiration from [18].
2.5.1 Fault
For fault current calculation purpose following assumptions are made-
• The power system is balanced before the fault occurs such that of the three sequence networks only the positive sequence network is active. Also as the fault occurs, the sequence networks are connected only through the fault location.
• The fault current is negligible such that the pre-fault positive sequence voltages are same at all nodes and at the fault location.
• All the network resistances and line charging capacitances are negligible.
Based on assumptions above the voltage at the faulted point will be denoted by V
fand current in the three faulted phases are I
fa ,I
fb and I
fc. Faults in 3 phase system are
Figure 2.3: Representation of a Faulted Line
divided into two groups of symmetrical and unsymmetrical faults. Symmetrical fault occurs when all 3 phases are grounded at the same time. In that case it is considered to be equivalent to 3 line-ground faults. A more detailed discussion of asymmetrical faults is provided here. For phase currents to be denoted by I
a, I
b, I
cand the sequence components for both currents and voltages are I
1, I
2, I
0and U
1, U
2, U
0where I
1represents the positive sequence, I
2represents the negative sequence and I
0represents the zero sequence. The same applies for voltages. The phase current can be expressed as
I
a= I
1+ I
2+ I
0I
b= a
2I
1+ aI
2+ I
0I
c= aI
1+ a
2I
2+ I
0(a)
It can be derived that, I
0= 1/3(I
a+ I
b+ I
c) I
1= 1/3(I
a+ aI
b+ a
2I
c)
I
2= 1/3(I
a+ a
2I
b+ aI
c) (b)
U
1= V − I
1Z
1U
2= −I
2Z
2Figure 2.4: Thevenin Equivalent Voltage Representation
U
3= −I
0Z
0(c)
Following section presents all three types of faults along with fault voltage and current plots.
2.5.2 Single Line to Ground Fault
A single line to ground fault occurred at node is depicted in the figure. Assuming that phase-a has touched the ground through an impedance Zf.
Figure 2.5: Single Line to Ground Fault
Since the system is unloaded before the occurrence of the fault
I
f c= I
f b= 0 (2.1)
Fault current and voltages for a single line to ground fault are shown in the figures below-
Figure 2.6: Single Line to Ground Fault- Fault Current
Figure 2.7: Single Line to Ground Fault- Fault Voltage
2.5.3 Line to Line Fault
Line to Line fault can be represented as shown in the figure below. In this the phases b and c got shorted through the impedance Z
f.
Figure 2.8: Line to Line Fault
For an unloaded system,
I
f a= 0 (2.2)
Since fault is between phases b and c,
I
f b= −I
f c(2.3)
Two phase fault current and voltage quantities are plotted in the figures below-
Figure 2.9: Line to Line Fault- Fault Current
Figure 2.10: Line to Line Fault- Fault Voltage
2.5.4 Three Phase to Ground Fault
In three-phase to ground fault can be represented by figure below-
Figure 2.11: Three Phase to Ground Fault
An example of fault current and voltage in the event of a three phase fault is given below
respectively-
Figure 2.12: Three Phase to Ground Fault-Fault Current
2.5.5 Consequences of Faults
The consequences of electrical faults in power systems strongly depend on the magnitude of the fault current, which in turn depends on the type of fault, the location of the fault, the system earthing, the source impedance, and the impedance of the fault. The duration of the fault is also of considerable importance when estimating the consequences of a fault.
Consequences of fault include-
1. Mechanical stresses- Conductor exert stress on each other which depends on square of current flowing in them. Fault current being higher than rated current the mechanical stress exerted is higher than what the line was designed for.
2. Thermal Stress â“ Losses within the conductor are given by I2R. Since current rises during fault the losses are higher. The higher energy that is to be dissipated in the conductors leads to higher thermal stress.
3. Arcing- If a fault is not cleared immediately, it may cause arcing within cables or within the equipment.
4. Cost- Fault current causes eventual wearing of insulation of cables. It also affects
tips of circuit breaker contacts and instrument transformers, transformers and gen-
erators as well. Replacing the system and the subsequent down time cause losses
Figure 2.13: Three Phase to Ground Fault
not just to the utility as well as to overall commercial activities based in the affected region.
