Harmonic Load Modeling: A Case Study of 33 KV Abuja Steel Mill Feeder
Harmonic Load Modeling: A Case Study of 33 KV Abuja Steel Mill Feeder
Harmonic Load Modeling: A Case Study of 33 KV Abuja Steel Mill Feeder
ABSTRACT
An in-depth study of the harmonic orders inherent in a power system network is required before mitigation
techniques are adopted. This paper studied the harmonic orders of the 33 kV Abuja Steel Feeder modeled as a
harmonic source using measured data. Readings of kW, kVar, kV and Hz were obtained using power quality analyzer
(PQube) to identify the trend and the harmonic unbalance on the feeder. Electrical Transient Analyzer Program
(ETAP) software package was deployed to perform Discrete Fast Transform (DFT) while the input parameters were
the resultant relative amplitudes and phase angles for both the current and voltage source models. The current source
model spectrum of the feeder under study revealed that the 11th and 13thharmonic orders have the highest percentage
of amplitude relative to the fundamental compared to the other harmonics. On the other hand, the voltage source
model spectrum showed that the 14th, 15th, 16th, 17th, 32nd, 33rd and 34thharmonic orders have higher percentage of
relative amplitude. However, only the 3rd and 5thharmonic orders were found to cause severe harmonic distortion of
voltage and current after harmonic analysis and frequency scan were performed on the ETAP platform.
Keywords: Harmonic order; Power quality; Transmission grid; ETAP; Tuned filters
network to address other power quality challenges, no the IEEE-12-pulse type 2. Similarly, Ali [10]investigated
harmonic filter or Flexible Alternating Current the propagation of harmonic current in a power
Transmission System (FACTS) devices for the purpose network using hypothetic model that injects harmonic
of addressing power quality and harmonic distortion current of 3rd, 5th, 7th and 9th orders developed on
on the power network has been implemented or Alternative Transient Program (ATP-Draw) software. A
installed as the Transmission Company of Nigeria major drawback of these idealized models is that they
network asset data revealed. This work is particularly do not give a true reflection of the network’s harmonics
significant to Nigeria power system operators where penetration and propagation on real time. Attempts
only voltage and frequency are being properly were made in the works of [11]-[13] to utilize field
monitored and controlled. measured data to model harmonic generating loads.
Harmonic load modeling of some Nigerian steel plants Various simulation software were used including ETAP
feeders using idealized models have been simulated as in Hagh et al. [11] wherein on-site harmonic current
and documented in research and development articles measurements were used to tune the computer based
[4][5]. However, modeling of these loads using harmonic current models to match and validate the
measured data obtained from power quality analysers measured data. Power System Computer Aided
on the Nigerian grid network appears very un-common. Design/Electromagnetic Transients including DC
Therefore, this study attempts to blaze the trail on this (PSCAD/EMTDC) was deployed in Venkatesh et al. [12]
all important power quality issue in the Nigerian grid to develop simulation models based on laboratory
network with 33kV Abuja Steel feeder as case study. measured waveforms. A generalized model for
distribution systems was successfully implemented for
2. REVIEW OF HARMONIC MODELING IN POWER practical case studies using the verified simulation
SYSTEM NETWORK models for various industrial and household electrical
Harmonic modeling is a mathematical way of predicting equipment and appliances. Similarly, DigSilent/Power
harmonic distortion levels and potential resonance on factory is the simulation software used in Ma et al.[13]
an electrical network based on available system data to evaluate the future levels of harmonic distortion in a
[6]. The steps required for making harmonics study power grid based on the present harmonic current
based on industrial research and experience involved source model developed from measured data.
