A methodology for optimal placement of distributed generation on meshed networks to reduce power losses for time variant loads.

PLAGIARISM

PUBLICATION

World energy outlook

Predictions from the '450' scenario for a reduction in greenhouse gases in 2035 require a growth in renewable energy sources from 13% to 27%. The share of energy sources translates from the scenario forecast to the technology mix for each scenario, as shown in Figure 1-2.

International demand for renewable energy

South African demand for renewable energy

Capacity and sizing alternatives
Location options
South Africa’s investment in renewable energy
Renewable power plant network assessment challenges

The DOE's acceleration of REI4P is commitment to start up renewable power plants (RPPs) to support the state-owned utility. The accelerated connection plan and technical impact of RE capacity on the Eskom power system led to the definition of the following research questions 1.5.

Table 1-2 – Capacity Options for Technology Mix over IRP Horizon [5]

Research hypothesis

Long-term equipment size and power vary for single-entity connections compared to multiple connections.

Research questions

Objectives and approach

Significance of research

Structure of dissertation

Chapter in perspective

Optimal placement techniques

This chapter aims to search scholarly works in the area of ideal placement of DG.

What critical electrical factors determine the optimal placement of DG to reduce power

In [17], a network unbundling algorithm was presented that enables radial protection in the presence of DG. While DG placement improves the network voltage profile, optimal placement and sizing play an important role in reducing system losses [29].

What is the importance of reactive power control for the integration DG ?

The layout and dimensioning of the DG unit was designed by mixed integer nonlinear programming, the aim of which was to improve the stability limit. This index method was implemented using a 69-node IEEE radial distribution network to determine the optimal layout and size of DG units.

How does DG penetration and concentration affect power system losses?

A method to detect and determine DG units to improve the voltage stability in the presence of probabilistic load and renewable DG generation was presented in [50]. The characterization of these sources and sinks by network impedance supports the research questions leading to the hypothesis.

Characteristics of time variant loads

Load types
Time variant loads

Utility planners must account for likely permutations of generation and load patterns when sizing and locating DG. High generation – low load; Low generation – high load and low generation – low load scenarios represent the critical study points for generation and load profiles.

Characteristics of time variant sources of renewable generation

Global wind patterns
Weibull distribution
Wind power equation
Wind power curve

Electrical characteristics of renewable wind generation

Real power
Reactive power

Solar PV - partially predictable source of renewable energy

Global solar irradiance
Solar power curve

In reality, the power output is affected by cloud disturbances, as shown in Figure 3-3 below. Depending on the time of day and cloud cover, the PV output varies as shown in Figure 3-3.

Figure 3-2 Direct Solar Radiation (KW/m2) – Southern Hemisphere [71]

Electrical characteristics of renewable solar PV generation

Real power
Reactive power

Concentrated solar plants

Bagasse

Hydro / pumped storage

Electrical characteristics of renewable non-solar PV and non-wind generation

Real power
Reactive power

RPCC is limited by the excitation current in the capacitive scale; from the stator currents to the inductive step. Unlike the PV inverter, the synchronous generator is a relatively clean source of electricity as very few harmonics are injected into the grid.

Dispatchable Vs non-dispatchable generation

Characteristics of power losses

Energy losses
Power losses

Characteristics of distribution network topology

Typology of distribution networks

Radial networks
Tie line radial networks
Meshed networks

Voltage profile of distribution networks
Thermal capability of distribution networks
Fault level characteristic of radial networks

For increased network reliability, radial network designs include switchable normally open link lines to support interconnection. They are placed at the beginning of the network, connected to the bypass circuit breaker for maintenance purposes. Another option is in the middle of a traditional trunked network; or at the end of the feeder.

The voltage profiles of radial and tie-line radial feeders are tapered from source to end of the line. Regulatory codes may limit the use of DG to resolve existing voltage problems as it may be too expensive for this purpose alone [87]. Strategic and optimal placement and sizing of DG on radial and tie-line radial networks can increase thermal transfer on the network.

Impact of distributed generation on distribution network topology

Department of energy’s renewable energy requirement

Eskom transmission connection strategy

Where concentrated RE sources are to be evacuated through EHV networks, the following connection strategy is considered. Passive voltage control through reactive power is limited by the impedance of the sub-transmission lines between the satellite, solar collector and MTS. Integration, shown in Figure 3-10, favors loss reduction as DG is consumed at the load center.

Connections for network expansion, while managing voltage control; reactive power control and loss minimization, should be robust.