2.5.6 Fault Clearing
Devices that are used for fault clearing are-
1. Fuses
2. Circuit Breakers
Fuses- Fuses work to limit the first peak current of the short circuit fault. Fuses tend to melt down in case of higher energy dissipation within them thus breaking off contacts between two terminals. The drawback in its use is, it has to be replaced once used.
Circuit Breaker- Circuit breakers break fault current based on natural zero crossing of
short circuit current. At that moment arcing between the tips of circuit breaker contacts
can be extinguished and current could be interrupted.
2.5.7 Current Limiters
Current limiting is used to protect power system from mechanical forces associated with the first current peak. Series reactors and solid state based current limiters are some examples of current limiters. Series reactors- The traditional method limiting fault cur- rent is to install series reactor in main circuit. The drawback of this installation being, additional losses that occur when load current passes through the reactor. It also fails to limit the fault current to the requisite extent. Solid State Current Limiter â“ Solid state devices do not depend on natural zero crossing to interrupt the current. When a signal is sent to the control circuit of solid state device, it immediately turns off the current through the device. Drawbacks of solid state limiter include 1. Cost of the com- ponents and 2. Probability of damage due to dissipation of energy within components when current is interrupted at peak value. Other current limiters include Switch and Fuse combination current limiter.
2.5.8 Fault Clearing System
Fault clearing system consists of relay protection system and circuit breaker.
2.5.8.1 Relay Protection System
Relay protection system consists of sensors/ transducers, trip coil, relays and auxiliary power supply. To detect fault, quantities that are commonly observed are currents and voltages. The measuring devices introduce a delay. According to IEC, a delay of 0.17 ms is allowed for a network working on frequency of 50 Hz.
Trip coil- Circuit breakers are operated by releasing energy stored in a large spring.
The latch which holds the spring is released by energizing a coil that provides a force that acts on the latch thus releasing it. The opening coil needs a rather large power to operate the latch. The power is supplied by a circuit to which the contact of the relay is connected. Relays- Relays are classified into Electromechanical, Solid State and Digital/Numerical Relays. Electromechanical relays were the first relays to be used in power systems for protection purposes. The operation of electromechanical relay depends on physical characteristics of current and voltage. Solid State Relay- Designed based on working of Operational Amplifier.
Digital/ Numerical Relays- These relays investigate fault condition based on computa-
tions carried out by microprocessor. These consist of a signal conditioning subsystem, a
conversion subsystem, and a digital processing subsystem.
2.5.8.2 Circuit Breaker
Circuit breaker as the name suggest are necessary to break the contacts. On detection of fault, contacts within the circuit breaker must open to interrupt the circuit. Separation of contacts is carried out by mechanically-stored energy. Different types of circuit breaker include air circuit breaker, air-blast circuit breaker, Oil circuit breaker, vacuum circuit breaker, SF6 circuit breaker etc. These breakers are used depending on breaking capacity of the breaker.
2.5.9 Protection Principles
Magnitude relay In this case magnitude of current or voltage is compared to a pre- determined threshold level. In this case, instantaneous value of the quantity cannot be used for comparison. RMS value of quantity is compared to threshold. Directional Relay “ Phase angle between current and polarizing quantity is estimated. It is possible to tell if the fault is forward or in reverse direction, based on the measure of angle as seen from relay. Impedance Relay “ Measuring value of voltage and current, estimation of impedance of protected object can be made. Differential Relay- Differential relay are used for equipment protection. It requires two sets of inputs to determine health of the system. Generators, Transformers and machines are protected by differential relays. Pilot relay- Pilot relay consists of any of the relays mentioned above along with communication between two or more adjacent relays.
2.5.10 Fault Identification
Positive sequence quantities are found in all fault types as well as normal system op- eration. Negative sequence quantities are found in line-line (LL) and line-ground (LG) faults. Zero sequence quantities are found only in line to ground faults. In overcurrent protection, zero sequence quantities are used to detect the ground faults. Zero sequence quantities are absent in balanced systems and other fault events. Hence it makes sense to use zero sequence quantities for identification of single phase to ground fault.
