JOURNAL OF OPTOELECTRONICS AND ADVANCED MATERIALS Vol. 18, No. 11-12, November - December 2016, p.

1007 - 1017

Single phase positive output super-lift luo converter fed

high power LED lamp with unity power factor and
reduced source current harmonics
J. GNANAVADIVELa,*, N. SENTHIL KUMARa, C. N. NAGA PRIYAb, S. T. JAYA CHRISTAa,
K. S. KRISHNA VENIa
a
Electrical and Electronics Engineering Department, Mepco Schlenk Engineering College, Sivakasi, Tamilnadu, India
b
Electrical and Electronics Engineering Department, Sri Vidya College of Engineering and Technology, Virudhunagar,
Tamilnadu, India

In this paper, a proposed single phase positive output super-lift Luo converter fed high power LED lamp for industrial
lighting with improved power quality at input side has been presented. This converter boosts up the input voltage with high
gain ratio. The proposed converter operates in continuous conduction mode. The design and simulation of 100 W LED
driver is carried out in Matlab/Simulink platform with constant load voltage of 48V. Closed loop control incorporates outer
voltage control and inner current control. A novel Adaptive Neuro Fuzzy Inference System (ANFIS) controller is proposed
for voltage control, which tunes the Proportional and Integral (PI) parameters and Hysteresis controller is used in current
loop. From the simulation results, it is found that the proposed LED lamp driver with controller combination gives the best
result with a low total harmonic distortion and unity power factor with no peak-over shoot at rated condition. The proposed
system is implemented in real time in order to validate the performance of the proposed system over a wide range of output
power with improved power quality at AC mains using Xilinx Spartan-6 XC6SLX25 FPGA board.

(Received December 10, 2015; accepted November 25, 2016)

Keywords: Adaptive Neuro Fuzzy Inference System, PI controller, Positive output super-lift Luo Converter, LED, Input Power Factor,
Total Harmonic Distortion

1. Introduction regulation such as IEC 61000-3-2 class C enacted to limit

the source current harmonics and to guarantee input power
Now-a-days, light emitting diode lamps are gaining factor of 0.9 at least.
popularity in both industrial and domestic usage. It can be Therefore, some steps have to be taken in order to
used in many applications even from street lights to car improve the power factor and the best method to
head lights. Improvement in the lighting technology from accomplish is to use power factor correction (PFC)
the past few years result in plenty of advantages such as converters as a LED driver circuit. The commonly used
high luminous efficiency, long life, low power power factor correction converters are buck, boost, buck-
consumption, environment friendly because there is less boost, Cuk, Sepic , Zeta and flyback converters along with
mercury contents, easily dimmable, low maintenance, diode bridge rectifier[13]-[17] and these converters can
flicker less start, robust in structure and least affected by operate in continuous conduction mode or discontinuous
vibrations[1]-[6]. Even though LED lamp offers many conduction mode to achieve better power quality in the
merits, it also has some demerits such as generation of supply side. To reduce the switching stress and conduction
harmonics in the input AC side and low input power factor losses, Boost-interleaved buck-boost converter is
because of switching devices used in the LED driver. LED employed in [13], and it results in THD of nearly 4.6%
lighting module consists of two major parts: LED driver which is within in the limit of IEC and IEEE standards.
and LED module. In many research articles they focused Fly back buck-boost converter [14] is introduced for
on the development of LED lamps, lighting control and low power applications with voltage follower approach in
illumination on the work surface [7] - [12]. Only few order to get low THD (4.8%) and improved power factor.
research papers so far published focus on the improvement (0.992). PFC Cuk converter is used with current multiplier
of power quality in the AC supply side of the LED lamp approach to maintain harmonics ranging from 3.85% to
driver. 7.42% with variation in the DC link voltage[15]. To
In general, an AC/DC converter uses as a diode bridge reduce the conduction losses, SEPIC converter is
rectifier and a large value of capacitor to obtain smooth dc employed in [16]. Isolated zeta converter is used for PFC
link voltage. This type of LED driver introduces highly correction in [17]. With this system, source current
distorted source current which results in high value of total harmonics gets reduced to 3.7% and also the power factor
harmonic distortion and very low input power factor. To gets improved in the front end. Three level boost converter
maintain the power quality, standards of the harmonic
1008 J. Gnanavadivel, N. Senthil Kumar, C. N. Naga Priya, S. T. Jaya Christa, K. S. Krishna Veni