obtaining the single line diagrams of the network, The second group of literature review proposed the use
highlighting the point of common coupling, of equivalent circuit models to represent characteristic
identification of harmonic sources and/or carrying out harmonic producing loads [14]-[16]. The Norton/
harmonic measurements, modeling the network Thevenin equivalent circuit provides a reliable model
components using a simulation software like ETAP and for harmonic current and voltage sources in a
comparing results obtained to national standards [7]. distribution network. The contribution of the power
Power quality meters may be installed on a number of utility equipment and customer loads to total harmonic
buses having large non-linear load in order to distortion observed at the point of common coupling
determine the level of unbalance. The locations of these (PCC) can separately be evaluated using field data and
measurements should be the Point of Common Coupling analytical techniques [14]. The major challenge of the
(PCC) which is the customer-utility interface. A Thevenin/Norton equivalent circuit models is the
minimum measurement period of one week is normally determination of the impedance and other parameters
recommended as it provides a representative loading of the power network. Evolutionary algorithm have
cycle for most industrial and commercial loads [8].To been tested and verified in the estimation of Thevenin/
obtain reliable harmonic model of any network, all Norton equivalent circuit parameters without knowing
components of the network must be precisely modeled. the network impedance [15]. An equivalent circuit that
In order to carry out a comprehensive harmonic considered the modeling of the harmonic current
modeling and simulation of the Abuja Steel feeder, two emission on the AC side of a diode bridge rectifier while
categories of literatures were identified and reviewed evaluating the harmonic voltage contribution of the DC
so as to establish the gaps and come up with a more side of the rectifier was successfully developed [16]. The
reliable result. The first group centres on harmonics drawback of modeling only the harmonic current was
modeling using traditional method otherwise called the successfully addressed for a more accurate harmonic
ideal current source models [9]–[16]. Khaled, et al. [9] analysis.
and Ali [10] deployed the use of idealized models. An In this research work, we aimed at modeling the current
important reference is Khaled, et al. [9] wherein and voltage harmonic content of the 33 kV Abuja Steel
harmonic model of the load under study was selected Feeder using ETAP software. A two week data was
from the ETAP library as a harmonic current source of obtained from power quality meter installed on the
Nigerian Journal of Technology, Vol. 36, No. 4, October 2017 1185
HARMONIC LOAD MODELING: A CASE STUDY OF 33 KV ABUJA STEEL MILL FEEDER M. N. Nwohu, et al
feeder at the point of common coupling (PCC). The The Abuja Steel feeder under case study is a 33 kV
ETAP Harmonic Analysis module was utilised to obtain feeder supplied by the 132/33 kV Suleja transmission
the harmonic load model of the Abuja Steel Feeder for substation, Niger State, Nigeria. It feeds the Abuja Steel
both harmonic current and voltage sources respectively. plant located about 15km away from the substation
The Thevenin/Norton Equivalent model was selected through a 33 kV distribution line. The plant is composed
for this research. This model involves the of two induction furnaces that produce high quality
representation of both the utility and customer sides steel from scrap metal with a capacity of 15 Tons each
with Thevenin/Norton equivalent circuits. Results and rated 33/1 kV.
obtained served as bases for further harmonic analysis Power Quality Analyser (PQube) produced by Power
and mitigation on the load under study. Its effect on Standard Lab was installed on the secondary side of
other power network components and loads were 45MVA transformer TR2 at Suleja 132/33 kV
revealed during harmonic load flow and frequency scan. transmission substation. This transformer supplies the
33 kV Abuja Steel Feeder. Two weeks data was
3. SULEJA POWER NETWORK DESCRIPTIONANDABUJA retrieved from the PQube’s Secure Digital (SD) memory
STEEL FEEDER SIMULATION card for the purpose of this study. The two weeks
The Suleja power network is composed of 132/33 kV spanned between 1st and 16th January, 2016. The data
transmission sub-station located in Niger State, Nigeria. was further analysed using Discrete Fourier Transform
The station is mesh connected to the National grid (DFT) analysis capability on the ETAP software’s
through four numbers of high voltage circuits namely harmonic analysis module to obtain the spectrum and
Abuja-Suleja 132 kV Line I, Abuja-Suleja 132 kV Line II, waveform of the harmonic model for the Abuja steel
Minna-Suleja 132 kV Line I and Minna-Suleja 132 kV feeder.