Positive sequence currents are typically used for the three Phase and LL faults. Posi-
tive sequence phase pickup current must be set above load current to prevent nuisance
tripping. This decreases the sensitivity to LL faults. Because there are no zero sequence
currents in LL faults, zero sequence ground fault protection will not work. In this sit-
uation, negative sequence currents found in LL faults are useful. Because no negative
sequence quantities exist in a balanced system, negative sequence pickups can be set
below load current [19].
Figure 2.14: Single Phase Fault Identification
Figure 2.15: Negative sequence quantities during two phase fault
For three phase fault identification, principle of Magnitude relay is employed. RMS value of current in each phase is compared to a threshold. Threshold for identification of a three phase fault is set to thrice the rated current, I
rated.
2.6 Multi-Agent Concept
Multi-Agent based systems are being increasingly used for monitoring and control of power system. Agent based system has inherent benefits of flexibility, extensibility, au- tonomy, reduced maintenance and makes possible to implement network control through intelligent decision making. For distributed control of a system, it is divided into mul- tiple subgroups and each subgroup is assigned an agent. The agent collects information for its assigned subgroup and communicates with other agents. The decision is arrived at by optimization of the information exchanges amongst cooperating agents. A central- ized system may be plagued by resource limitations, performance bottlenecks, or critical failures, an MAS is decentralized and thus does not suffer from the "single point of fail- ure" problem associated with centralized systems. Application of Multi-agent concept also covers distribution network automation and self healing grids. The paper titled
"Multi-agent application in protection coordination of power system with distributed generations" [20] presents use of agent technology applied to substations protection co- ordination of power system. The architecture of MAS could be represented as shown in the figure below. The agent receives information from its surrounding environment, applies the set logic and issues command through actuator as shown in the figure.
Figure 2.16: Negative sequence quantities during two phase fault
In its thesis, Sandia National Laboratories concluded that, "agents are a valuable frame-
work for operating microgrids. Intelligent agents offer a secure, distributed means to
manage electric power production, enforce system policy, and respond to unexpected
events in the power network" [21]. In multi-agent systems, several agents work in tan-
dem to achieve the specified goal of the system. The agents perceive and react to their
environment and communicate amongst each other. This ensures autonomous behaviour
of agents and rules out need of any human intervention. Multi-Agent concept has previ-
ously been applied for power system control and management applications. Management
of distribution system with distributed generation with the help of Multi-Agent system
was shown in [22] for dynamic and efficient power dispatch. In [20], it has been used for
protection coordination and [23] used it for system restoration purpose.
Methodology Employed
In this chapter the working process is described. The distribution network under study is presented. It is followed by steps taken for simulating the network in SimPowerSys- tems. The fault localization algorithm is then presented and eventually implementation of multi-agent based protection algorithm in SimPowerSystems is described.
3.1 Distribution Network
The distribution network simulated, consists of a 60:11 kV OLTC Distribution Trans- former which steps down voltage from 60 kV to 11 kV. The network operates at 2 voltage levels and consists of 11 Buses and 8 distribution lines of lengths ranging from 100 m to 2.4 km. There are 8 loads connected to the network as shown in the diagram. Total load connected to the network is 18.45 MW with reactive power consumption of 8.93 MVAR at a power factor of 0.9. Transformer connected is Yg-D11 type transformer with rated primary and secondary voltages of 60 kV and 11 kV respectively and rated power of 50 MVA. The distributed generators are connected at Buses 5 and 7 and of capacity 5 MW and 10 MW respectively. The network studied is depicted in the figure below.
32
Figure 3.1: Distribution Network Diagram
Figure 3.2: Network Model in Simulink
The model simulated in SimPowerSystems is as shown below:
Figure 3.3: Grid Model in Simulink
3.2 Simulink Modelling of Distribution Network
The Simulink SimPowerSystems was used for modelling of network. The components used are-
1. Three Phase Programmable Voltage source 2. Three Phase Two Winding Transformer 3. Three Phase Stub Lines
4. Three Phase Dynamic Loads
5. Three Phase VI Measurement Blocks 6. Three Phase Fault
7. Three Phase Breakers
3.2.1 Three Phase Programmable Voltage Source
3 Phase programmable voltage block is used in the network to specify external grid. As there is a need of a swing bus in the network, Load flow of the programmable voltage source is set to swing.