and low frequency PWM converter are used for PFC in Among the various techniques of artificial
LED drive applications [18], [19]. intelligence, fuzzy logic is the most popular one for
Bridgeless topology was introduced, in order to controlling a system. An intelligent controller, with proper
reduce the conduction losses possessed by the design, works very well even with an approximate model
conventional rectifier topologies. Bridgeless Sepic of the system. The adaptive-fuzzy controller is used to
converter is employed at the front end to improve the improve the system’s dynamic response and the
power quality indices which results in further decline in parameters of PI controllers are adjusted online. The
THD in the order of 2.16% [20]. The harmonic distortions conventional PI controller produces constant Kp, Ki values
are only 2.12% in case of bridgeless Zeta converter [21]. which need to be changed during continuous system
In [22] bridgeless Luo converter topology is presented in variations [34]. The intelligent agent based Adaptive
order to improve the power factor and the harmonics in Neuro-Fuzzy Inference System (ANFIS) was developed to
this system is found to be 4%. Even though, the above make system stable even under external disturbances [35] -
mentioned topologies operate with minimized conduction [38].
losses, they account for high switching losses due to The main objective of the proposed system is to
increased number of components. Also, for these systems design a single phase positive output super-lift Luo
the voltage transfer gain is low and ripple content in output converter based LED driver with high power factor and
voltage is high. reduced THD even under source voltage and load
In this paper, a single phase positive output super-lift fluctuations. Fuzzy supervised ANFIS controller is used to
Luo converter is considered to improve the quality of tune the PI parameters. This properly tuned system makes
power and to mitigate various power quality issues. Super- it to offer plenty of advantages such as regulated output
lift Luo converter is extensively used because of its voltage with very low overshoot, less settling time, very
excellent response even during load variations , supply low harmonics and unity power factor. The proposed work
voltage fluctuations and the voltage lift technique is used is rationalised by comparing the results obtained in
in many high voltage gain converters in which the output simulation with the PI-hysteresis and fuzzy tuned PI-
voltage increases stage by stage progressively[23]-[27]. hysteresis.
Though the gain ratio increases in geometric progression
on stage-by-stage in super-lift technique, they are quite
complex to design. To overcome this disadvantage, an 2. Proposed topology of LED driver
efficient positive output super-lift Luo converter is
introduced to enhance the voltage transfer gain in power Fig. 1 shows the circuit diagram of single phase
law terms [28]. positive output super-lift Luo converter. The proposed
In regard to the controller part, generally PI controller system consists of a single phase diode bridge rectifier,
is used which offers several advantages such as stability inductors (filter and boost), fast recovery diodes, DC
even for large line and load variations, reduced steady capacitors and switching component. The proposed system
state error, robustness, good dynamic response and easy is designed properly and their performances are
implementation[29]. Though PI controller posses lot of investigated for a wide range of variations in load side and
advantages, the main challenge is involved in the design of supply side.
current loop. Because, it requires an accurate mathematical
model of the plant otherwise its performance during load
disturbance and parameter variation are not satisfactory
[30]. Whereas Hysteresis current controller does not
require any accurate mathematical modeling when the
controller is being designed [31]. Also this technique has
an advantage of yielding instantaneous current control,
which results in a very fast response and increased switch
reliability. Fig. 1. Proposed Single phase positive output super-lift Luo
The hysteresis current control method is widely used converter based LED driver
because of its ease in implementation and quick current
controllability [32]. The hysteresis band current control is 2.1 Modes of operation of proposed LED driver
robust and it provides excellent dynamics and fastest circuit
control response with minimum hardware. The voltage
controller performance is further improved by designing Based on the switching action of the proposed LED
good input and output membership function of the fuzzy driver circuit, the modes of operation can be classified into
logic controller. Traditional ways for designing controllers two types and they are given below.
of DC/DC converters are based on small signal model and
the validation of small signal model is limited by changes
in the operation point [33].
 Single phase positive output super-lift luo converter fed high power LED lamp with unity power factor … 1009