Figure 3-10: Transmission Connection Strategy – Option I [95]

Typical Distribution connection options

Utility technical challenges

Critical electrical factors that determine optimal location to reduce power loss
Practical power flow simulation methodology applied for placement of DG for loss
The importance of reactive power support for integration of distributed generation 31

They show high reactive power losses when operating at high thermal demands, due to I2X losses. A lower Thevenin equivalent source impedance provides higher error levels and allows connections with higher DG capacity. Error rates can be too high and reduce the size of the connection. Due to the above problems, the integration and strengthening of the network will increase the costs.

This research aims to develop a practical approach for optimal placement and sizing of DG with the aim of reducing power losses. Sizing of DG at potential reverse flow locations should ideally be moved at higher load factors [100]. Incorporating dispatchable DG technologies can reduce reverse current flows, especially on older protection technology networks.

Table 3-2: Extra High-Voltage (EHV) Point of Connection (POC)

Chapter in perspective

Modeling these parameters to analyze optimal placement of DG are important factors in the assessment process.

Objective of method

Problem statement

Technical constraints

Scope of method

Assumptions

Exclusions

Optimal placement methodology

Overview
Pre-DG evaluation
Defining power system busbars
Ranking power system busbars
Profiling power system busbars for DG sizing
Activating DG for placement
Evaluate DG sizing for voltage variation
Optimal selection process

Pre-DG evaluates and analyzes a convergent DIgSILENT Power Factory © (DPF) version 15.1.4 utility network file. The pre-DG survey took place before August 2015 and the post-DG survey took place in September and October 2015. Optimal selection of candidate rails based on a comparative overview of optimal placement as described in section 4.7.8.

The method of defining identifies load and non-load (network) busbars suitable for DG connection. Profiling load and network rails to DG profiles classifies the expected impact of load and loss reduction. Voltage variation tests performed on load and network busbars confirm the grid code requirement for generation loss.

Figure 4-1: Optimum placement method 4.7.2 Pre-DG evaluation

Analysis of results

Chapter in perspective

The utility test network shows the interconnection of a typical network supplying industrial, commercial and rural load.

Transmission grid

Sub-Transmission

Grid I and Grid II
Line Parameters

The electrical network was simulated with the conductor impedances and conductor lengths shown in Table 8-1 and Table 8-2 in 8.3.1.

System overview prior to distributed generation

Load types and profiles
Sub – transmission grid loading
System fault level

High-voltage fault level
Medium voltage fault level

System voltage level

High-voltage busbar system per unit values
Medium voltage busbar system per unit values

Active power load flow
Reactive power load flow
Sub-transmission phase angles – power factor
Sub-transmission grid loss

The absence of reactive power from a solar PV plant implies other means of reactive compensation. Utilities can capitalize on inherent reactive power during these cycles so as not to negatively impact peak output capacity. This is an important finding that seeks to optimize the reactive power flow through solar DG without additional reactive power compensation.

Reactive power losses recorded the difference in flow between "from" and "to" busbars, shown in Table 8-10 and Table 8-11. Examination of the pre-DG phase angle of the 11kV ground bus, obtained from Table 8-3 row 17, shows the power factor of 0.965. Active power is reduced from the DG's supply; reactive power remains unchanged, resulting in a negative impact on the upstream power factor.

Figure 5-1 Daily Load Profiles 5.3.2 Sub – transmission grid loading

Distributed generation profile characteristic

Subtracting the DG profile from the load profiles gives the resultant "duck" shaped profile shown in Figure 5-3. The negative real power represents the excess DG that will flow upstream, from the connection point, to feed the adjacent load centers. The reverse power flow will cause losses if the mismatch between the load profile and the DG profile is large.

Voltage variation results

Voltage variation results – Earth 11kV busbar
Sub-transmission load flow – post DG 11kV Earth bus
Sub-transmission grid loss – post DG 11kV earth

The pre- and post-DG comparison at the 66 kV Earth bus shows the deepening of the midday profile as the DG injection increases. Earth 11kV BB: Output Current, Active Power in MW Earth 11kV BB: Output Power, Reactive Power in Mvar. Earth 66kV BB: Output Current, Active Power in MW Earth 66kV BB: Output Power, Reactive Power in Mvar.

The 11 kV ground collector hosted a relatively small DG solar plant due to its relatively low fault rate. The introduction of 8 MWs of DG reduced network losses by 9.5% for real power and 13% for reactive power. The analysis showed different levels of network losses depending on the load profile; type of load; error rate.