Figure 3.4: Three Phase Programmable Voltage Source
Figure 3.5: Generator Type
3.2.2 Three Phase Two Winding Transformer
The external source is connected to a 3 phase 2 winding transformer which steps down 60 kV primary voltage to 11 kV secondary voltage. Rated capacity of Transformer is 50 MVA.
Figure 3.6: Three Phase Two Winding Transformer
Figure 3.7: Transformer Parameters
3.2.3 Three Phase Stub-Lines
Artemis Stubline is part of Artemis Toolbox. It has been used keeping in mind real time
simulation of network. Stubline enables a user to connect multiple phase networks. The
inductance and resistance parameters are specified based on length of each line segment
of network. Based on value of sample time Ts, the block estimates value of capacitance
of the line.
Figure 3.8: Three Phase StubLine
The block is suitable for a short line. As a rule of thumb, if length of line is smaller than
Ts*30000, Stubline is preferred over other blocks in Artemis Toolbox.
3.2.4 Three Phase Dynamic Load
Three Phase dynamic load implements dynamic load with active power and reactive power as function of voltage or controlled from external input. In this case the load is assumed to varying with the change of time hence active and reactive power.
Figure 3.9: Three Phase Load
Figure 3.10: Sample Load Profile
3.2.5 Three Phase VI Measurement Block
The Three-Phase V-I Measurement block is used to measure instantaneous three-phase voltages and currents in a circuit. When connected in series with three-phase elements, it returns the three phase-to-ground or phase-to-phase peak voltages and currents [24].
Figure 3.11: Three Phase VI Measurement Block
3.2.6 Three Phase Fault
The Three Phase Fault block was used to introduce fault in the network.
Fault is introduced when First input to the block is 1. Input 2 specifies type of the fault.
Here a 3 phase symmetrical fault is introduced. For Input 2 = 1, the block will introduce
fault in phase A of the network and for fault type set to 2, the block will be introduced in
phases A and B. The connection of phases could be altered to introduce fault in specific
phase in the network.
Figure 3.12: Three Phase Fault Settings
Figure 3.13: Three Phase Fault Block
3.2.7 Three Phase Breakers
Three phase breaker implements three-phase circuit breaker opening at zero crossing of current.
Figure 3.14: Three Phase Breaker Block
The Three-Phase Breaker block implements a three-phase circuit breaker where the open- ing and closing times can be controlled either from an external Simulink signal, or from an internal control timer. External signal equal to 1 closes the breaker and the signal equal to 0 opens it. High value of Snubber resistor is to minimize current in the network in case of opening of breaker. Few basic IGBT based solid state breakers were introduced to minimize transient in simulation results.
Figure 3.15: Three Phase Breaker Block Parameters
3.3 Fault Identification and Localization
Based on the discussions in previous chapter fault identification and localization algo- rithm was formulated. The diagram below depicts the algorithm employed for identifi- cation and localization purpose.
Figure 3.16: Fault Identification and Localization Algorithm
The algorithm presented above is for an individual node agent as depicted in the diagram below. The node agent receives measurements from the bus; it identifies the type of
Figure 3.17: Agent Architecture
fault based on the steps mentioned in algorithm. It then communicates with its adjacent node agents, as shown in the diagram; to identify the faulty section of network. After identification of faulted phase, trip signal is issued to respective breaker to isolate the faulted phase.
3.4 Implementation of Fault Identification and Localization in Simulink
The algorithm shown in figure 3.16 was implemented in Simulink. The task to be per- formed is divided into two parts- Fault Identification and the next Fault Localization.
3.4.1 Fault Identification
Components that were used for Fault Identification purpose are-
1. RMS- Root Mean Square of a signal.
2. Switch
3. Sequence Analyser
Simulink representation of Fault Identifier is as shown in the figure below- Detailed
description of Fault Identification is as follows- Identification of Three Phase fault is done
by Magnitude Relay principle where RMS value of current is compared to a threshold
set three times the rated value.
Figure 3.18: Fault Identification
Figure 3.19: Three Phase Fault Identification
Since a 3 phase fault is nothing but a combination of 3 single phase faults the network showed above works well to identify a 3 phase fault. 2 Phase fault is identified by comparing negative sequence component of the current with a threshold value. Since negative sequence quantities are found in Line- Line faults, the same could be utilized to identify 2 Phase faults.