Mode 1: 3. Design of the proposed high power factor

LED Driver

The following assumptions are considered to design

the proposed high power factor LED driver circuit.
1. All the device components of the proposed LED
driver circuit are ideal one.
2. Under normal operating condition Light Emitting
Diode behaves as a resistor.
The inductance L1 is determined by the peak-to-peak
ripple current ΔI L1 . Therefore, the inductance L1 can be
Fig. 2 Mode 1 operation
represented as,
V0  2Vin
L1  (5)
I L1 f
During mode 1 operation, the switch S is considered
to be in ON state thereby, capacitor C1 gets charged to Vs.
The current iL1 flowing through inductor increases with The capacitor value can be calculated by the formula
voltage Vs. The input current is will be equal to iL1 + i c1.
given below,
The mode 1 operation is depicted clearly through Fig.2.
1  k  V
The characteristic equation during mode 1 operation are C1  Cd  0
(6)
given below, fR lamp V0

di The output voltage equation of proposed LED driver

vs  L1  vc1 (1)
dt is,
2k
V0  Vin (7)
dv v 1 k
Cd 0  0  0 (2)
dt R lamp
The voltage across the diode bridge rectifier is given
by,
Mode 2: 2  Vm 
Vin  (8)
3.14
where
k is the duty cycle
Vo is the DC output voltage
Vm is the peak value of the input voltage
Vs is the supply voltage
Vin is the voltage across the diode bridge rectifier
ΔI L is the peak to peak ripple inductor current(5% of
1

inductor current I L )
1
Fig. 3. Mode 2 operation
f is the switching frequency
ΔVo is the output ripple voltage (2% of output voltage Vo )
Fig.3 clearly shows the mode 2 operation of single Rlamp is the LED lamp resistance under normal operating
phase positive output super-lift Luo converter based LED condition
drive. During mode 2 operation, switch S is considered to By using the design equations, the values of circuit
be in OFF state, diode D1 is reversed biased and D2 components are calculated and they are listed in Table 1.
becomes forward biased. During mode 2 operation, input
current will be equal to iL1 (=ico).The characteristic Table 1. Design parameters of proposed LED driver.
equations are,
Parameter Specification
di L1 Supply voltage (Vs) 20 V
vs  L1  vc1  v0 (3) LED lamp voltage(VLamp) 48 V
dt
LED lamp power(Po) 100 W
dv0 v Supply frequency(fL) 50 Hz
Cd  0  i L1 (4) Inductor Lf 1.4 mH
dt R lamp Inductor L1 1.6 mH
Capacitor C1= Cd 2000 µF
1010 J. Gnanavadivel, N. Senthil Kumar, C. N. Naga Priya, S. T. Jaya Christa, K. S. Krishna Veni

4. Block diagram of the proposed controller 5. ANFIS tuned PI controller for high power
factor LED driver
The control structure of the proposed single phase
positive output super-lift Luo converter based LED driver The motive behind the design of ANFIS tuned PI
is shown in Fig.4. The proposed LED driver control controller is to achieve the power quality indices such as
system consists of voltage controller, power estimator, reduction in source side harmonics, unity power factor and
phase locked loop and current controller. The output the load side parameters like minimum settling time and
voltage is sensed and it is compared with the reference rise time with no peak overshoot. The PI parameters are
voltage. After comparison, the obtained error signal is fed tuned online using ANFIS controller to improve the
to the voltage controller. Then, the output signal from the system performance such as high input power factor, less
power estimator block and the control signal from the harmonic distortion.
voltage controller are added. The obtained value is then ANFIS tuned PI controller needs no separate
multiplied with the PLL output to get reference inductor mathematical modelling of the system similar to Fuzzy
current. The rectified sinusoidal signal is multiplied by the logic controller. Here, the data to be loaded is trained
amplitude of command current to form the reference again and again according to system response to get zero
inductor current. The reference inductor current is error from the controller so as to produce perfect results
expressed as, both at input side and output side of the converter. Fig.5
 vi  shows the block diagram of ANFIS tuned PI controller for
i ref   0 0  K p verr  K i  verr  sin t (9) LED driver. The control signal from ANFIS controller is
v 
 s.peak  given to the PI controller so that it tunes the PI parameters.