Figure 5-4 11kV Earth Busbar Voltage Variation Result

Summary of post-DG on utility network

Utility grid – post-DG

Two busbars in this voltage range do not qualify for selection, namely Ruby and LinkedIn. A review of the 22kV results by DG suggests all negative real and reactive losses in the network for selection. Larger dimensioning of DGs showed higher network losses when they were not placed close to the load busbars.

Increased power losses with larger DG sizes are notable for 88 kV rail voltages, shown in Table 5-7. The reduction in actual power losses is attributed to the 88kV Basket bus being closer to the load center. A review of the post-DG table for 132 kV rails does not nominate any rail for selection.

Table 5-3: 11kV Busbar - Post-DG Results Bus Name

Overview

Summary of findings

Assessing the research questions

What critical electrical factors determine the optimal placement of DG to reduce

Network topology
System impedance and system fault level
Equipment rating
Voltage control of active and reactive power flow
Load types and characteristics
Solar PV characteristics

What practical loss optimization power flow method can be applied to place DG on
What is the importance of reactive power control for the integration DG?
How does DG penetration and concentration affect power system losses?

Inadequate sizing of transformer and line capacities will overload, resulting in increased real and reactive power losses. As Q decreases, the voltage decreases as the square, referring to causes of voltage drift under low reactive power reserves. Voltage control through transformer tap-changers and reactive power compensation controls constant voltage within prescribed limits.

Reactive power compensation is essential for the integration of DG to address the reduction of reactive power flows. Comparisons were made for each pre- and post-connected DG validating the changes in phase angle; power factor and resulting reactive power current. As the penetration of DG increases to reduce active power, larger reactive power compensation units must be added to the network.

Relevance of research

Penetration and concentration of DG was explained in chapter 2 and demonstrated chapter 5 section and 5.6.7 shows the results for DG integration with 66kV busbars.

Assessing the hypothesis

Future research

Conclusion

Recommendation

Harrison, “Improved Resource Utilization for Voltage Regulation with Distributed Generation,” IEEE Transactions on Power Systems, vol. Atwa, “Optimal placement and sizing method to improve voltage stability margin in a distribution system using distributed generation,” IEEE Transactions on Power Systems, vol. Román, “Assessing Power Distribution Losses for Increasing Penetration of Distributed Generation,” IEEE Transactions on Power Systems, vol.

Reactive Power Compensation," in Electric Power Generation, Transmission, and Distribution, Third Edition, 5 vols., CRC Press, 2012, p. Lee Willis, "Power Delivery Systems," in Power Distribution Planning Reference Book, Second Edition, 0 vols. ., CRC Press, 2004. Khadkikar, “Planning of Active Distribution Networks Considering Multi-General Director Configurations,” IEEE Transactions on Power Systems, vol.

Transmission grid

Sub-transmission grid I

Sub-transmission grid II

Line parameters

Pre-DG system overview

Sub – transmission grid loading
High voltage fault level
Medium voltage fault level
High voltage busbar system per unit values
Medium voltage busbar system per unit values
High voltage active power load flow
Medium voltage active power load flow
High voltage reactive power load flow
Medium voltage reactive power load flow
Phase angle results for pre-DG
Sub-Transmission Grid Loss

Post DG assessment

Diamond 132kV busbar assessment for DG

Sub-transmission load flow – post DG 132kV diamond bus
Voltage variation results – diamond 132kV busbar
Sub-transmission grid loss – post DG 132kV diamond

Basket 33kV busbar assessment for DG

Sub-transmission load flow – post DG 33kV Basket bus
Voltage variation results – 33kV Basket busbar
Sub-Transmission Grid Loss – Post DG 33kV Basket

Phase angle results for post-DG connection to 11kV Earth busbar

The maximum export capacity of 98 MW is injected into the 132 kV diamond busbar and changes the load profile shown in Figure 8-7. The following set of graphs shows the impact of injecting 89 MW of real power into the Diamond 132 kV bus in 10% increments (X-axis). The 132kV diamond voltage variation test by Method 1 reports a pre-DG voltage of 1.035 p.u and a post-DG voltage of 1.031 p.u.

The DG dimensioning for the 132 kV diamond rail shows an increase in grid losses of 98% for real power and 61% for reactive power. Comparison shows the reduction in real power at the peak of solar PV generation. Solar PV injected into the 33kV Basket bus will have the effect of reducing the real power requirement of the 88V Basket busbar.