Line II. While the Abuja circuits emanate from the
Ketampe 330/132 kV transmission station in Abuja, 4. METHODOLOGY
Minna circuits emanate from Minna 132 kV Data collected from the harmonic measurement of the
transmission substation(primary supply source for both 33 kV Abuja Steel plant was carefully studied. The load
stations is ShiroroHydo-electric Power station in Niger variation within the two weeks period was analyzed
state). The Suleja 132/33 kV sub-station has three and the maximum total harmonic distortion as
power transformers rated as follows: 7.5MVA 132/11 percentage of the fundamental voltage and current was
kV transformer TR1 feeding five 11 kV feeders compared for the fourteen day period. Day seven shows
(Gauraka, Dikko, Minna Road, NNPC and Water Works the highest level of total harmonic distortion of voltage
11 kV feeders); 45MVA 132/33 kV transformer TR2 (VTHD) with a peak and average of 13.1% and 5.8%
feeding two 33 kV feeders (Abuja Steel and Field Base respectively. Generally, the harmonic distortions for all
33 kV Feeders); and 60MVA 132/33 kV transformer the days reveal a high level of distortion above tolerable
TR3 feeding two 33 kV feeders (Jere and Kantoma 33 kV limit of 2.5% as specified in Nigerian Grid code, 3.5% for
feeders). IEC 61000-4-30 (2003) and 5% for IEEE Std. 519
Suleja Power Network has three 33/11 kV Injection (1992)[17]-[19].
sub-stations located at Kantoma, Jere and Field Base. Therefore, the seventh day was selected for the
Kantoma 33/ 11 kV Injection substations have two harmonic load modelling as it provides the probable
power transformers each rated 15MVA with four 11 kV worst case scenario for the feeder. The loading of Abuja
feeders namely Dawaki, Hassan Dalhatu, RafinSanyi and Steel feeder was found to be near maximum on this day.
Madalla; Field Base 33/11 kV injection substation has Furthermore, ETAP deploys Discrete Fourier Transform
one power transformer rated 15MVA with two feeders (DFT) technique to model the distorted waveform and
namely Suleiman Barau and Tommy; while Jere 33/11 spectrum of any load model under case study. PQube’s
kV Injection substation has two power transformers harmonics data contains Amplitude, Phase Angle and
rated 7.5 MVA feeding three feeders namely Inter-harmonics Amplitude of both current and voltage
SabonLugbe, SabonDikko and Nasara 11 kV feeders. All captured separately on a ten minutes interval daily for
the other major network equipment like circuit the three phase lines herein referred to as Line 1 (Red
breakers, isolators and buses were also modeled and Phase), Line 2 (Yellow Phase) and Line 3 (Blue Phase).
simulated on the ETAP software package. The fourteen These three phase amplitudes and phase angles (in
11 kV distribution feeders were modeled as aggregate degrees) were combined to produce a resultant data for
loads based on load readings collected and analyzed the Abuja Steel line at every capture. Daily average
from field data. Single Line Diagram of the Suleja Power resultant values of amplitude and phase angle were
Network modeled for this research is shown in the computed for each harmonic order and then selected
attached Appendix. for simulation. Amplitudes of all other harmonic order
Nigerian Journal of Technology, Vol. 36, No. 4, October 2017 1186
HARMONIC LOAD MODELING: A CASE STUDY OF 33 KV ABUJA STEEL MILL FEEDER M. N. Nwohu, et al
were presented as percentage of the peak amplitude of summation of the impedance of the power transformer
the fundamental which is the 1st order harmonic supplying the Abuja Steel feeder (45MVA transformer
referred to as H1. Although the PQube records up to the TR ) and the network’s short circuit impedance
63rdorder Harmonic, this research truncates harmonics obtained from the grid data.
above 40th order since the higher order harmonics were
observed to have low amplitudes. 5. HARMONIC LOAD MODEL OF THE ABUJA STEEL
Fourier series expresses every periodic function as FEEDER
given in equation 1[11], [14], [15]: The ETAP software has a robust library that allows for
the building and to customize harmonics sources and
( ) ∑( cos sin ) ( ) store on the library’s shelves The relative amplitude
and phase angle of each harmonic order from the 1st to
K represents frequency of the Kth harmonics order.
40th were the input parameters to the model. The
Continuous Fourier transform is given by:
software then performs DFT to generate the spectrum
and waveform.