verr  vref  v0 (10)

where, Kp–Proportional gain, Ki-Integral gain, Vref -

reference voltage, Vo - output voltage

Fig. 5. Block diagram of ANFIS tuned PI controller

for LED driver

The structure of the ANFIS network is composed of

set of units and connections arranged in network layers
from layer1 to layer5 as shown in Fig. 6. ANFIS controller
contains four important blocks such as fuzzification,
knowledge base, normalisation of fuzzy rules and
defuzzification.
Layer 1: It consists of input variables in fuzzy form
which has triangular or bell shaped membership functions.
Layer 2: It is the membership layer and it checks for
the weights of each membership functions. It receives the
input values from the first layer and based on membership
functions; represent the fuzzy sets of the respective input
Fig. 4. Control diagram of the proposed single phase
positive output super-lift Luo converter based LED driver variables.
Layer 3: It is called as rule layer and it receives input
from the previous layer. Each neuron in this layer
The reference inductor current is then compared with performs the pre-condition matching of the fuzzy rules.
the actual inductor current and fed into the inner current This layer computes the activation level of each rule and
loop controller. For voltage control, ANFIS tuned PI is the number of layers equals to the number of fuzzy rules.
proposed and for current control, hysteresis control Each node in this layer calculates the weights which will
generates PWM signal, to maintain proper illumination of be normalized.
LED modules. The power switch ‘S’ is operating at 10kHz Layer 4: It is the defuzzification layer which provides
frequency. the output values resulting from the inference of rules.
Layer 5: It is called as the output layer which sums up
all the inputs coming from layer4 and transforms the fuzzy
classification results into a crisp value.
 Single phase positive output super-lift luo converter fed high power LED lamp with unity power factor … 1011

Fig. 6 Layers in ANFIS structure

The IF-THEN rules used are given below.

Rule 1: If e is A1 and e is B1 then 6. Simulation results and analysis of single
f1  P1  e  R1  e  S1 phase positive output super-lift Luo
converter based LED driver with ANFIS
Rule 2: If e is A2 and e is B2 then tuned PI voltage controller and hysteresis
f2  P2  e  R 2  e  S2 current controller
Rule i-1: If e is Ai-1 and e is Bi-1 then
fi 1  Pi 1  e  R i 1  e  Si 1
Fig. 7 shows the simulation diagram of single phase
positive output super-lift Luo converter based LED driver
Rule i : If e is Aj and e is Bj then with ANFIS tuned PI voltage and hysteresis current
fi  Pi  e  R i  e  Si controller. The error and change in error are fed as inputs
where, to ANFIS voltage controller. Here ANFIS voltage
controller is designed to produce two control signals which
e  Vref  Vo
are tuned by it with respect to the variations of the system.
d(Vref  Vo ) The tuned values from the controller are multiplied with
e  the error signal and are added. Then the resultant signal is
dt added with the output from the power estimation block
A and B are the input variables of the fuzzy membership and then multiplied with | sin 𝜔𝑡 | to generate reference
function, e and Δe are error and change in error and q is inductor current signal. The actual inductor current is
the number of membership functions for the fuzzy inputs e compared with the reference signal and given to hysteresis
and Δe. 𝑓𝑖 is the linear consequent function defined in current controller which limits the signal in the band limit
terms of inputs e and Δe. Pi, Ri and Si are consequent of 0.001.
parameters of an ANFIS fuzzy model. Same layer nodes of
an ANFIS model have similar functions.
1012 J. Gnanavadivel, N. Senthil Kumar, C. N. Naga Priya, S. T. Jaya Christa, K. S. Krishna Veni

Fig. 7. Simulation diagram of single phase positive output super-lift Luo converter based LED
driver with ANFIS tuned PI voltage controller and hysteresis current Controller

(a)
(a)

(b)
(b)
Fig. 9. Membership function for (a) input 1 (error) (b) input
Fig. 8. a) Initial Rule base (b) Final rule base 2(change in error)

Fig.8 shows the rule base model obtained before and

after training the data into the ANFIS respectively.
 Single phase positive output super-lift luo converter fed high power LED lamp with unity power factor … 1013

Then the gradient vector is accumulated for each

input–output training dataset. At the end of the backward
pass for all training datasets, the parameters are updated by
gradient descent method. Once updating of premise and
consequent parameters are completed, proper sets of
membership function and rule base are selected for fuzzy
inference system. After proper rules are selected and fired,
the control signal required to obtain the optimal output is
generated. To train the ANFIS controller, the network is
(a)
trained using MATLAB Simulink tool box. The desired
output is trained using the function ‘ANFIS’ in the
MATLAB toolbox. From the training, a fuzzy inference
system with adjusted membership function has been
obtained.

