Particle Swarm Optimization Approach for Distributed Generation Planning to Improve the Voltage Profile Margin


Master's Thesis, 2017
132 Pages, Grade: 4.00

Free online reading

TABLE OF CONTENTS

ACKNOWLEDGMENTS

LIST OF FIGURES

LIST OF TABLES

ABSTRACT

LIST OF ABBREVIATIONS

LIST OF SYMBOLS

CHAPTER
1.1 Introduction
1.2 Statement of the Problem
1.3 Research Gap and Research Questions
1.4 Objective
1.5 Scope
1.6 Limitation
1.7 Thesis Organization

CHAPTER
2.1 Background
2.2 Literature Survey
2.3 Distributed Generation (DG)
2.4 Distributed Generation Resources
2.4.1 Dispatchable Energy Resources
2.4.2 Non-Dispatchable Energy Resources
2.5 Distributed Generation Models
2.6 Power Distribution Systems Configurations
2.7 Power System Stability
2.7.1 Rotor Angle Stability and Frequency Stability
2.7.2 Voltage Stability
2.8 Effect of DG unit on Voltage Profile and Loss of the System
2.9 Power Flow Analysis in Distribution System
2.9.1 Backward / Forward Sweep Algorithm
2.10 Optimization Techniques
2.10.1 Analytical Method
2.10.2 Exhaustive Method
2.10.3 Linear Programming (LP) Method
2.10.4 AC Optimal Power Flow (OPF) Method
2.10.5 Metaheuristics Method
2.10.6 Multi-Objective Programming Method
2.10.7 Probabilistic Method

CHAPTER
3.1 General
3.2 Proposed Methodology
3.3 Methodological Tools

CHAPTER
4.1 General
4.2 DG Integration in IEEE 12 Bus Test System
4.2.1 Load Flow Analysis of IEEE 12 Bus Test System
4.2.2 Results of Optimal Allocation of DG unit in IEEE 12 Bus Test System
4.3 DG integration in IEEE 15 Bus test System
4.3.1 Load Flow Analysis of IEEE 15 Bus Test System
4.3.2 Results of Optimal Allocation of DG unit in IEEE 15 Bus Test System
4.4 DG integration in IEEE 33 Bus test System
4.4.1 Load Flow Analysis of IEEE 33 Bus Test System
4.4.2 Results of Optimal Allocation of DG unit in IEEE 33 Bus Test System
4.5 Discussion

CHAPTER
5.1 General
5.2 DG Integration in Bhaktapur - Balaju 21 Bus Distribution System
5.2.1 Load Flow Analysis of Bhaktapur - Balaju 21 Bus System
5.2.2 Results of Optimal Allocation of DG unit in Bhaktapur - Balaju 21 Bus System
5.3 Discussion

CHAPTER
6.1 Conclusion
6.2 Recommendations for Future Works

REFERENCES

APPENDIX A
PARTICLE SWARM OPTIMIZATION

APPENDIX B
IEEE - 12 BUS SYSTEM DATA [58]
IEEE - 15 BUS SYSTEM DATA [59]
IEEE - 33 BUS SYSTEM DATA [60]
BHAKTAPUR - BALAJU 21 BUS SYSTEM DATA [61]

APPENDIX C
RESULTS OF IEEE - 12 BUS SYSTEM
RESULTS OF IEEE - 15 BUS SYSTEM
RESULTS OF IEEE - 33 BUS SYSTEM
BHAKTAPUR - BALAJU 21 BUS SYSTEM

APPENDIX D
CANDIDATE BUS VOLTAGE FOR IEEE - 12 BUS SYSTEM
REAL POWER LOSS PROFILE FOR IEEE - 12 BUS SYSTEM
CANDIDATE BUS VOLTAGE FOR IEEE - 15 BUS SYTEM
REAL POWER LOSS PROFILE FOR IEEE - 15 BUS SYSTEM
CANDIDATE BUS VOLTAGE FOR IEEE - 33 BUS SYTEM
REAL POWER LOSS PROFILE FOR IEEE - 33 BUS SYSTEM
CANDIDATE BUS VOLTAGE FOR BHAKTAPUR - BALAJU 21 BUS SYTEM
REAL POWER LOSS PROFILE FOR BHAKTAPUR - BALAJU 21 BUS DISTRIBUTION SYSTEM

DEDICATION

To my parents,

Ram Narayan Byar and Sumintra Devi Byar My Sister and Brother - in - Law Kumari Shakti Singh and Shyam Prakash Mahato My Brother and Sister - in - Law Vijay Singh and Rakhi Purbey Singh And To

All of my Friends

ACKNOWLEDGMENTS

First of all, I would like to thank to the thesis supervisor Associate Prof. Brijesh Adhikary, Head of the Department of Electrical and Electronics Engineering, Associate Prof. Dr. Hari Prasad Neopane, Head of the Department of Mechanical Engineering, and Associate Dean, School of Engineering, Prof. Dr. Ing. Ramesh Kumar Maskey at my principal institute Kathmandu University (KU) for their technical and social supports as a main referent.

My humble gratitude goes to Assistant Prof. Shailendra Kumar Jha, Kathmandu University, Assistant Prof. Rojesh Dahal, Nepal Engineering College and Er. Dipendra Mandal for their kind suggestions and encouragements with their professional and technical guidance.

My hearty appreciation goes to mechanical department attendant Mr. Shyam Tolange to help for the documentation, my colleagues specially Ashish Shrestha, Kshitiz Khanal, and friends at Kathmandu University.

My acknowledgement goes to my Kathmandu University, Turbine Testing Lab (TTL), KU and Khwopa College of Engineering for providing technical support to complete my dissertation.

Ajay Singh

List of Figures

Figure 1: Typical AC power supply system[1]

Figure 2: Distributed Generation sources [16]

Figure 3: Classification of steady-state models of different DG technologies[24]

Figure 4: Radial Distribution System(RDS) [25]

Figure 5: Distribution System - Network configuration[25]

Figure 6: Distribution System - Loop Configuration[25]

Figure 7: Single line diagram of radial distribution network

Figure 8: Flowchart for sizing and optimal location of DG units

Figure 9: IEEE 12 bus test system

Figure 10: IEEE 12 bus system voltages profile

Figure 11: Active and Reactive power losses in IEEE 12 bus system

Figure 12: DG unit size in kVA integrated to IEEE - 12 bus system

Figure 13: IEEE 12 bus system voltages profile before and after DG allocation

Figure 14: Active power losses in each bus of IEEE 12 bus system without and with DG allocation

Figure 15: Total active and reactive power losses of IEEE 12 bus system without and with DG allocation

Figure 16: Total active power loss of network at different location of DG unit in IEEE - 12 bus system

Figure 17: IEEE 15 bus test system

Figure 18: IEEE 15 bus radial distribution system bus voltages profile

Figure 19: Active and Reactive power losses in IEEE 15 bus system

Figure 21: Active power losses in each bus of IEEE 15 bus system without and with DG allocation

Figure 22: Total active and reactive power losses of IEEE 15 bus system without and with DG allocation

Figure 23: Total active power loss of network at different location of DG unit in IEEE - 15 bus system

Figure 24: IEEE 33 bus test system

Figure 25: IEEE 33 bus radial distribution system bus voltages profile

Figure 26: Active and Reactive power losses in IEEE 33 bus system

Figure 27: IEEE 33 bus system voltages profile before and after DG allocation

Figure 28: Active power losses in each bus of IEEE 33 bus system without and with DG allocation

Figure 29: Total active and reactive power losses of IEEE 33 bus system without and with DG allocation

Figure 30: Total active power loss of network at different location of DG unit in IEEE - 33 bus system

Figure 31: Bhaktapur - Balaju 21 bus distribution system

Figure 32: Bhaktapur - Balaju 21bus radial distribution system bus voltages profile

Figure 33: Active and Reactive power losses in Bhaktapur - Balaju 21 bus distribution system

Figure 34: : Bhaktapur - Balaju 21 bus system voltages profile before and after DG allocation

Figure 35: Active power losses in each bus of Bhaktapur - Balaju 21 bus system without and with DG allocation

Figure 36: Total active and reactive power losses of Bhaktapur - Balaju 21 bus system without and with DG allocation

Figure 37: Total active power loss of network at different location of DG unit in Bhaktapur - Balaju 21 bus system

LIST OF TABLES

Table 1: Backward / Forward Sweep Method [38]

Table 2: Maximum Loss reduction at individual buses

Table 3: DG size and location with power loss for different IEEE buses

Table 4: Voltage improvement of candidate buses in different bus system

Table 5: DG size and location with power loss for different Bhaktapur - Balaju 21 bus system

Table 6: Voltage improvement of candidate buses in Bhaktapur -Balaju 21 bus system

Table 7: Bus and line data of IEEE 12 bus system

Table 8: Bus and line data of IEEE 15 bus system

Table 9: Bus and line data of IEEE 33 bus system

Table 10: Bus and line data of Bhaktapur - Balaju 21 bus system

Table 11: Result of optimization for IEEE - 12 bus system

Table 12: Result of optimization for IEEE - 15 bus system

Table 13: Result of optimization for IEEE - 33 bus system for DG unit limitation

Table 15: Result of optimization for Bhaktapur - Balaju 21 bus system for DG unit limitation..

ABSTRACT

In power system, research area is widening and one is Distributed Generation (DG) integration in distribution system. It restructures the power system and also helps in distribution system planning. This dissertation addresses the issue of voltage margin improvement by integrating DG unit and, therefore, this dissertation presents the model to enhance the candidate bus voltage of distribution system by optimal integration of DG units. The technical analysis is performed for radial distribution system (RDS).

The proposed algorithm evaluates the base condition of loss and bus voltage and then generate populations of solutions at first stage. Optimized of DG unit size and locate the optimal bus can be done with the particle swarm optimization (PSO) method. The proposed paradigm works with the objective that minimize the active power loss under a constraint of voltage limit and DG size limit.

In IEEE - 12 bus system, optimal DG unit size is 306.01 kVA at power factor (pf) of 0.85 located at bus 9. The real and reactive power losses are reduced by 84.06 % and 84.39 % respectively. In IEEE - 15 bus system, the real and reactive power losses are reduced by 72.84 % and 76.31 % by integration of 1036.54 kVA, at pf 0.85, DG unit located at bus 4. In IEEE - 33 bus system, optimal location of DG unit is bus 26 and size of DG unit is 2564.84 kVA, at pf 0.85 that reduces the real and reactive power loss by 67.89 % and 62.91 %. In all buses system the candidate bus voltages are within voltage regulation limit.

In Bhaktapur - Balaju 21 bus system, bus 3 to bus 21 have voltage lower than 0.9500 pu (per unit). The optimal size of DG unit is found 521.56 kVA that delivers active power of 517.51 kW and located at bus 14. This improves the bus voltage profile and all buses voltage are within specified limit. Real and reactive power losses are reduced by 90.58 % and 89.89 % respectively.

Finally, the conclusion is that DG unit can locate optimally with optimized size to the distribution system for the voltage profile improvement of the buses and also the total loss of system is reduced. Ii is also concluded that the model can be used for any radial distribution systems.

LIST OF ABBREVIATIONS

illustration not visible in this excerpt

LIST OF SYMBOLS

illustration not visible in this excerpt

CHAPTER 1 INTRODUCTION

1.1 Introduction

There is rapid increase in consumption of electricity in major sectors (Residential, Commercial, Industrial, Transportation, and Agriculture) which leads the electricity field to find more ways of electrical energy production with its efficient use and management. But, the traditional electrical generation system consists of large power plants (such as hydropower, thermal, nuclear power plants) and these power plants are located very far from load centers and power transfer to the end - use consumers are done through the transmission and distribution system. The typical power transmission system is shown in Figure 1 [1]. The growing load demand necessitates new generation power plants and expansion of transmission and distribution system which is neither be recommended for economic or environmental perspective. Therefore, integration of small generating plants at the load centers can be developed to meet the load demand. The small scale power generation plants connected at the load center or consumer side are called as distributed generation (DG) and these are fueled by renewable energy resources.