Single Phase to ground fault is determined by taking RMS value of sum of all phase
quantities. This sum has a very small value of the range 10-6 otherwise. In case of
Line to Ground fault, this quantity assumes a larger value and can be compared to a set
threshold.
Figure 3.20: Two Phase Fault Identification
Figure 3.21: Single Phase to Ground Fault
The block that contains these elements is as shown below. Since the identification is done based on current, current is given as input to the block and F4, the output serves to tell if fault has occurred or not. The output is in the form of an integer where, integer 1- specifies Single Phase to Ground Fault, integer 2 specifies Line to Line fault and integer 3 specifies Symmetrical 3 Phase to ground fault. Once the fault is identified the output F is then given to a Multi-Agent based Fault Localization and Phase Selection block.
3.4.2 Fault Localization and Phase Selection
Fault Localization and Phase selection is done to identify the location of fault by dis- tributed algorithm and to identify suitable breaker to be opened to isolate the faulted section. Each bus is considered as an Agent, with current, voltage and power as its components.
Each bus contains number of controllers based on number of adjacent buses it is connected
to. The relevant controller is operated when conditions set in it are met, opening the
relevant breaker. The first task for the controller is to localize the fault. Controller is
fed type of fault in the network. Based on nature of fault i.e. Single phase, 2 Phase or
Symmetrical different approaches are considered.
Figure 3.22: Multi-Agent based controller for Breaker46
Figure 3.23: Multi-Agent based controller for Breaker42
3.4.2.1 Single Phase Fault
Localization of single phase fault is done by comparing difference of RMS values of sum
of adjacent line currents as shown in the figure below-
Figure 3.24: Localization of Single Phase to Ground Fault
Based on the comparison, location of 1 phase to ground fault is determined. The faulted
Figure 3.25: Single Phase Fault- Phase Selection
phase on selection will give output equal to 0 which could be used to isolate the breaker relevant to the phase.
3.4.2.2 Line to Line Fault and Three Phase to Ground Fault
As mentioned in the algorithm, Line to Line fault phases and 3 phase faults are localized
based on direction of active power flowing in the line connecting two buses. Identical
blocks are used for this purpose. The blocks are as shown in the figure below-
Figure 3.26: Two Phase Fault Localization
All 3 outputs of the block are constant equal to 1 when there is no fault. As a fault is reported, the switch connects the output to terminal 1. Details of the block connected at terminal 1 are as follows-
Figure 3.27: Fault Localization Algorithm for Phase A
Based on reversal of active power among buses, respective breakers are sent output
signal equal to zero, isolating the faulty phases. The same procedure is applied for
all buses,except the terminal ones. For the terminal buses, in case of fault, the power
flowing through the bus becomes zero. In that case it is not possible to apply this
method. Instead breaker of terminal buses is coupled with breaker connecting the same
line to adjacent pre-terminal bus.
Results
In this chapter the results that obtained with the methods described in the previous chap- ter are compiled, and compared with the theory presented in the Related Work chapter.
Two scenarios are considered for the case study. In first case fault is introduced at the farthest point of the network, on Bus 10. In scenario two, fault is simulated at Bus 4. In both cases fault was introduced at 0.15 seconds, the simulation running time was 0.25 sec.
4.1 Scenario 1- Fault at Bus 10
Fault introduced on Line 6 connecting Bus 6 and Bus 10.
4.1.1 Three Phase fault
Fault identification in case of fault on Line 6.
50
Figure 4.1: Three Phase Fault Identification on Line 6
The trip signal issued to isolate faulted region.
Figure 4.2: Trip Signal for Fault on Line 6
Breakers corresponding to Line 6 are issued trip signal. Opening of the breakers will
lead to fault being isolated. Interruption of fault current in this case verifies isolation of
faulted line.
Figure 4.3: Fault current after issuing trip signal
Response at various buses, bus 6 and 10 respectively-
Figure 4.4: Voltage, Current and PQ at Bus 6
Figure 4.5: Voltage, Current and PQ at Bus 10
The current, voltage and PQ responses show that Line 6 has been isolated bringing these quantities to zero at bus 10.