( ) ∫ ( ) ( )
∑ ( )
√ ( )
Fig. 2. Harmonics Current Source Model (Spectrum) of
Abuja Steel Feeder
tan ( )
The Thevenin/Norton Equivalent model was selected 5.1 Current Source Model
for this research as the power electronics circuits of the The spectrum in Figure 2 reveals that 11thorder
steel rolling mill and induction furnaces inject harmonic harmonics creates the highest level of distortion with a
current to the power network. Similarly, the network value of 3%. It is closely followed by the 13thorder
contributes its own component of harmonic current to harmonics with 2.3%. Its waveform is as shown in
the PCC. Therefore, harmonic vector method applied Figure 1. Fig. 3 and Fig. 4 depict the wave form and
here involves the representation of both the utility and spectrum respectively up to 40th harmonic order for the
customer sides with Thevenin/Norton equivalent feeder as a current source model.
circuits. Thevenin and Norton equivalent circuits were However, after harmonic analysis of the entire Suleja
used to separately compute the harmonic voltage and Distribution network with Abuja steel feeder modelled
current contributions to the PCC. The Suleja power as a harmonic current source, the propagation of
network’s reference impedance was calculated from the harmonics was observed on the various network buses,
Nigerian Journal of Technology, Vol. 36, No. 4, October 2017 1187
HARMONIC LOAD MODELING: A CASE STUDY OF 33 KV ABUJA STEEL MILL FEEDER M. N. Nwohu, et al
Fig. 7. Comparison of VTHD from Harmonics Analysis [8] Roger, C. G., Mark, F. M., Surya, S., Wayne, H. B.
Electrical Power Systems Quality, McGraw-Hill, 2004.
using Current Source Model and voltage Source Models
of the Abuja steel feeder.
Nigerian Journal of Technology, Vol. 36, No. 4, October 2017 1189
HARMONIC LOAD MODELING: A CASE STUDY OF 33 KV ABUJA STEEL MILL FEEDER M. N. Nwohu, et al
[9] Khaled, M. E., Hanafy, M. I., Al-Fahd M “Harmonic [14] Gül Ö and Gündoğdu T “Harmonic Contributions of
Modelling, Analysis and Mitigation of Kuwait EHV Utility and Customer Based on Load Model Using
Electrical Network” IEEE Electrical Engineering, Field Measurements” Journal of Power and Energy
pp.145-152, 2009. Engineering, 3, pp. 11-28, 2015.
[10] Ali S A “Modeling of power networks by ATP-Draw [15] Bilandžija D Barukčić M Ćorluka V “Using an
for harmonics propagation studies” Transactions on Evolutionary Algorithm for Harmonic Load Modeling
Electrical and Electronics Materials, pp. 283-290, by Norton and Thevenin Equivalents” International
2013. Journal of Electrical and Computer Engineering
Systems, pp. 39-45, 2014.
[11] Hagh, M. T., Milani, A. R., Azimizadeh, M. R..
“Investigation of Harmonic Current Source Effect on [16] Ünverdi E Yildiz A B “Equivalent Circuit Model
Distribution and Transmission Line Capacity and Containing AC and DC Side Harmonics of Rectifier
Losses: Case Study” International Journal on Circuits” 4th International Conference on Power
Technical and Physical Problems of Engineering Engineering, Energy and Electrical Drives,Istanbul,
(IJTPE), pp. 81-85, 2011. pp. 582-586, 2013.
[12] Venkatesh, C., Kumar D. S., Sarma, S., Sydulu, M. [17] The Grid Code for the Nigerian Electricity
“Modelling of Nonlinear Loads and Estimation of Transmission System. - Version 01. 2005
Harmonics in Industrial Distribution
[18] IEC 61000-3-6: Assessment of emission limits for
System” Fifteenth National Power Systems
distorting loads in MV and HV power systems -
Conference (NPSC), pp. 592-597, Bombay, 2008.
technical report part 3. 1996.
[13] Ma Z You G & Xu Z “Harmonic evaluation of grid
[19] IEEE Std. 519: Recommended Practices and
with multiple harmonic sources based on
Requirements for Harmonic Control in Electric Power
DIgSILENT” International Conference on Power
Systems, 1992.
System Technology, pp. 2332 – 2337, 2014.