The growth of electricity demand availabilities of distributed generation resources and technologies and recent environmental policies provide the basis to use the distributed generation in combination of utility structure. This also provides the utility for restructure planning and modular power technologies that can quickly respond electricity demand. Some positive benefits of the installation of DG system are energy loss reduction, voltage profile augmentation and reliability enhancement [2]. Generally, DG systems are designed purely to deliver power to consumers, introducing DG in utility structure will change the characteristics of distribution network system due to the bi-directional power flow [3].

illustration not visible in this excerpt

Figure 1: Typical AC power supply system\Y\

1.2 Statement of the Problem

Distribution system network does not suffer any stability problems when substation supplies the demanded active power and reactive power to the load. However, stability problems are encountered in distribution system network by power supply scarce or mismatches and that can be remedial with introducing integration of DG units. While distribution system integrated with DG, static type DG units that are equipped with power electronics converters and rotary type DG units that are equipped with directly connected induction motor both affects the voltage stability margin of the system.

The higher penetration of DG units in utility structure can create a problem of reverse power flow through the line because of higher voltage at consumer side [4]. The improper placement and size of DG units may lead to the instability and more power losses in a system [5]. The integration of large amount of distributed generation resources may cause operational conflicts for a power distribution network [6]. The integration of DG units has impact on power quality among which harmonic resonance is the most significant. The output of transformer and grid impedance cannot be ignored in a grid having weak buses. When the impedance of grid and interfaced DG impedance is matched, then inverter voltage decays rapidly while grid voltage and inverter current are the major cause of harmonic resonance. The voltage of system increases beyond its typical limit 1±0.05 pu due to parallel harmonic resonance. Current in the system increases beyond its thermal limit due to series resonance.

1.3 Research Gap and Research Questions

Now-a-days DG is more popular in power system research for modular power station that can be used as power supply system to meet the electricity demand. Research area widens for DG unit integration into distribution system. The placement of DG unit into radial distribution system and its optimal size for improvement of power system quality can be studied. Study and investigations of DG unit impacts on distribution network operation can be made. The distribution system of Nepalese Power System is radial type and impacts of DG unit penetration in radial distribution system can be studied. In Nepal, the DG unit is placed and / or planned to be placed at substation of consumer side to meet the demand. Oversized DG unit is designed for the placement which may cause overvoltage at substation and no any improvement of voltage at farthest end of radial distribution system. There is a research gap for optimal DG unit size and planning for DG unit location to improve the voltage profile of the system.

The research questions are:

1. How DG unit size and location can be determined improve voltage profile margin in radial distribution system?
2. Is DG unit location being appropriate to improve voltage profile margin and reduce system loss?

1.4 Objective

Motivated from such above problems, the major objective of this thesis is to study enhancement of voltage stability margin of distribution system by optimal integration of DG units. The specific objectives are:

- Load flow analysis of distribution system using sweep algorithm.
- Optimal sizing and placement of DG units in the distribution system using particle swarm optimization (PSO) techniques.

1.5 Scope

This thesis is advantageous for the planning of power system with distributed generation and also restructure a distribution system with two-way power flow. The planning is obstructed with many factors and most common factors are proper sizing of DG unit and its allocation is radial distribution system. The optimization problem of DG unit placement and its capacity determination can be done using this study.

1.6 Limitation

This dissertation discusses on the DG unit allocation (size and location) in the distribution system but limited to technical discussion on voltage profile improvement and power loss minimization. It does not consider thermal limit, current handling capacity, network sensitivity effect of impedance matching and harmonic resonance. It only provides information about single DG allocation in radial distribution system (RDS), not the multiple DG allocation in RDS or any other distribution network configuration. It only specifies the size of DG i.e. active power (P) and reactive Power (Q) of DG, not the what type of DG actually can be used. It does not specify the protection system that can be used for DG unit in association with grid distribution system. The reconfiguration of distribution system affects the power flow in networks and this reconfiguration is also not studied in this thesis. The variable renewable energy (VRE) cannot supply constant power at all the times and that effect is not considered.

1.7 Thesis Organization

The dissertation consists of six chapters and the remainder of the thesis are as follows:

Chapter 1 presents a brief introduction of power system along with DG. The problems occurred while DG integration into grid is discussed and presents the objective of this dissertation along with scope and limitation of the dissertation.

Chapter 2 presents a literature review on background of distributed generation, power system distribution configurations, power system stability, power flow analysis in distribution system and optimization techniques. In distributed generation, the different types of DG based on specification of P or Q and renewable and non - renewable resources based DG types are discussed. The different configurations used for distribution system are discussed. The power system stability and its types are introduced and also power flow method or algorithms are discussed. It also provides different optimization techniques used in traditional and modern systems and its mathematical formulation.

Chapter 3 proposes the methodology for DG allocation and optimal size for improvement of candidate buses voltage with minimized power loss. It first introduces the base condition calculation of voltages and power loss and then formulation of optimization method for allocation of DG unit. It provides the information about the methodological tools used for proposed paradigm.

Chapter 4 studies the DG allocation in IEEE - 12, IEEE - 15, and IEEE - 33 bus based on proposed methodology. In this chapter, demonstration of results is provided for IEEE test bus system.

Chapter 5 presents the study of DG allocation in Bhaktapur - Balaju 21 bus distribution system network and highlights the improvement of buses voltage with DG allocation. It also discusses on the behaviour of power loss and its reduction by DG allocation.

Chapter 6 provides the conclusion and summary with recommendations on future works.

CHAPTER 2 LITERATURE REVIEW

2.1 Background

The electricity production is in the flow of distributed - centralized and - distributed again with the development of technology and its advancement. Initially, the electric utilities were established in open territories and provide service in isolated mode granted with de facto monopolies without connection with other generating sources. The centralization process of generating sources was made by adjoining on another and interconnected grid system were made with various of peak load sharing, backup power and serving more areas with economic prices of electricity. The different distributed generation resources were innovated with technological advancement in power industry which again restructured the power utility decentralized. The deregulation of electricity made market competitive and introduction of DG unit in distribution system is more pronounced.

2.2 Literature Survey

N. Acharya et al. [5] proposes analytical approach to allocate the DG unit for minimization of total losses calculated from exact loss formula. It investigates the minimum losses with placement of DG unit in IEEE - 30, IEEE - 33, and IEEE - 69 bus system. IEEE - 33 bus distribution system has initially total active power loss of 211. 20 kW without and DG interfacing. The active power loss of the system is reduced by 47.33 % for IEEE - 33 bus system when DG unit size of 2.49 MW placed at bus 6 which is optimal location. M. Jegadeesan and V. Keerthana [7] also presents an analytical approach for DG unit placement with the objective that active and reactive power loss of the system should be reduced and voltage profile of the candidate buses should be improved.

M. Vatani et al. [8] presents the combination of analytical and genetic algorithm methods is used for optimal allocation of multiple DGs in a distribution network to minimize the system losses. In this study, the DGs active power, power factor, and location are simultaneously considered during distribution network losses minimization. The proposed method is applied to IEEE - 33 bus and IEEE - 69 bus test distribution systems with two scenarios i) unity power factor mode of DG operation, and ii) non - unity power factor mode of DG operation. For IEEE - 33 bus system, the proposed method results the optimal location of DG unit at bus 6 with 2706.73 kW power delivery at unity power factor by the DG unit. Active power loss of the system is found 99.22 kW and shows that the proposed method has better performance in loss reduction over the analytical method and / or improved analytical method. This paper also shows the result that DG unit allocation at bus 30 which delivers an active power 1844.85 kW at power factor 0.767 lagging can reduce the IEEE - 33 bus network loss to 60.29 kW. The combined analytical and GA method has better performance over analytical, loss sensitivity factor (LSF) and exhaustive load flow (ELF) methods.

M. Sohi et al. [9] presents the Bee Colony Optimization (BCO) optimization method for optimal DG placement and sizing. The objective function of the propsed method is cost function based on loss reduction and capacity of the lines as indexes. With one DG unit, the optimization method proposed is tested in IEEE - 33 bus system which results DG placed at bus 5 of unit size 2.3970 MW. Active power loss at this condition is found 0.060 MW. In this paper multiple DG unit placement is also presented. Multiple DG unit placement is also presented in this paper which shows that the total real power loss of the system gets reduced by DG unit placement.

K. Vinothkumar andM.P. Selvan [10] presented Fuzzy embedded genetic algorithm for distributed generation planning. The proposed method employs fuzzy set theory and the genetic algorithm for formulation and evaluation of a multi-objective function, respectively, for optimal planning of distributed generator units. In this paper, the operation of distribution system and optimal allocation of DG unit considered under constraints voltage deviation index (VDI), real power loss index (PLI), reactive power loss index (QLI), line loading index (LLI), and short circuit index (SCI). Proposed method has objective function to minimize CDI, PLI, and QLI under constraint of SCI and LLI must be within its maximum limit. Using proposed method, IEEE - 25, IEEE - 33, and IEEE - 66 bus system are studied and it is concluded that presented method can be used for identification of grid integration points and size of DG units.

F.J. Ruiz-Rodriguez et al. [11] considers the technical constraint as voltage regulation in radial distribution system when distributed generation as photovoltaic system is integrated into grid. This paper proposes a binary PSO method that can be applied for DG sizing and sitting into the distribution system.

K. R. Guerriche et al. [12] proposes PSO method for optimal placement and sizing of distributed generation units in distribution system. The methodology of this paper is mainly focused on maximum loss reduction and voltage profile improvement. In this paper, the efficiency and robustness of proposed PSO is shown for DG unit allocation and enhancing loadability of radial distribution system.

M. Gomez-gonzalez et al. [13] introduces a hybrid approach which employs discrete particle swarm optimization (PSO) and optimal power flow. It considers different technical constraints, such as voltage limits, thermal limits on lines and transformers, operational and planning limits and maximum level of penetration of DG with objective function as sum of costs of connecting a traditional generator, traditional thermal generator cost, and unit cost of DG.

Taha Jabbar Sahib et al. [14] proposes PSO method for DG unit allocation in radial distribution system. The objective function for the proposed method is minimization of total power losses and voltage profile improvement without any constraints violation. The constraints under consideration are voltage regulation limit and current handling capacity of line. The results in this paper shows that the total real power losses can be minimized significantly by allocating the optimal size of DG in the optimum placement.

2.3 Distributed Generation (DG)

In literature, large number of terms and definitions is used in relation to distributed generation. Anglo-American countries often use the term ‘embedded generation’, North American countries the term ‘dispersed generation’, and in Europe and parts of Asia, the term ‘decentralized generation’ is applied for the same type of generation [15]. The different organization working for electrical, electronics and energy system defines distributed generation as:

IEEE defines DG as “the generation of electricity by facilities that are sufficiently smaller than central generating plants so as to allow interconnection at nearly any point in a power system”. CIGRE and CIRED provides more precise definition of DG based on size, location and type. It defines DG as “all generation units with a maximum capacity of 50MW to 100 MW, that are usually connected to the distribution network and that are neither centrally planned nor dispatched”. CIRED defines DG to be “all generation units with a maximum capacity of 50 MW to 100MW that are usually connected to the distribution network”. Ackermann’s definition is the most generic one, because there is no limit on the DG size and capacity. The definition covers the location of the DG. It defines DG as “an electric power generation source connected directly to the distribution network or on the customer side of the meter” [16].

2.4 Distributed Generation Resources

Based upon the distributed generation technologies used in distribution systems, it is mainly two types: dispatchable and non-dispatchable. The Figure 2 shows the different DG resources [16].