(b)

Fig. 10. Training error Vs Epochs for (a) Kp (b) Ki

(a)
Fig. 9 (a) and (b) show the membership functions of
input error and change in error respectively after training
the data. Fig. 10 (a) and (b) show the training of Kp and Ki
values using Neuro-Fuzzy system. The values are trained
up to 10 Epochs and the error value is minimised to a very
low value which is shown in the Fig. 10.
Fig. 11 shows the layer description of ANFIS
obtained during the training of data in ANFIS. In each
iterations, two sequences of operations occur; they are
forward pass and a backward pass sequence. During (b)
forward pass, after an input data is processed, the node
outputs are updated layer by layer until, layer 4 is reached. Fig. 12. (a) Input voltage and input current waveforms
This process is repeated for all training input–output (b) Output voltage waveform for ANFIS tuned PI voltage
datasets and then, the consequent parameters are identified and hysteresis current controller for proposed LED lamp
by least square estimation. In backward pass, the driver at rated lamp power
derivative of the error signals with respect to each node
propagates from the output end towards the input end. Fig. 12 (a) shows the input voltage and input current
waveforms which are in phase with each other gives unity
power factor and the source current distortion is reduced.
Fig. 12 (b) shows the output voltage waveform for ANFIS
tuned PI voltage and hysteresis current controller and it is
observed that there is no peak overshoot and the settling
time is 0.55sec.
Fig. 13 (a) shows the FFT analysis of ANFIS tuned PI
voltage and hysteresis current controller and THD is
reduced to 0.44% during rated LED lamp load condition.
Fig. 13 (b) shows the FFT analysis at 10% of LED lamp
load with a THD of 1.87% for ANFIS tuned PI voltage
and hysteresis current controller.
From the Table 2, it is observed that the output
voltage is regulated and maintained constant at 48 V, input
power factor is close to unity and source current
harmonics reduces from 1.87% to 0.44% for a variation of
Fig. 11. Layers in ANFIS load from 10% to 100%. It is observed that even under
load variations the THD value lies within 5% for ANFIS
tuned PI voltage and Hysteresis current controller.
1014 J. Gnanavadivel, N. Senthil Kumar, C. N. Naga Priya, S. T. Jaya Christa, K. S. Krishna Veni

Table 3. Performance analysis of single phase positive

output super-lift Luo converter based LED lamp driver
for ANFIS tuned PI voltage and hysteresis current
control under supply voltage variations

Supply Output Input Source Source

voltage voltage power current Current
(V) (V) factor THD(%) (A)
13 48 0.9979 2.79 10.39
15 48 0.9981 1.97 8.025
(a) 20 48 1 0.44 5.797
25 48 0.9999 0.96 4.472
30 48 0.9997 1.24 3.669
35 48 0.9996 1.40 3.125

Table 3 shows the performance analysis of single

phase positive output super-lift Luo converter based LED
lamp drive under supply voltage variations. The THD
value decreases from 2.79% to 0.44% for supply voltage
(b) variations from 13V to 35V. It is also observed that the
THD value is less than 5% for a wide variation in supply
Fig. 13. (a) FFT analysis at 100% of LED lamp load voltage.
power (b) FFT analysis at 10% of LED lamp load power
for ANFIS tuned PI voltage and hysteresis current Table 4. Source current harmonics for various
controller controllers for a wide variation in LED lamp load power

Source current THD (%)

Table 2. Performance analysis of single phase positive LED lamp
output super-lift Luo converter based LED driver for Fuzzy tuned ANFIS tuned
load PI -Hys
ANFIS tuned PI voltage and hysteresis current controller PI- Hys PI - Hys
under load variations power(W)
100 2.72 1.21 0.44
Source 90 3.19 1.46 0.49
LED lamp Output Input Source
current 80 3.28 1.58 0.51
load voltage power current
THD
power(W) (V) factor (A) 70 3.54 1.72 0.53
(%)
100 48 1 0.44 5.797 60 3.72 1.91 0.66
50 5.61 2.24 0.76
90 48 1 0.49 5.193
40 6.21 2.63 0.85
80 48 1 0.51 4.622
30 7.65 3.75 0.93
70 48 1 0.53 4.013 20 8.78 4.96 1.17
60 48 1 0.66 3.471 10 19.6 7.94 1.87
50 48 1 0.76 2.894
Table 4 shows the comparison of source current
40 48 1 0.85 2.324 harmonics obtained by varying the load from 10% to
30 48 0.9999 0.93 1.761 100% for various controller combinations. For PI voltage
and hysteresis current controller, the THD reduces from
20 48 0.9999 1.17 1.182 19.6% to 2.72%. For fuzzy tuned PI voltage and hysteresis
10 48 0.9926 1.87 0.611 current controller, the THD decreases from 7.94% to
1.21%. Then to improve the response further, ANFIS
tuned PI voltage controller and hysteresis current
controller is employed. From the Table 4, it is evident that
the ANFIS tuned PI voltage and hysteresis current
controller combination gives the best result of 1.87% to
0.44% THD, even for a wide range of load variations.
 Single phase positive output super-lift luo converter fed high power LED lamp with unity power factor … 1015