4.1.2 Line to Line Fault
Results of fault identification process for Line to Line fault on Line 6 are-
Figure 4.6: Fault Identification for Line to Line Fault at Bus 6
Bus 4 and 6 identify the two phase fault that lies downstream to them. With the help of fault localizing agents, the exact location of faulted line is located and relevant breakers are opened. The trip signal issued are presented in the figure below- Trip signals are
Figure 4.7: Trip Signal issued for Line to Line Fault at Bus 6
issued to breakers B610 and B106 which isolate Line 6. As the faulted phases are cut off, the fault current can be shown as-
Figure 4.8: Fault Current for Line to Line Fault at Bus 6
Plots of voltage, current and PQ waveforms during Line to Line fault at bus 6 are-
Figure 4.9: Voltage, Current and PQ at Bus 6 for Line to Line fault
Figure 4.10: Voltage, Current and PQ at Bus 10 for Line to Line fault
Figure 4.11: Voltage, Current and PQ at Bus 4 for Line to Line fault
The operation at Bus 4 remains unaffected after opening of breakers on Line 6.
4.1.3 Single Phase Fault
Fault Identification in case of single phase fault-
Figure 4.12: Fault Identification for Line to Ground Fault at Bus 6
Buses 4, 6 and 10 identify the single phase fault. With the help of fault localizing agents, the exact location of faulted line is located and relevant breakers are opened. The trip signal issued are presented in the figure below- Trip signals are issued to relevant breakers
Figure 4.13: Trip Signal issued for Line to Ground Fault at Bus 6
on Line 6. As the faulted phases are cut off, the fault current can be shown as-
Figure 4.14: Fault Current for Line to Ground Fault at Bus 6
Plots of voltage, current and PQ waveforms during Line to Ground fault at bus 6 are-
Figure 4.15: Voltage, Current and PQ at Bus 6 for Line to Ground fault
Figure 4.16: Voltage, Current and PQ at Bus 10 for Line to Ground fault
Figure 4.17: Voltage, Current and PQ at Bus 4 for Line to Ground fault
4.2 Fault on Line 4
Fault is introduced at Bus 6 on line 4 connecting Bus 4 and 6.
Figure 4.18: Network Simulation for fault on Line 4
Figure 4.19: Three Phase Fault Identification on Line 4
4.2.1 Three Phase fault
Fault identification in case of fault on Line 4. The trip signal issued to isolate faulted region.
Figure 4.20: Trip Signal for Fault on Line 4
Breakers corresponding to Line 4 are issued trip signal. Opening of the breakers will
lead to fault being isolated. Interruption of fault current in this case verifies isolation of
faulted line.
Figure 4.21: Fault current after issuing trip signal
Response at various buses, bus 4 and 6 respectively-
Figure 4.22: Voltage, Current and PQ at Bus 4
Figure 4.23: Voltage, Current and PQ at Bus 6
The current, voltage and PQ responses show that Line 4 has been isolated.
Figure 4.24: Line to Line Fault Identification on Line 4
4.2.2 Line to Line fault
Fault identification in case of Line to Line fault on Line 4. The trip signal issued to isolate faulted region.
Figure 4.25: Trip Signal for Line to Line Fault on Line 4
Breakers corresponding to Line 4 are issued trip signal. Opening of the breakers will
lead to fault being isolated. Interruption of fault current in this case verifies isolation of
faulted line.
Figure 4.26: Fault current after issuing trip signal
Response at various buses, bus 4 and 6 respectively-
Figure 4.27: Voltage, Current and PQ at Bus 4
Figure 4.28: Voltage, Current and PQ at Bus 6
4.2.3 Line to Ground Fault
Fault identification in case of Line to Ground fault on Line 4.
Figure 4.29: Line to Line Fault Identification on Line 4
The trip signal issued to isolate faulted region.
Figure 4.30: Trip Signal for Line to Ground Fault on Line 4
Breakers corresponding to Line 4 are issued trip signal. Opening of the breakers will
lead to fault being isolated. Interruption of fault current in this case verifies isolation of
faulted line.
Figure 4.31: Fault current after issuing trip signal
Response at various buses, bus 4 and 6 respectively-
Figure 4.32: Voltage, Current and PQ at Bus 4
Figure 4.33: Voltage, Current and PQ at Bus 6
Figure 4.34: Voltage, Current and PQ at Bus 2