2.4.1 Dispatchable Energy Resources

Generally, dispatchable renewables are constantly available (apart from maintenance needs) for production of energy at high capacity factor [17]. These energy resources are presented briefly as:

Micro - turbines

Micro-turbine generator system consists of a gas turbine, a compressor, and an ac generator that is to be capable of delivering clean energy from a wide variety of fuels and low emissions [18]. The capacity of micro-turbine ranges from several kilowatts up to megawatts. The advantages of micro-turbines are that they are small in size, have high efficiency, low noise, low gas emission, and low installation and maintenance costs. Micro - turbines generate a very high frequency power (1500-4000 Hz) compared to the nominal power frequency, which is 50Hz or 60 Hz. Therefore, the turbines require power electronics converter (ac-dc-ac converter) to interface them to the grid or to operate them in parallel with other types of DG sources [16], [18].

illustration not visible in this excerpt

Figure 2: Distributed Generation sources [16]

Internal Combustion Engines

A reciprocating engine that produces a work from a combustion of working fuel is Internal Combustion Engine [19]. The Internal Combustion Engines are majorly used in transportation and generator applications [20]. Now - a - days Internal Combustion Engine combined with synchronous generator or induction generator is the most commonly applied in DG technology and it can be connected directly to the grid without any power electronics interfacing [21].

Combustion Engines

Majorly gas turbine is used as combustion engine for the application of DG. Combustion engine are fueled by natural gas which are burnt externally that rotates the turbine based on pressure difference [19].

Fuel Cell

An electromechanical device that coverts chemical energy into electrical or thermal energy. It is unlike a battery in that it does not need to be charged for the consumed materials during the electrochemical process since these materials are continuously supplied to the cell. The efficiency of these cells is significantly high, about 40-60% when used for electricity production. Moreover, when the exothermic heat is combined with electrical power (CHP), the overall efficiency can reach more than 80% [16].

2.4.2 Non-Dispatchable Energy Resources

Non-dispatchable energy resources, also known as Variable Renewable Energy (VRE), depends on meteorological conditions. These resources are variable due to temporal availability of resources with uncertainty, location-specific properties due to the geographical availability of resources, and low marginal costs since the resources are freely available [20].

Photovoltaics (PV)

Solar energy is the most common renewable energy resource followed by wind energy, both because of their ubiquity [22]. A typical silicon PV module is made up of 36 or 72 cells depends upon the voltage. The modules are then connected in series/parallel configurations to form a solar array that is used to generate electrical energy from the sunlight. Photovoltaics DG units are interfaced with power system by means of power electronics converters / inverter [16], [22].

Wind Turbine

The produced mechanical energy from wind turbine is proportional to the area of wind turbine blade, density of wind, cube of velocity of wind. This mechanical energy can be used for specific tasks such as grinding grain or pumping water or for driving a generator. Generator converts the mechanical energy into electrical energy, using ac generators (induction and synchronous machines) or dc machines that are attached to the wind turbine [16][23].

2.5 Distributed Generation Models

Different DG technologies can generally be classified into three main types based on their active and reactive power generation characteristics, as illustrated in Figure 3 [24]. It consists of three types of DG in which type I has unspecified active power (P), type II has unspecified reactive power (Q) and type III has unspecified both active and reactive power i.e. unspecified PQ. Combining different energy sources with different energy converters represents special DG generation characteristics for each configuration. The bounds of the decision variables, the active and reactive DG powers, are specified for each DG type [24].

illustration not visible in this excerpt

Figure 3: Classification of steady-state models of different DG technologies[24]

2.6 Power Distribution Systems Configurations

An important characteristic of distribution systems is their configuration, or how their lines are connected and basically three common configurations of distribution systems: radial, loop and network [25].

illustration not visible in this excerpt

Figure 4: Radial Distribution System(RDS) [25]

Radial Distribution System (RDS)

The radial distribution system configuration is shown in Figure 4 [25]. The characteristics [26], [27] of this distribution configurations are:

- power flows in one direction,
- lowest capital cost,
- lowest reliability,

A distribution system in which there is only one path between costumer and substation is called radial feeder system. The advantages and disadvantages of RDS are:

Advantages

- Low cost: this is least expensive type of distribution,
- Simplicity of analysis: It is easiest to analyze and operate the system.,
- Less reliable: Any equipment failure will interrupt service to all consumers downstream from it,

A cleaver design and planning of radial distribution system can achieve a fair degree of reliability even without much addition of cost.

In some cases, feeder systems are constructed as a network and operates radially, In Y-connected radial systems, the neutral conductor is connected through all open switch points forming a network connecting feeder and substations.

Network Distribution System

The networked configuration distribution system shown in Figure 5 [25] and has following characteristics[26], [27]:

- more interconnected between two points,
- more than one path and some lines form loops within the system,
- more reliable,
- highest cost,

The distribution network involves multiple paths between all points in the network. Power flow between any two points is usually split among several paths, and if a failure occurs it instantly and automatically re-routes itself. Rarely does a distribution network involve primary voltage level network design, in which all or most of the switches between feeders are closed so that the feeder system is connected between substations.

The major advantages are that it provides very high level of reliability. The loss of any source will not interrupt the flow of power to any customers. The multiple failure of sources can occur with little or no interruption. Among their disadvantages, feeder network systems costs considerably higher and much more complicated analysis and operating procedures.

illustration not visible in this excerpt

Figure 5: Distribution System - Network configuration[25]

Loop Distribution System

The loop configured distribution system shown in Figure 6 [25] and has following characteristics [26], [27]:

- fall in between the two (radial and network) in terms of cost and reliability,
- two radial systems separated by a normally open switch,
- two paths from substation to the load,

Distribution systems can be operated as loop systems in which the two paths consist in between the consumers and substation. There is being a “null point” somewhere on the loop where no power passes. This layout is basically dynamic radial system with open point (null point) shifting as loads change. A loop must be able to meet all power and voltage drop requirements when fed from only one end, not both.

Advantages

- High reliability than radial system,

- In terms of complexity, a loop feeder system is only slightly more complicated than a radial system,

- Major disadvantages are capacity and cost of the loop system,

illustration not visible in this excerpt

Figure 6: Distribution System - Loop Configuration[25]

2.7 Power System Stability

The stability of a power system is defined as “a property of a power system that enables it to remain in a state of equilibrium under normal operating conditions and to regain an acceptable state of equilibrium after being subjected to a disturbance” [28]. The operating condition of a power system is described according to physical quantities, such as the magnitude and phase angle of the voltage at each bus, and the active/reactive power flowing in each line. If these quantities are constant over time, the system is in steady state; if they are not constant, the system is considered to be in disturbance [29]. The disturbances can be small or large depending on their origin and magnitude. For instance, small variations in load and generation are types of small disturbances, but faults, large changes in load, and loss of generating units are types of large disturbances [29]. Any fault occurring in the distribution system might cause voltage and angle instability [30].

2.7.1 Rotor Angle Stability and Frequency Stability

Rotor angle stability is “the ability of interconnected synchronous machines in a power system to remain in synchronism under normal operating conditions and after being subjected to a disturbance” [31]. Frequency stability is the ability of a power system to maintain the frequency within an acceptable range following a system upset that results in a significant imbalance between generation and load [31]. The DG units are assumed to be dispatchable in order to control both the reactive and real power during the transition from grid-connected to autonomous mode. The results showed that interfaced DG units can effectively enhance power quality and also maintain angle stability. However, the micro-grid was assumed to be a balanced system [16]. The analysis of frequency stability and rotor angle stability both are not the scope or objective of this thesis.

2.7.2 Voltage Stability

Voltage stability refers to the ability of a power system to maintain steady and acceptable voltages at all buses in the system after being subjected to a disturbance from a given initial operating condition. The main factor causing voltage instability is inability to meet the reactive power demand [28]. It is helpful to classify voltage stability into two categories: large - disturbance voltage stability and small -disturbance voltage stability. Large - disturbance voltage stability is concerned with the system’s ability to control voltages following large disturbance such as system faults, loss of load, loss of generation [28]. The voltage stability can be evaluated by two different methods of analysis: static and dynamic [16], [28].

1. Static Analysis

This method examines the viability of the equilibrium point represented by a specified operating condition of the power system. This method allows the examination of a wide range of system conditions. The electric utility industry depends on P-V and Q-V curves in order to determine stability at selected buses. The static method is evaluated by means of a variety of techniques such as

a) P - V and Q - V curves

P-V and Q-V curves are generated by executing a large number of power flows using power flow methods. In this case, a power system is typically modeled with non-linear differential algebraic equations [16], [32].

illustration not visible in this excerpt

Where,

x represents state vector including bus voltage magnitude (V) and angle (5).

X is a parameter vector that represents the real and reactive power demand at each load bus.

b) V - Q Curves

In this method, the network is represented by a power flow equation that can be linearized, as given in equation (2) [16].

illustration not visible in this excerpt

Where,

AP, AQ, AV and A9 are incremental change in real power, reactive power, bus voltage magnitude and bus voltage angle respectively.

J is Jacobian Matrix.

With the real power of load constant, the incremental change in real power AP = 0. Then using partial inversion of equation (2) gives,

illustration not visible in this excerpt

The V-Q sensitivity can be calculated by solving equation (3). The V-Q sensitivity at a bus represents the slope of the Q-V curve at a given operating point. A positive V-Q sensitivity is indicative of stable operation, and a negative sensitivity is indicative of unstable operation.

2. Dynamic Analysis

Dynamic analysis can show the real behaviour of the system such as loads (dynamic and static), DG units, automatic voltage and frequency control equipment, and the protection systems. The overall power system is represented by a differential equations, as given in equation (5). set of first order

illustration not visible in this excerpt

and set of algebraic equations

illustration not visible in this excerpt

With the known initial conditions (X0, V0).

2.8 Effect of DG unit on Voltage Profile and Loss of the System

Tap changing transformers in substation and / or capacitor banks are used to maintain the voltage regulation limit of the system under an assumption that power flows from substation to the loads. The DG unit placement in distribution system impacts the power flow and voltage conditions at consumers of the utility equipment. Distribution system operation and DG unit characteristics decides positive and negative effect of DG unit [33], [34]. The improper placement of DG unit may lead to more power loss of the system and overvoltage at consumer side. The power flow from consumer side to substation is not needed at normal operation of distribution system. This is being caused by improper placement of DG unit [4], [5], [33]. In comparison of capacitor bank, DG unit is capable of delivering both active and reactive powers and hence DG unit will provide substantial reduction in losses.

2.9 Power Flow Analysis in Distribution System

The power-flow analysis of a distribution feeder is similar to that of an interconnected transmission system [35]. The distribution networks because of the some of the following special features fall in the category of ill-condition [36].

- Radial or weakly meshed networks,
- High R/X ratios,
- Multi- phase, unbalanced operation,
- Unbalanced distributed load and/or distributed generation,

Due to the above factors the Newton Raphson (N-R), Gauss Siedel (G-S) and other transmission system algorithms are failed with distribution network. Because a distribution feeder is radial, iterative techniques commonly used in transmission network power-flow studies [35]. Instead, an iterative technique specifically designed for a radial system is used. The sweeping algorithm is iterative technique and has the advantages of less computation effort and calculation time compared to the N - R and G - S methods [37].

2.9.1 Backward / Forward Sweep Algorithm

The unique path from any given bus back to the source in RDS is the key feature exploited by the backward / forward sweep algorithm [38]. The sweep algorithm consists of two basic steps, backward sweep and forward sweep, which are repeated until convergence is achieved. The backward sweep is primarily a current or power flow summation with possible voltage updates. The forward sweep is primarily a voltage drop calculation with possible current or power flow updates [38].

Table 1: Backward / Forward Sweep Method [38]

illustration not visible in this excerpt

In distribution system, the sweeping process includes different methods and are power summation method, current summation method, and admittance summation method [37].

Power Summation Method

The sweep algorithm using power summation method [36] is described in detail. The single line diagram of radial distribution network is shown in Figure 7 considering the active power (Pk) and reactive power (Qk) that flows through branch from node ‘k’ to node ‘k+1’.

Initially, a flat voltage profile is assumed at all nodes i.e., 1.0 pu.

Backward Sweep

The power flows through each branch is calculated in backward direction from last node and is given as equation (7),

illustration not visible in this excerpt

where,

illustration not visible in this excerpt

[illustration not visible in this excerpt]are the effective real and reactive power flows from node ‘k+1’ and [illustration not visible in this excerpt] are loads that are connected at node ‘k+1’.

Forward Sweep

The voltage magnitude and angle at each node are updated in forward direction using equation (8),

illustration not visible in this excerpt

where,

Ik is the current flows through branch and Zk is the impedance of branch connected from node ‘k’ to node ‘k+1’.

The backward and forward sweep equations are calculated iteratively until it converges.

illustration not visible in this excerpt

Figure 7: Single line diagram of radial distribution network

Current Summation Method

Given the voltage of the root bus and an initial voltage guess of other buses, the algorithm takes three steps for each iteration [37]:

Step 1: Nodal current calculation:

The current injection at each node ‘i' is calculated using equation (9).

Figure 7: Single line diagram of radial distribution network

Where,

i = 1, 2, 3, ..., n.

Si is the power at node i. Vi is the voltage at node I and yi is the shunt admittance at node i.

Step 2: Backward sweep:

Starting from the last ordered branch, current flow Ji in branch l is calculated using equation (10):

illustration not visible in this excerpt

Where,

l = b, b-1, .., 1.

Iir is the current injection of node ir calculated from step 1, [illustration not visible in this excerpt] is the currents in branches emanating from node ir.

Step 3: Forward Sweep:

Starting from the root bus, the node voltages are updated using equation (11):

illustration not visible in this excerpt

Where,

l =1, 2, ..., b.

is and Ir denote the sending and receiving end of branch l, Zi is the series impedance of branch l.

2.10 Optimization Techniques

Over the years, in order to solve the problem, many traditional and mathematical methods are studied. These methods include [39] branch exchange algorithm (BEA), optimal flow pattern (OPF) algorithm, mathematical optimization theory algorithm and artificial intelligent (AI) algorithm. Among them, artificial intelligent algorithm is widely used in distribution network reconfiguration in recently years, such as artificial neural network (ANN), genetic algorithm (GA), simulated annealing algorithm (SA), ant colony Optimization (ACO) and particle swarm optimization (PSO). The different approaches for optimization of distributed generation in distribution systems are presented in this section.

2.10.1 Analytical Method

This method is classical method previously used for optimization of DG unit. If only a given demand-generation snapshot scenario is taken into account, a specific technical aspect (or objective function) can be formulated analytically in such a way that it is possible to find the most beneficial DG unit capacity [40]. This type of analysis, focusing particularly on real power losses, was used in [5] and [41]. However, while power losses can be studied in passive networks considering peak load scenarios as is traditionally done distribution networks with DG plants require the assessment of energy losses. The incorporation of operational solutions such as coordinated voltage control or generation curtailment cannot be done either. Consequently, although analytical approaches are straightforward alternatives to assess DG unit siting and sizing, care must be taken as the results are only indicative and scenario limited [40]. The limitation of this process is that single DG unit can only be evaluated at a time.

2.10.2 Exhaustive Method

A single technical issue, such as voltage rise or power losses, can also be approached by exhaustively exploring the entire (or most of the) search space corresponding to the locations and sizes of DG plants that could be connected to a distribution network [40]. However, the actual benefit brought by exhaustive analyses is that it is possible to cater for a number of technical issues and constraints. Indeed, with this more direct approach the objective function can be the combination of parameters or indices that represent different technical and non-technical aspects, although it will be very time consuming [40], [42].

2.10.3 Linear Programming (LP) Method

Linear programming (LP) has also been employed to address the capacity allocation and energy optimization issues. Fundamentally, the use of linear programming entails a linearization of the power flow or the linearization of the results from an ac power flow. It has been demonstrated through simulation that the resulting approximation inevitably introduces an error, but not a significant one in the context of discrete turbine sizes [40]. A linear programming formulation of optimal power flow (OPF) is employed to assess the control of multiple DG plants. The objective employed is to minimize the annual active generation curtailment cost. Ac power flow is employed to calculate linearized sensitivity factors. The sensitivities are employed to characterize a range of constraints, such as voltage, thermal and short circuit limits. The method is formulated as a linear program and solved with the objective of maximizing the capacity of DG, subject to typical network constraints and taking account of N-1 configurations [24], [43], [44]. An advantage of LP is that it offers significant potential for development of operational methods and is a robust optimization method [40].

2.10.4 AC Optimal Power Flow (OPF) Method

The ac OPF [45] is used to widely acknowledge by the electric industry for powerful optimization tool. he ac OPF is a nonlinear programming (NLP) problem, for which many solution methods exist including some which are highly specialized to OPF problems. The ac OPF formulation can be adapted to have different objectives and constraints according to the study being carried out [40]. A number of solution methods can be adopted to solve the ac OPF problem: from special linear programming formulations to branch and bound techniques. Commercial solvers specialized for NLP problems include CONOPT, that uses a generalized reduced gradient, and, KNITRO, that uses interior points. Although no practical method exists which can guarantee to find the global optimum of a non-convex NLP, local optima can be found in most cases [40].

2.10.5 Metaheuristics Method

A metaheuristic method is defined as an iterative generation process which guides a subordinate heuristic by combining intelligently different concepts for exploring and exploiting the search space. Learning strategies are used to structure information in order to efficiently find near-optimal solutions [40]. There are numerous metaheuristic algorithms: ant colony optimization (ACO), artificial bee colony optimization (ABC), tabu search (TS), particle swarm optimization (PSO), simulated annealing (SA) including genetic algorithms (GA). All these algorithms have been used to solve the problem of optimal allocation of DG [40], [46], [47] and the PSO method is described in detail.

Artificial ants in ACO algorithm can be seen as probabilistic construction of heuristics that generate solutions iteratively by taking into account of accumulated past search experience: pheromone trails and heuristic information on the instance under solution. [48].

ACO and ABC are based on the dynamic of the social insect population. The interactions are executed via multitude of various chemical and/or physical signals (e.g., bee dancing during the food procurement, ants’ pheromone secretion, and performance of specific acts, which signal the other insects to start the same actions). The final product of different actions and interactions represents the behavior of social insect colony. ABC algorithms has been used for determining the optimal DG-unit’s size, power factor, and location in order to minimize the total system real power loss [9], [49].

TS is a metaheuristic that guides a heuristic method to expand its search beyond local optimality, with the systematic prohibition of some solutions to prevent cycling and to avoid the risk of being trapped in local minima. New solutions are searched in the set of the points reachable with a suitable sequence of local perturbations (neighborhood). One of the most important features of TS is that a new configuration may be accepted even if the value of the objective function is greater than that of the current solution. To prevent cycling, some moves are marked “tabu” for a number of iterations; the length of the tabu list, the tabu-tenure, fixed or variable, guides the optimization [40].

The SA is an algorithm that combines combinatorial search with a very simple metaheuristic that follows the cooling process of materials. Following an appropriate cooling schedule, the SA has the potential to avoid local minima and converges to the global minimum in a reasonable computing time. The parameters to tune are the annealing temperature, the number of iterations at constant temperature and the cooling strategy. SA annealing has been used for multi-objective optimization to minimize energy losses, polluting emissions and contingencies [40].

GA mimics the process of evolution. The most promising individuals have greater chances of transmitting their genes to offspring. By so doing, the population, generation by generation, improves and, if the premature convergence is avoided, for instance, with a random mutation, the algorithm converges. GA have been used by the first authors that pioneered the problem of the optimal integration of distributed energy resources in the distribution system and since then it has been preferred to other metaheuristic algorithms. The reason of the success is that GA is intrinsically suited to solve location problems. The coding of a solution can be very simple, a binary vector with as many positions as the number of bus candidate to DG connection. The classical GA operators (selection, crossover and mutation) can be used simply and effectively with few or no changes. As in many metaheuristic algorithms, high values of penalty factors can be added to the fitness function of those individuals that do not comply with the constraints [40].

PSO makes use of a velocity vector to update the current position of each particle in the swarm. The position of each particle is updated based on the social behavior that a population of individuals, the swarm in the case of PSO, adapts to its environment by returning to promising regions that were previously discovered [40]. The particle swarm optimization concept consists of, at each time step, changing the velocity (accelerating) each particle toward its pbest and gbest locations (global version of PSO). Acceleration is weighted by a random term, with separate random numbers being generated for acceleration toward pbest and gbest locations [50]. The process for implementing the global version of PSO is as follows [50]:

i. Initialize a population (array) of particles with random positions and velocities on d dimensions in the problem space.

ii. For each particle, evaluate the desired optimization fitness function in the variables.

iii. Compare particle's fitness evaluation with particle's pbest. If current value is better than pbest, then set pbest value equal to the current value, and the pbest location equal to the current location in d-dimensional space.

iv. Compare fitness evaluation with the population's overall previous best. If current value is better than gbest, then reset gbest to the current particle's array index and value.

v. Change the velocity and position of the particle according to equations (12) and (13) respectively.

illustration not visible in this excerpt

The acceleration constants c1 and c2 in equation (10) represent the weighting of the stochastic acceleration terms that pull each particle towards pbest and gbest positions. Therefore, adjustment of these constants changes the amount of "tension" in the system. Low values allow particles to roam far from target regions before being tugged back, while high values result in abrupt movement toward, or past, target regions.

vi. Loop to step ii) until a criterion is met, usually a sufficiently good fitness or a maximum number of iterations (generations).

2.10.6 Multi-Objective Programming Method

Multi-objective optimization problems are solved by two fundamentally different groups of techniques. The first set of techniques uses preference information and the iterative repetition of a single-objective optimization problem, usually solved by GA. The most common techniques of this first group are the weighted-sum method and the s-constrained method [40].

2.10.7 Probabilistic Method

Uncertainties related to DG are due to two main aspects: the variable nature of the primary energy source and the possible unavailability of the unit when it is required to generate. The combination of these aspects may lead to generation deficit, which can heavily compromise the security, reliability and quality of power supply. The increase in the complexity of distribution systems with DG requires the assessment of the random nature of network failures and generation availability [40], [51]. In order to adequately address the uncertainties introduced by DG integration to distribution systems, probabilistic methods can be applied for network planning and optimization. Besides, stochastic models of renewable resources must be developed in order to represent the influence of the primary energy source variability on generation availability [52]. However, probabilistic approaches have two most discussed disadvantages: the large amount of data required and the potential difficulty in interpreting the results and thus make decisions based on such results. The probabilistic reliability evaluation or system planning requires the adequacy analysis of several system states or expansion alternatives that are performed by optimization methods [40], [51], [53].

METHODOLOGY

3.1 General

The successful planning of integration of DG in distribution system must be covered by all static and dynamic analysis in technical and economic aspects. DG unit is optimized when system has minimum loss under a given constraint of technical aspects, environmental aspects and economic aspects. Only the technical aspects of voltage stability or marginal improvement of voltage and network minimum loss are considered for optimal DG placement in radial distribution system and the other environmental and economic aspects are not considered in this dissertation. The methodology uses iterative techniques of sweeping process for load flow analysis and PSO, a metaheuristic approach, for optimization of size and location of DG unit. The objective function as a function of active power loss which is to be minimum under given constraints: the voltage regulation limits and DG size limits. The details of proposed methodology are described herein.

3.2 Proposed Methodology

The complete methodology for analysis and planning of DG integration in radial distribution system is represented as flow diagram in Figure 8. The technical analysis performed for RDS insights into the weakness and strengths of the system with power losses in each buses, and respective bus voltage and hence the integration planning process become decisive for the outcome.

The proposed methodology starts with the input of line data and bus data of the RDS. The voltage of each buses of network is set to 1.00 pu and then the branch currents are calculated by backward sweep propagation using equation (9) and (10). The magnitude of voltage and phase angle is updated using equation (11) in forward sweep propagation. The backward and forward sweep propagation is iteratively continued until the convergence criteria is met. In the proposed methodology, convergence criterion is to iterate the backward and forward sweep propagation for 100 iterations. The updated voltages and currents with the computed power losses are taken for base condition values for the RDS network without DG integration.

illustration not visible in this excerpt

Figure 8: Flowchart for sizing and optimal location of DG units

The minimum and maximum sizes of DG units are defined and then the population of particles are generated within these minimum and maximum values. The constraints under consideration given in equation (14) and (15) are considered arbitrarily which defines the limitation of size of DG unit considering that active power supplied by the DG unit is less than or equal to 60 % of the total load with DG unit power factor at 0.85 and constraint defined for voltage regulation limit in equation (16) is considered from rule 40 of Electricity Rules, 2050 (1993) [54] and IEEE standard [55].

illustration not visible in this excerpt

The objective function or fitness function of proposed methodology is to minimize the total active power loss under given constraints and conditions. The objective function is the sum of total losses of the RDS and is given in equation (17). The power loss (Pioss) of the branch is evaluated from branch current (I) and branch resistance (R) as [illustration not visible in this excerpt]

illustration not visible in this excerpt

The fitness function is evaluated for each particle of population and after the initialization of personal best (Pbest) and global best (Gbest), the position and velocity of a particle is updated using given equations (18) and (19) and then Pbest and Gbest are updated.

illustration not visible in this excerpt

The optimized solution is evaluated after number of iterations defined. In this thesis, number of iteration is defined 1000 and the inertia weight (w) is varied for each iteration using equation (20). The inertia weight has maximum and minimum value of 0.4 and 0.9 respectively. The optimal solution gives the minimum loss under subjected constraints, position of DG with its’ size and individual bus voltages and losses. The detail of PSO is represented in APPENDIX A.

3.3 Methodological Tools

The tools used for proposed methodology is MATLAB (MATrix LABoratory) developed by MathWorks and is a multi-paradigm numerical computing environment that works with matrix and array mathematics directly. In MATLAB, Math operations are distributed across the computer’s cores, library cells are heavily optimized, and all the codes are just-in-time compiled [56]. The algorithm is coded in a MATLAB scripting language and processed for testing and debugging in PC with 64GB RAM, 2.54 GHz processor.

DISTRIBUTED GENERATION (DG) UNIT INTEGRATION PLANNING IN IEEE BUS TEST SYSTEM

4.1 General

As described in chapter 3, the proposed methodology uses real power loss as an objective function which should be minimized under given constraints. The proposed methodology for optimization of DG size and its placement formulation, DG integration planning in RDS, is coded in MATLAB scripting language and is tested for standard test systems IEEE - 12, IEEE - 15 and, IEEE - 33 bus distribution system. The IEEE bus test systems are standard distribution systems and are considered for the benchmark of the study in this thesis for validation and verification of the model and method. The three test systems IEEE - 12, IEEE - 15 and, IEEE - 33 bus system are chosen because of its different branching characteristics and load variation at buses. The proposed methodology code is firstly verified for IEEE standard bus systems and then it is applied for Nepalese distribution networks. In this thesis, the Bhaktapur -Balaju 21 bus distribution system network is considered for study purpose. This system has 11 kV distribution voltage and located at Bhaktapur and Balaju territory.

The DG types defined are type I (Unspecified P), type II (Unspecified Q), and type III. However, in this thesis, type III (Unspecified PQ) is chosen for the placement of DG unit in IEEE - 12 bus, IEEE - 15 bus and IEEE - 33 bus test systems are majorly. There is a specified minimum and maximum values of P and Q. Under maximum and minimum values, the P and Q are optimized. This may be any dispatchable and / or non-dispatchable type renewable energy resources that can be able to deliver specified active and reactive power.

4.2.1 Load Flow Analysis of IEEE 12 Bus Test System

The radial distribution system network of IEEE 12 bus system is presented in Figure 9 and it has no any point where current is sub-divided into more branches. The farthest end bus of this RDS is bus 12. The reference bus 1 has no any load connected to it and it acts as substation for the distribution system. It is assumed and also found that the voltage of reference bus is always 1.0000 pu with no any power losses into it. Maximum load of the system is found at bus 2 with active and reactive load of 60 kW and 60 kVAR respectively. The total load of the system is 435 kW and 405 kVAR distributed at different buses non - uniformly. The line data and load data of the network is tabulated in Appendix B.

illustration not visible in this excerpt

Figure 9: IEEE 12 bus test system

The voltage profile and real and reactive power losses in IEEE - 12 bus distribution system resulting from the from the forward / backward sweep method of load flow using base voltage of 11 kV and base power of 100 MVA are presented in Figure 10 and Figure 11 respectively . The maximum bus voltage is found 1.0000 pu of reference bus i.e. bus 1, and minimum bus voltage is found 0.9467 pu for bus 12. Buses 10, 11 and 12 have voltage less than 0.9500 pu. Voltage of the buses are in decreasing fashion from reference bus 1 to the end bus 12. The maximum loss is found at bus 5 with active power loss 4.2019 kW and reactive power loss of 1.7575 kVAR. The current is also decreasing fashion and there is a variation of resistance and reactance of the branches and thus I2 R and I2 X is varied for each bus and found maximum losses at bus 5. The total active power loss of the system is found 20.2862 kW and reactive power loss of 8.0344 kVAR.

illustration not visible in this excerpt

Figure 10: IEEE 12 bus system voltages profile

illustration not visible in this excerpt

Figure 11: Active and Reactive power losses in IEEE 12 bus system

Bus Number

4.2.2 Results of Optimal Allocation of DG unit in IEEE 12 Bus Test System

The paradigm described in section 3.2 is applied for IEEE - 12 bus distribution system with objective function as active power loss which should be minimized for the entire system and loss must be less than the base case loss of 20.2862 kW and 8.0344 kVAR under the constraints that (i) bus voltage limit (1.0000 ± 0.05) pu, and (ii) active power supply from DG unit must be less than 60 % of total active power load at power factor of 0.85 lagging (DG size ≤ 261 kW, 162 kVAR). The optimal size of DG unit (Figure 12) is found 260.28 kW and 160.93 kVAR located at bus 9 which results the minimum loss under provided constraints, improved voltage of each buses. After the allocation of DG unit at bus 9, the maximum bus voltage is found 1.0000 pu of reference bus and minimum bus voltage is found 0.9914 pu of bus 6. The bus voltages profile plot of the system for without DG unit and with DG unit placed as optimal location is presented in Figure 13. Voltage of buses 10, 11 and 12 are improved and also other buses voltage are also improved. All the voltage of buses lies within a specified limit with DG unit allocaction.

illustration not visible in this excerpt

Figure 13: IEEE 12 bus system voltages profile before and after DG allocation

Active power loss in each buses is presented in Figure 14 where it can be seen that active power loss at bus 5 is drastically reduced by installation of DG unit. The other buses 2, 3, 4, and 8 also have more losses initially and those losses are also reduced. The maximum loss reduction at individual buses of IEEE - 12 bus system is represented in Table 2. Though bus 5 has maximum loss reduction of 3.9261 kW, bus 7 has maximum loss reduction of 96.43 %. After allocation of DG unit at bus 9, the maximum active power loss and reactive power loss are found 0.8234 kW and 0.3477 kVAR at bus 2. The total active and reactive power losses of the system with DG unit allocation are 3.2335 kW and 1.2545 kVAR respectively and is presented in Figure 15 with corresponding losses without DG unit. When optimal DG unit size 306.01 kVA, at pf 0.85, is located at different buses then the active power loss of the system is presented in Figure 16. The active power loss varies with DG unit location due to voltage of the buses at that case and current flowing during that time. The total network active power loss is minimum when DG is located at bus 9 and maximum when DG unit is located at bus 2. The active power loss of the system is reduced by 84.06 % and reactive power loss of the system is reduced by 84.38 % with installation of DG unit (DG size = 260.28 kW, 160.93 kVAR) at bus 9. Individual candidate buses voltage for different location of DG units are represented in APPENDIX D.

illustration not visible in this excerpt

Figure 14: Active power losses in each bus of IEEE 12 bus system without and with DG allocation

Table 2: Maximum Loss reduction at individual buses

illustration not visible in this excerpt

illustration not visible in this excerpt

Figure 15: Total active and reactive power losses of IEEE 12 bus system without and with DG allocation

illustration not visible in this excerpt

Figure 16: Total active power loss of network at different location of DG unit in IEEE – 12 bus system

4.3 DG integration in IEEE 15 Bus test System

4.3.1 Load Flow Analysis of IEEE 15 Bus Test System

The RDS network of IEEE - 15 bus system is different from IEEE - 12 bus system in terms of branching. IEEE - 15 bus system is presented in Figure 17 has branches from bus 2, bus 3, and bus 4 with buses 6, 7, 8, 9, and 10 connected to bus 2, buses 11, 12 and 13 connected to bus 3, and buses 14 and 15 connected to bus 4. In this case also the reference bus i.e. bus 1 has voltage of 1.0000 pu and no any load connected to it. The total load of the system is 1489.2 kW and 922.96 kVAR. The line data and load data of the network is tabulated in Appendix B.

When a forward / backward sweep method of load flow is used for this system with base voltage and power of 11kV and 100 MVA respectively, the resulting voltage of each bus is presented in Figure 18 and active and reactive power losses are presented in Figure 19.

The maximum bus voltage is found 1.0000 pu of reference bus and minimum bus voltage is found 0.9456 pu of bus 13. There is an irregular variation of bus voltages with the voltage of bus 12, bus 13, and bus 15 have voltage 0.9469 pu, 0.9456 pu, and 0.9498 pu respectively and they are lower than specified voltage constraint’ lower limit 0.9500 pu (1 - 0.05pu).

illustration not visible in this excerpt

Figure 18: IEEE 15 bus radial distribution system bus voltages profile

illustration not visible in this excerpt

Figure 19: Active and Reactive power losses in IEEE 15 bus system

The RDS has maximum active and reactive power losses are found 37.5653 kW and 36.7436 kVAR respectively at bus 2 and minimum losses are found of 0.0437 kW and 0.0295 kVAR at bus 10. The bus 2 has many branches connected to it and therefore it is heavily loading in terms of current and thus have more power loss than other buses.

The total active and reactive power losses of the system are found 60.4562 kW and 56.3421 kVAR respectively. In this RDS, active power loss and reactive power loss has almost same pattern at buses.

4.3.2 Results of Optimal Allocation of DG unit in IEEE 15 Bus Test System

With objective function as minimum loss for the entire system and loss must be less than the base case active power loss of 60.4562 kW under the constraints that (i) bus voltage limit (1.0000 ± 0.05) pu, and (ii) active power supplied from DG unit must be less than or equal to 60 % of total real power load at power factor of 0.85 (DG size ≤ 893.52 kW, 553.78 kVAR). The optimal size of DG unit is found 890.54 kW and 530.26 kVAR located at bus 4 which results the minimum loss under provided constraints.

The bus voltage profile plot for base case (without DG unit) and improved case (With DG unit) is represented in Figure 20. With DG allocation at bus 4, the maximum bus voltage is found 1.0000 pu of reference bus and minimum bus voltage is found 0.9750 pu of bus 7. All buses voltages are improved and also the voltage of buses 12, 13, and 15 are improved. The candidate buses voltage with DG unit are improved and they are within the voltage regulation limit.

Active power losses of each bus before and after DG unit allocation are presented in Figure 21. Before DG allocation, the active power loss of bus 2 has maximum contribution in total active power loss of the system as seen from Figure 21. After DG unit ((DG size = 890.54 kW, 530.26 kVAR) allocation at bus 4, there is drastic reduction in real power loss at each buses. The active power loss at bus 2 reduces to 6.0017 kW and other buses 3, 6, and 10 have active power loss less than the base case i.e. before DG allocation. When DG unit is allocated at bus 4 then the total active and reactive power losses are 16.4193 kW and 13.3493 kVAR respectively represented in Figure 22 and corresponding losses are reduced by 72.82 % and 76.31 %. When DG unit of size 1036.45 kVA, at 0.85 pf, is located at different buses then the total system loss is represented in Figure 23. When DG is located at bus 13, the real power loss of the system is maximum and it is minimum when DG is located at bus 4. Individual candidate buses voltage for different location of DG units are represented in APPENDIX D.

illustration not visible in this excerpt

Figure 20: IEEE 15 bus system voltages profile before and after DG allocation

illustration not visible in this excerpt

Figure 21: Active power losses in each bus of IEEE 15 bus system without and with DG allocation

illustration not visible in this excerpt

Figure 22: Total active and reactive power losses of IEEE 15 bus system without and with DG allocation

4.4 DG integration in IEEE 33 Bus test System

4.4.1 Load Flow Analysis of IEEE 33 Bus Test System

The RDS configuration of IEEE 33 bus system presented in Figure 24 has branches sub-divided from bus 2, bus 3, and bus 6. The bus 2 has branch that includes four buses 19, 20 21, and 22, bus 3 has branch that includes three buses 23, 24, and 25, and bus 6 has branch which includes eight buses from bus 26 to bus 33 as shown in Figure 24. The total load of the system is 3715 kW and 2300 kVAR with maximum active load of 420 kW at bus 24 and bus 25 and maximum reactive load of 600 kVAR at bus 30, and minimum active load of 45 kW at bus 11 and minimum reactive load of 10 kVAR at bus 15. In this system also, the reference bus is bus 1 and has no any load with bus voltage 1.0000 pu. The line data and load data of the network is tabulated in Appendix B.

Load flow of the system is one with base voltage of 12.66 kV and base power of 100 MVA. The load flow analysis using forward / backward sweep algorithm is used which results the voltage, current, active and reactive power flow, and active and reactive power losses at each bus. The voltage profile, active and reactive power losses of each bus is presented in Figure 25 and Figure 26 respectively.

The maximum value of bus voltage is found 1.0000 pu of reference bus and minimum bus voltage is 0.9134 pu of bus 18 which is located at farthest end of the system. The bus voltage from bus 1 to bus 18 is going on decreasing order and then increases for bus 19, and after bus 19 the voltage is again decreases gradually because bus 18 is the farthest end of the RDS configuration. The voltage of buses (bus 6 to bus 18 and bus 26 to bus 33) lies in between 0.9134 pu to 0.9498 pu and are lower than specified voltage constraint’ lower limit. The bus voltage profile of IEEE - 33 bus system is presented in Figure 25. It has total 21 buses whose voltage does not satisfy the voltage regulation limit.

illustration not visible in this excerpt

Figure 25: IEEE 33 bus radial distribution system bus voltages profile

The maximum power losses in this configuration are found 51.5711 kW active power loss at bus 3 followed by bus 6 whose active power loss is 38.0256 kW. Real power loss at bus 3 is responsible for 25.54 % of the total real power loss of the system. The maximum reactive power losses and 32.8256 kVAR reactive power loss at bus 6 followed by 26.2668 kVAR at bus 3. Reactive power loss at bus 6 is responsible for 24.38 % of the total reactive power loss of the system. Minimum power losses are found 0.0132 kW and 0.0205 kVAR at bus 33. All the buses have real (or reactive power loss) below 5 kW (or 5 kVAR) except buses 2, 3, 4, 5, 6, 28 and 29 whose losses are greater. Bus 2 handles more current than other buses and thus the active power loss is greater for that bus and bus 6 is also heavily loaded and the reactive power loss at that bus is greater than other buses. The total active and reactive power losses of the system are found 201.8927 kW and 134.6416 kVAR respectively.

illustration not visible in this excerpt

Figure 26: Active and Reactive power losses in IEEE 33 bus system

4.4.2 Results of Optimal Allocation of DG unit in IEEE 33 Bus Test System

The paradigm described in section 3.2 is applied for IEEE - 33 bus distribution system with objective function as minimum loss for the entire network and loss must be less than the base case loss of 201.8927 kW and 134.6416 kVAR under the constraints that

(i) bus voltage limit (1.0000 ± 0.05) pu, and (ii) active power supplied by DG unit must be less than or equal to 60 % of total load at power factor of 0.85 (DG size ≤ 2229 kW, 1381.4 kVAR). The optimal size of DG unit is found 2188.95 kW and 1336.75 kVAR located at bus 26 which results the minimum active and reactive power loss of the system 64.8273 kW and 49.9337 kVAR respectively. The voltage profile plot of the system presented in Figure 27. All buses voltage is improved and lies within voltage regulation limit. When DG unit is allocated at bus 26, the maximum bus voltage is found 1.0000 pu of reference bus 1 and minimum bus voltage is found 0.9583 pu of bus 18.

illustration not visible in this excerpt

Figure 27: IEEE 33 bus system voltages profile before and after DG allocation

Active power losses of each bus before and after DG unit allocation are presented in Figure 28. Before DG allocation, the active power loss of bus 3 has maximum contribution followed by bus 6, bus 4, and bus 5 in total active power loss of the system as seen from Figure 28. After DG unit ((DG size = 2188.95 kW and 1336.75 kVAR) allocation at bus 26, there is drastic reduction in real power loss at each buses. The active power loss at bus 3 reduces to 5.7833 kW and other buses 4, 5, and 6 have active power loss less than the base case i.e. before DG allocation. When DG unit is allocated at bus 26 then the total active and reactive power losses are 64.8273 kW and 49.9337 kVAR respectively represented in Figure 29 and corresponding losses are reduced by 67.89 % and 62.91 %. When DG unit of size 2564.84 kVA, at 0.85 pf, is located at different buses then the total system loss is represented in Figure 30. When DG is located at bus 22, the real power loss of the system is maximum and it is minimum when DG is located at bus 26. Individual candidate buses voltage for different location of DG units are represented in APPENDIX D.

illustration not visible in this excerpt

Figure 28: Active power losses in each bus of IEEE 33 bus system without and with DG allocation

illustration not visible in this excerpt

Figure 29: Total active and reactive power losses of IEEE 33 bus system without and with DG allocation

illustration not visible in this excerpt

Figure 30: Total active power loss of network at different location of DG unit in IEEE - 33 bus system

4.5 Discussion

The DG unit size, location and corresponding real and reactive power loss and reduction with DG allocation is tabulated and represented in Table 3. The method proposed, in this dissertation, targets to utilize DG units for voltage improvement of candidate buses and loss minimization is applied for IEEE - 12, IEEE - 15, IEEE - 33 bus distribution system. In all IEEE buses distribution system, the results indicate that the DG unit allocation improve the voltage and also reduces the system losses. The system losses increase by improper placement of DG unit and voltage profile of the system gets reduced or overvoltage is seen as presented in APPENDIX D. Therefore, PSO method of optimization with sweep method of load flow analysis can be used to determine the size of DG unit and its best location for loss minimization. The DG units can be placed most sensitive bus to improve the bus voltages without violating DG unit size limit, the current and voltage limits.

Table 3: DG size and location with power loss for different IEEE buses

illustration not visible in this excerpt

illustration not visible in this excerpt

The voltage improvement of candidate buses in IEEE - 13, IEEE - 15, and IEEE - 33 bus systems are represented in Table 4. Initially the 3 buses of IEEE - 12 bus system 10, 11 and 12 have voltage lower than 0.9500 pu and those are improved. Similarly, candidate buses 12, 14, and 15 of IEEE - 15 bus voltage are improved and also for IEEE - 33 bus system there are 21 candidate buses 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 26, 27, 28, 29, 30, and 31 whose voltage are lower than 0.9500 pu and those candidate buses voltage are improved by DG allocation. The voltage without DG unit and with DG unit allocation are represented in Table 4.

CHAPTER 5 DISTRIBUTED GENERATION PLACEMENT IN NEPALESE BUS SYSTEM

5.1 General

The proposed paradigm is applied on IEEE - 12, IEEE - 15, and IEEE - 33 bus system described in chapter 4 and the results are found to satisfactory. The benchmark of study is completed with results of improved voltage and minimum power loss when DG is allocated at most sensitive candidate bus. The methodlogy is applied to the Bhaktapur - Balaju 21 bus distribution system to locate the DG and also size of DG to improve the voltage of the system with minimized power loss. In this chapter, the base case of the system is studied using forward / backward sweep load flow method and then size of DG and its location is determined with improved voltage and power losses in each case. The optimized size of DG is recommended to the most sensitive bus which provide minimum power loss under voltage limits.

5.2.1 Load Flow Analysis of Bhaktapur - Balaju 21 Bus System

The radial distribution system network of Bhaktapur - Balaju 21 bus system is presented in Figure 31 located at Bhaktapur - Balaju territory. The RDS is sub-divided into branches. The bus 2, bus 3, bus 6, bus 16, bus 17 and, bus 18 have branches connected single bus to each of them. The reference bus 1 has no any load connected to it and it acts as substation for the distribution system. The total load of the system is 865.13 kW and 121.41 kVAR with maximum load 145.6 kW and 29.6 kVAR at bus 11. It is assumed and also found that the voltage of reference bus is always 1.0000 pu with no any power losses into it. The line data and load data of the network is tabulated in Appendix B.

illustration not visible in this excerpt

Figure 31: Bhaktapur - Balaju 21 bus distribution system

The voltage profile and power (active and reactive) losses of the Bhaktapur - Balaju 21 bus system resulting from forward / backward sweep method of load flow is presented in Figure 32 and Figure 33 respectively. In load flow, the base value of voltage and power are 11 kV and 100 MVA respectively. The maximum and minimum voltage of the system is found 1.000 pu at reference bus i.e. bus 1 and 0.8640 pu at bus 13 respectively. Buses 3 to 21 have voltage lower than 0.9500 pu which is specified voltage constraint’ lower limit. In this system the load varies randomly from one bus to another and thus current also and therefore the voltage profile of the system has randomness. The buses which have under-voltage should to be improved such that all buses voltage lies within (1±0.05) pu.

The maximum power losses in this configuration are found 5.7354 kW active power loss at bus 3 and 4.9510 kVAR reactive power loss at bus 6. The total active and reactive power losses of the system are found 38.9215 kW and 28.0974 kVAR respectively. Losses at bus 6, 8, 9 and 10 have more contribution for the total system loss.

illustration not visible in this excerpt

Figure 32: Bhaktapur - Balaju 21bus radial distribution system bus voltages profile

illustration not visible in this excerpt

Figure 33: Active and Reactive power losses in Bhaktapur - Balaju 21 bus distribution system

5.2.2 Results of Optimal Allocation of DG unit in Bhaktapur - Balaju 21 Bus System

The paradigm described in section 3.2 is implemented for Bhaktapur - Balaju 21 bus distribution system with objective function as total loss of the system should be minimum and loss must be less than the base case active power loss of 38.9215 kW under the constraints provide that (i) candidate bus voltage limit (1.0000 ± 0.05) pu, and (ii) active power supplied by DG unit must be less than or equal to 60 % of total active load at a power factor of 0.85. The optimal size of DG unit is found 517.51 kW and 64.87 kVAR located at bus 14 which results the minimum loss under provided constraints, improved voltage of each buses, and minimized power losses. After the allocation of DG unit (DG size = 517.51 kW, 64.87 kVAR) at bus 14, the maximum bus voltage is found 1.0000 pu of reference bus and minimum bus voltage is found 0.9852 pu of bus 7 and bus 11. Voltage of bus 3 to bus 21 is improved and lies within the voltage regulation limit. The bus voltages profile of base case and improved bus voltages profile are presented in Figure 34.

Active power loss in each buses, and total active power and reactive power without DG unit allocation and with DG unit allocation is represented in Figure 35 and Figure 36 respectively. The active power losses at bus 3, bus 6, bus 8, bus 9, bus 10, and bus 13 are drastically reduced by installation of DG unit and active power loss of other buses are also reduced significantly. After allocation of DG unit at bus 14, the maximum active power loss and reactive power loss in the system are found 0.5868 kW and 0.5066 kVAR at bus 6. The total active and reactive power losses of the system with DG unit allocation are 3.6647 kW and 2.8400 kVAR respectively. The active power loss of the system is reduced by 90.58 % and reactive power loss of the system is reduced by 89.89 % with installation of DG which indicate significant reduction in the power losses. A DG unit of size 521.60 kVA, at power factor of 0.99, is located at different buses of Bhaktapur - Balaju 21 bus distribution system then the maximum real power loss is found 30.8688 kW when DG unit is located at bus 2 and minimum real power loss is found 3.6647 kW when DG unit is located at bus 14. The network total real power loss when DG unit is allocated at different buses is represented in Figure 37. Individual candidate buses voltage for different location of DG units are represented in APPENDIX C.

illustration not visible in this excerpt

Figure 34: : Bhaktapur - Balaju 21 bus system voltages profile before and after DG allocation

illustration not visible in this excerpt

Figure 35: Active power losses in each bus of Bhaktapur - Balaju 21 bus system without and with DG allocation

illustration not visible in this excerpt

Figure 36: Total active and reactive power losses of Bhaktapur - Balaju 21 bus system without and with DG allocation

illustration not visible in this excerpt

Figure 37: Total active power loss of network at different location of DG unit in Bhaktapur - Balaju 21

illustration not visible in this excerpt

5.3 Discussion

The proposed paradigm finds the optimal size of DG unit and its location in the distribution using PSO approach by simultaneous minimization of active power losses. The assize of DG found 517.51 kW and 64.87 kVAR located at bus 14 reduces the active power loss by 90.58 % and is presented in Table 5. The buses voltage are improved with DG allocation and results with DG unit and without DG unit voltage are represented in Table 6. Results represented in Table 6 indicate that the candidate buses 3 to 21 have lower voltage at initial and those voltages are improved by DG allocation.

Table 5: DG size and location with power loss for different Bhaktapur - Balaju 21 bus system

illustration not visible in this excerpt

Table 6: Voltage improvement of candidate buses in Bhaktapur -Balaju 21 bus system

illustration not visible in this excerpt

CHAPTER 6 CONCLUSION AND RECOMMENDATION

6.1 Conclusion

The DG unit integration in distribution system affect the operation of the distribution system and also improper placement of DG unit cause mal-operation of the system. DG units in distribution network is more beneficial when optimally placed and sized. In this thesis, therefore, an optimization algorithm of particle swarm optimization approach with the iterative technique of sweep approach for load flow is presented for the integration of DG unit to improve candidate bus voltage and minimize the total loss of the system. The proposed algorithm evaluates the losses and bus voltage of base condition. There is random generation of populations of solutions at first stage of particle swarm optimization method and then particle moves in search space and then find the best solution. It works with the objective that minimize the active power loss under a constraint of voltage regulation limit and DG size limit. It is tested for standard IEEE test distribution system and then applied to the Bhaktapur - Balaju 21 bus distribution system which is Nepalese distribution system. From the results, explained in chapter 4, for IEEE - 12, 15, and 33 bus distribution systems, it is clearly seen that voltage profile of the system is improvement and the active power loss of the system gets reduce by using the proposed model. When the proposed methodology is used for Bhaktapur - Balaju 21 bus distribution system then the results, explained in chapter 5, shows that voltage profile of the system is improved and also the total loss of system is reduced.

Finally, it is concluded that DG unit can locate optimally with optimized size into the radial distribution system for the voltage profile improvement of the buses and also for total system loss reduction.

6.2 Recommendations for Future Works

The possible future directions works based on the framework of this thesis are as follows:

- The algorithm can be made for time varying loads. The multiple objective functions can be used for optimization of DG unit.
- In conjunction with proposed method, the reconfiguration of distribution system and capacitor placement can be investigated.
- Loss repression index can be introduced in the investigation of DG unit placement.
- The role of smart grids can be examined with DG units in planning, reconfiguration and voltage control of the distribution system.
- Modelling of the load for planning and / or operational issues can be analyzed.
- Protection system for DG unit and distribution system can be studied.

REFERENCES

[1] W. Khan, “Typical AC power System,” 2016. [Online]. Available: http://www.electricaltechnology.org/2013/05/typical-ac-power-supply-system-scheme.html. [Accessed: 10-Oct-2016].

[2] P. Chiradeja and R. Ramakumar, “An Approach to Quantify the Technical Benefits of Distributed Generation,” IEEE Trans. Energy Convers., vol. 19, no. 4, pp. 764-773, 2004.

[3] A. Keane, “Integration of Distributed Generation,” University College Dublin, 2007.

[4] C. L. Masters, “Voltage rise: the big issue when connecting embedded generation to log 11 kV overhead lines,” POWER Eng. J., vol. 16, no. 1, pp. 512, 2002.

[5] N. Acharya, P. Mahat, and N. Mithulananthan, “An analytical approach for DG allocation in primary distribution network,” Int. J. Electr. Power Energy Syst., vol. 28, no. 10, pp. 669-678, 2006.

[6] E. Liu and J. Bebic, “Distribution System Voltage Performance Analysis for High-Penetration Photovoltaics Distribution System Voltage Performance Analysis for High-Penetration Photovoltaics,” Natl. Renew. Energy Lab., no. February, 2008.

[7] M. Jegadeesan and V. Keerthana, “Optimal Sizing and Placement of Distributed Generation in Radial Distribution Feeder Using Analytical Approach,” Int. J. Innov. Res. Sci. Eng. Technol. , pp. 1 -12, 2016.

[8] M. Vatani, G. B. Gharehpetian, M. J. Sanjari, and D. Solati Alkaran, “Multiple distributed generation units allocation in distribution network for loss reduction based on a combination of analytical and genetic algorithm methods,” IET Gener. Transm. Distrib., vol. 10, no. 1, pp. 66-72, 2016.

[9] M. F. Sohi, A. Javidaneh, and M. Shirdel, “Applying BCO Algorithm to Solve the Applying Algorithm to Solve the Optimal DG BCO Placement and Sizing Problem Optimal DG Placement and Sizing Problem,” in IEEE 5th international Power Engineering and optimization Conference, 2011, no. June, pp. 6-7.

[10] K. Vinothkumar and M. P. Selvan, “Fuzzy Embedded Genetic Algorithm Method for Distributed Generation Planning,” Electr. Power Components Syst., vol. 39, no. 4, pp. 346-366, 2011.

[11] F. J. Ruiz-Rodriguez, M. Gomez-Gonzalez, and F. Jurado, “Binary Particle Swarm Optimization for Optimization of Photovoltaic Generators in Radial Distribution Systems Using Probabilistic Load Flow,” Electr. Power Components Syst., vol. 39, no. 15, pp. 1667-1684, 2011.

[12] G. K. Ras and T. Bouktir, “Optimal Allocation and Sizing of Distributed Generation with Particle Swarm Optimization Algorithm for Loss Reduction,” Rev. des Sci. la Technol. - RST-, vol. 6, no. July, 2015.

[13] M. Gomez-gonzalez, A. Lopez, and F. Jurado, “Optimization of distributed generation systems using a new discrete PSO and OPF,” Electr. Power Syst. Res., vol. 84, no. 1, pp. 174-180, 2012.

[14] T. J. Sahib, Z. Jano, I. H. Mohamed, and U. Teknikal, “Optimum Allocation of Distributed Generation using PSO : IEEE Test Case Studies Evaluation,” Int. J. Appl. Eng. Res., vol. 12, no. 11, pp. 2900-2906, 2017.

[15] T. Ackermann, G. Andersson, and L. Soder, “Distributed generation: A definition,” Electr. Power Syst. Res., vol. 57, no. 3, pp. 195-204, 2001.

[16] R. S. Al-Abri, “Voltage stability analysis with high distributed generation penetration,” University of Waterloo, 2012.

[17] IEA ETSAP and IRENA, “Renewable Energy Integration in Power Grids,” 2015.

[18] F. S. Pai and S. J. Huang, “Design and Operation of Power Converter for Microturbine Powered Distributed Generator with Capacity Expansion Capability,” IEEE Trans. Energy Convers., vol. 23, no. 1, pp. 110-118, 2008.

[19] R. C. Flagan and J. H. Seinfeld, “Internal Combustion Engines,” in Fundamentals of air pollution engineering, Prentice-Hall, Inc., 1988, pp. 226289.

[20] B. R. Zogg et al., “Using Internal-Combustion Engines For Distributed Generation,” ASHRAE J., pp. 76-80, 2007.

[21] R. C. Dugan, M. F. McGranaghan, S. Santoso, and H. W. Beaty, Electrical Power Systems Quality, Second Edi. McGrawHill, Inc., 2002.

[22] A. Shrestha, A. Singh, K. Khanal, and R. K. Maskey, “Potentiality of Off-grid Hybrid Systems for Sustainable Power Supply at Kathmandu University Campus,” in IEEE 6th International Conference on Power Systems (ICPS), 2016, pp. 1-6.

[23] “Wind turbine paper,” 2012.

[24] K. Mahmoud, N. Yorino, and A. Ahmed, “Optimal Distributed Generation Allocation in Distribution Systems for Loss Minimization,” IEEE Trans. Power Syst., vol. 31, no. 2, pp. 960-969, 2016.

[25] A. Bin Humayd, “Distribution System Planning with Distributed Generation : Optimal versus Heuristic Approach,” University of Waterloo, 2011.

[26] T. Gonen, Electric Power Distribution System Engineering, Second Edi. NewYork: McGrawHill, Inc., 1986.

[27] H. L. Willis, Power Distribution Planning Reference Book, Second Edi. NewYork: Marcel Dekker, 2004.

[28] P. Kundur, Power System Stability And Control, First Edit. California: McGrawHill, Inc., 1993.

[29] J. J. Grainer and W. Stevenson, Power System Analysis. Singapore: McGrawHill, Inc., 1994.

[30] N. Hatziargyriou and M. Donnelly, “CIGRE technical brochure on modeling new forms of generation and storage,” Cigre, Tf, no. November, 2000.

[31] P. Kundur et al., “Definition and Classification of Power System Stability IEEE/CIGRE Joint Task Force on Stability Terms and Definitions,” IEEE Trans. Power Syst., vol. 19, no. 3, pp. 1387-1401, 2004.

[32] H. Chiang and R. J.- Jumeau, “Toward a Practical Performance Index for Predicting Voltage Collapse in Electric Power Systems,” vol. 10, no. 2, pp. 584592, 1995.

[33] S. Kumar, “Optimal Sizing and Placement of Generator Units in Distributed Generation System by Particle Swarm Optimization (PSO) Techniques,” 2012.

[34] P. P. Barker and R. W. de Mello, “Determining the Impact of Distributed Generation on Power Systems: Part 1 - Radial Distribution Systems,” Power Eng. Soc. Summer Meet. 2000. IEEE, vol. 3, no. c, pp. 1645-1656 vol. 3, 2000.

[35] W. H. Kersting, Distribution System Modelling and Analysis. NewYork: CRC

Press, 2013.

[36] J. A. M. Rupa and S. Ganesh, “Power Flow Analysis for Radial Distribution System Using Backward / Forward Sweep Method,” vol. 8, no. 10, pp. 16211625, 2014.

[37] F. Teng, “Implementation of a Voltage Sweep Power Flow Method and Comparison with Other Power Flow Techniques,” Swiss Federal Institute of Technology (ETH) Zurich, 2014.

[38] R. D. Zimmerman, “Comprehensive Distribution Power Flow: Modeling, Formulation, Solution Algorithms and Analysis,” Cornell University, 1995.

[39] S. Luo, W. Cui, and J. Yang, “Distribution Network Optimization Reconfiguration Based on Hybrid Algorithm,” 2011 Int. Conf. Electr. Control Eng., pp. 2194-2197, 2011.

[40] A. Keane et al., “State-of-the-Art Techniques and Challenges Ahead for Distributed Generation Planning and Optimization,” IEEE Trans. Power Syst., vol. 28, no. 2, pp. 1 -1, 2012.

[41] C. Wang and M. H. Nehrir, “Analytical Approaches for Optimal Placement of Distributed Generation Sources in Power Systems,” IEEE Trans. Power Syst., vol. 19, no. 4, pp. 2068-2076, 2004.

[42] L. F. Ochoa, A. Padilha-Feltrin, and G. P. Harrison, “Evaluating Distributed Time-Varying Generation Through a Multiobjective Index,” IEEE Trans. Power Deliv, vol. 23, no. 2, pp. 1132-1138, 2008.

[43] Q. Zhou and J. W. Bialek, “Generation Curtailment to Manage Voltage Constraints in Distribution Networks,” Gener. Transm. Distrib. IET, vol. 1, no. 2, pp. 492-498, 2007.

[44] C. J. Dent, L. F. Ochoa, G. P. Harrison, and J. W. Bialek, “Efficient secure AC OPF for Network Generation Capacity Assessment,” IEEE Trans. Power Syst., vol. 25, no. 1, pp. 575-583, 2010.

[45] A. Hauswirth, T. Summers, J. Warrington, J. Lygeros, A. Kettner, and A. Brenzikofer, “A Modular AC Optimal Power Flow Implementation for Distribution Grid Planning,” in IEEE Eindhoven PowerTech, 2015.

[46] G. P. Harrison, A. Piccolo, P. Siano, and A. R. Wallace, “Hybrid GA and OPF evaluation of network capacity for distributed generation connections,” vol. 78, pp. 392-398, 2008.

[47] E.-G. Talbi, A. Nakib, and A. Fusar, “Hybrid Metaheuristic for Annual Hydropower Generation Optimization,” in IEEE 28th International Parallel & Distributed Processing Symposium Workshops, 2014, pp. 412-419.

[48] L. Lessing, I. Dumitrescu, and T. Stutzle, “A Comparison Between ACO Algorithms for the Set Covering Problem,” in 4th International Workshop, 2004, pp. 1-24.

[49] F. S. Abu-mouti and M. E. El-Hawary, “and Sizing in Distribution Systems via Artificial Bee Colony Algorithm,” IEEE Trans. Power Deliv., vol. 26, no. 4, pp. 2090-2101, 2011.

[50] R. C. Eberhart and Yuhui Shi, “Particle swarm optimization: developments, applications and resources,” Proc. 2001 Congr. Evol. Comput. (IEEE Cat. No.01TH8546), vol. 1, pp. 81-86, 2001.

[51] W. S. Andrade, C. L. T. Borges, and D. M. Falcao, “Modeling reliability aspects of distributed generation connected to distribution systems,” 2006 IEEE Power Eng. Soc. Gen. Meet., p. 6 pp., 2006.

[52] V. F. Martins and C. L. T. Borges, “Active distribution network integrated planning incorporating distributed generation and load response uncertainties,” IEEE Trans. Power Syst., vol. 26, no. 4, pp. 2164-2172, 2011.

[53] R. C. Green, L. Wang, M. Alam, C. Singh, and S. S. S. R. Depuru, “An examination of artificial immune system optimization in intelligent state space pruning for LOLP estimation,” NAPS 2011 - 43rdNorth Am. Power Symp., 2011.

[54] “Electricity Rules , 2050 ( 1993 ),” Nepal.

[55] IEEE Recommended Practice for Electric Power Distribution for Industrial Plants. IEEE Power System Engineering Committee, 1993.

[56] “MATLAB Designed For Engineers and Scientists,” 2016. [Online]. Available: https://www.mathworks.com/products/matlab/why-

matlab.html?s_tid=hp_brand_whymatlab#reason5. [Accessed: 14-Oct-2016].

[57] S. S. Rao, Engineering Optimization: Theory and Practice. 2009.

[58] S. Mishra, D. Das, and S. Paul, “A Simple Algorithm for Distribution System Load Flow with Distributed Generation,” in IEEE International Conference o Recent Advances and Innovations in Engineering, 2014.

[59] G. V. N. Lakshmi, A. J. Laxmi, and V. V. Reddy, “Optimal Allocation and Sizing of Multiple Distributed Generators using Genetic Algorithm,” in International Conference on Advances in Communication, Network, and Computing, 2014.

[60] A. Wazir and N. Arbab, “Analysis and Optimization of IEEE 33 Bus Radial Distributed System Using Optimization Algorithm,” J. Emerg. Trands Appl. Eng., vol. 1, no. 2, 2016.

[61] D. Bhujel, B. Adhikary, and A. K. Mishra, “A Load Flow Algorithm for Radial Distribution System with Distributed Generation,” in IEEE International Conference on Science Engineering, and Technology, 2012.

APPENDIX A

PARTICLE SWARM OPTIMIZATION

Particle Swarm Optimization based on behaviour of social organisms such as colony or swarms of insects, such as ants, termites, bees, and wasps, a flock of birds, or a school of fish. A particle denotes a bee in a colony or a bird in a flock. Each individual or particle in swarm using its own intelligence and the collective or group intelligence of the swarm [57].

In multi-variable optimization, each particle is located randomly in multi-dimension design space. Each particle has its own position and velocity in designed space and remembers the best position (based on objective function) that has been discovered.

Computational Implementation of PSO in Distribution System

Consider a minimization objective function: Minimize f(X) [illustration not visible in this excerpt]

With limit of DG unit size DG [illustration not visible in this excerpt]

[illustration not visible in this excerpt] are lower and upper bounds on X respectively. The PSO procedure is implemented through following steps:

Generate the random population of X in a range of [illustration not visible in this excerpt] of size of swarm (N). Set initial particles Xi(0), X2(0), X3(0),., Xn(0) and perform load flow to evaluate objective function for each particle.

All particles will be moving to the optimal point with a velocity and, Initially, all particle velocities are assumed to be zero.

For [illustration not visible in this excerpt]particle, evaluate following parameters.

a. Historical best value of Xj(i) is the best position Pbest, with lowest value of objective function, encountered by particle j in all the previous iterations.

b. Gbest, with the lowest value of the objective function [illustration not visible in this excerpt], encountered in all the previous iterations by any of the N particles.

c. Determine the value of velocity for particle j in [illustration not visible in this excerpt] iteration and then update the position of [illustration not visible in this excerpt] particle using following equations:

illustration not visible in this excerpt

Where c1 and c2 are individual and social learning coefficients.

d. Check the DG size constraints and Voltage constraints. If not meet the constraint, then generate swarm and goto step 2.

e. Evaluate objective function for a new particle [illustration not visible in this excerpt]

4. Repeat step 4 until maximum iteration reached and convergence criteria not satisfied.

APPENDIX B

IEEE - 12 BUS SYSTEM DATA [58]

Table 7: Bus and line data of IEEE 12 bus system

illustration not visible in this excerpt

IEEE - 15 BUS SYSTEM DATA [59]

Table 8: Bus and line data of IEEE 15 bus system

illustration not visible in this excerpt

IEEE - 33 BUS SYSTEM DATA [60]

Table 9: Bus and line data of IEEE 33 bus system

illustration not visible in this excerpt

BHAKTAPUR - BALAJU 21 BUS SYSTEM DATA [61]

Table 10: Bus and line data of Bhaktapur - Balaju 21 bus system

illustration not visible in this excerpt

APPENDIX C

RESULTS OF IEEE - 12 BUS SYSTEM

Table 11: Result of optimization for IEEE - 12 bus system

illustration not visible in this excerpt

RESULTS OF IEEE - 15 BUS SYSTEM

Table 12: Result of optimization for IEEE - 15 bus system

illustration not visible in this excerpt

RESULTS OF IEEE - 33 BUS SYSTEM

Table 13: Result of optimization for IEEE - 33 bus system for DG unit limitation

illustration not visible in this excerpt

BHAKTAPUR - BALAJU 21 BUS SYSTEM

Table 14: Result of optimization for Bhaktapur - Balaju 21 bus system for DG unit limitation

illustration not visible in this excerpt

APPENDIX D

CANDIDATE BUS VOLTAGE FOR IEEE - 12 BUS SYSTEM

(When 306.01 kVA of DG unit is located at different buses of IEEE – 12 bus system)

REAL POWER LOSS PROFILE FOR IEEE – 12 BUS SYSTEM

(When 306.01 kVA of DG unit is located at different buses of IEEE – 12 bus system)

illustration not visible in this excerpt

CANDIDATE BUS VOLTAGE FOR IEEE - 15 BUS SYTEM

(When 1036.45 kVA of DG unit is located at different buses of IEEE – 15 bus system)

illustration not visible in this excerpt

REAL POWER LOSS PROFILE FOR IEEE – 15 BUS SYSTEM

(When 1036.45 kVA of DG unit is located at different buses of IEEE – 15 bus system)

illustration not visible in this excerpt

CANDIDATE BUS VOLTAGE FOR IEEE – 33 BUS SYTEM

(When 2564.84 kVA of DG unit is located at different buses of IEEE – 33 bus system)

illustration not visible in this excerpt

REAL POWER LOSS PROFILE FOR IEEE – 33 BUS SYSTEM

(When 2564.84 kVA of DG unit is located at different buses of IEEE – 33 bus system)

illustration not visible in this excerpt

CANDIDATE BUS VOLTAGE FOR BHAKTAPUR – BALAJU 21 BUS SYTEM

(When 521.60 kVA of DG unit is located at different buses of Bhaktapur – Balaju 21 bus system)

illustration not visible in this excerpt

REAL POWER LOSS PROFILE FOR BHAKTAPUR - BALAJU 21 BUS DISTRIBUTION SYSTEM

(When 521.60 kVA of DG unit is located at different buses of Bhaktapur – Balaju 21 bus system)

illustration not visible in this excerpt

132 of 132 pages

Details

Title
Particle Swarm Optimization Approach for Distributed Generation Planning to Improve the Voltage Profile Margin
College
Kathmandu University  (School of Engineering)
Course
Planning and Operation of Energy System
Grade
4.00
Author
Year
2017
Pages
132
Catalog Number
V376132
ISBN (Book)
9783668551015
File size
3031 KB
Language
English
Tags
particle, swarm, optimization, approach, distributed, generation, planning, improve, voltage, profile, margin
Quote paper
Ajay Singh (Author), 2017, Particle Swarm Optimization Approach for Distributed Generation Planning to Improve the Voltage Profile Margin, Munich, GRIN Verlag, https://www.grin.com/document/376132

Comments

  • No comments yet.
Read the ebook
Title: Particle Swarm Optimization Approach for Distributed Generation Planning to Improve the Voltage Profile Margin


Upload papers

Your term paper / thesis:

- Publication as eBook and book
- High royalties for the sales
- Completely free - with ISBN
- It only takes five minutes
- Every paper finds readers

Publish now - it's free