Table 5. Power factor for various controllers for a wide proposed controller for LED driver gives the best result
variation in LED lamp load power with a decrease in THD from 1.97% to 0.44%.

Table 7. Power factor for various controllers with a wide

Power factor variation in supply voltage for proposed LED driver
LED lamp
load
ANFIS Power factor
power(W) PI- Fuzzy tuned Source
tuned PI- ANFIS
Hys PI -Hys voltage Fuzzy tuned
Hys PI-Hys tuned PI -
100 0.9995 0.9999 1 (V) PI -Hys
Hys
90 0.9992 0.9998 1 15 0.9966 0.9996 0.9997
80 0.9991 0.9998 1 20 0.9995 0.9999 1
70 0.9990 0.9997 1 25 0.9991 0.9997 0.9999
60 0.9988 0.9996 1
30 0.9990 0.9996 0.9998
50 0.9983 0.9995 1
35 0.9986 0.9995 0.9997
40 0.9976 0.9993 1
30 0.9959 0.9989 0.9999
20 0.9910 0.9972 0.9999 By varying the supply voltage from 15V to 35V, the
10 0.9706 0.9919 0.9926 power factor is noted for various controller combinations
and is tabulated in Table 7. The power factor value
increases from 0.9966 to 0.9995 for PI voltage and
Table 5 gives the comparison of power factor hysteresis current controller. With fuzzy tuned PI voltage
obtained by varying the LED lamp load power from 10% and hysteresis current controller, the power factor value is
to 100% for various controller combinations. With PI improved from 0.9995 to 0.9999. Finally ANFIS tuned PI
voltage and hysteresis current controller, the power factor voltage controller with hysteresis current controller is
value increases from 0.9706 to 0.9995. Fuzzy tuned PI devised to improve the overall system performance. From
voltage and hysteresis current controller gives a power the Table 7, it can be found that for ANFIS tuned PI
factor value of 0.9919 for 10% load and 0.9999 for rated voltage and hysteresis current controller, the power factor
load. To improve the system response ANFIS tuned PI value increases from 0.9997 to 1.
voltage and hysteresis current controller combination is
employed and an increase in power factor from 0.9926 to
1 is obtained for an increase in load from 10% to 100%. 7. Experimental results of proposed high
power factor LED driver

Table 6. Source current harmonics for various To verify the validity of the proposed ANFIS-tuned PI
controllers with a wide variation in supply voltage for voltage controller and hysteresis current controller, a
proposed LED driver single-phase POSLLUO converter based LED driver has
been built and tested in the laboratory. It is shown Fig. 14.
Source current THD (%) The power switching devices and various components of
Source the prototype include an MUR360 input rectifier bridge,
ANFIS
voltage Fuzzy tuned IRF250 power MOSFET switch, a 1.6mH boost inductor,
PI-Hys tuned
(V) PI -Hys a 2000 µF output filter capacitor, an FPGA Spartan-6
PI - Hys
15 6.60 2.20 1.97 controller, an HCPL-7840 voltage sensor, a WCS 2705
20 2.72 1.21 0.44 hall effect current sensor and a LED panel.
25 1.54 1.99 0.96
30 2.42 2.57 1.24
35 3.14 2.64 1.40

Table 6 shows the comparison of source current

harmonics by varying the supply voltage from 15V to 35V
below and above the rated input voltage (20V). With the
PI voltage and hysteresis current controller, the harmonics
decreases from 6.60% to 1.54%. For fuzzy tuned PI
voltage and hysteresis current controller, the THD value
decreases from 2.64% to 1.2%. To improve the system
performance further, the ANFIS tuned PI voltage
controller is used at voltage side with hysteresis controller
at current side. From the Table 6, it is evident that, the Fig. 14. Hardware setup of proposed LED driver
1016 J. Gnanavadivel, N. Senthil Kumar, C. N. Naga Priya, S. T. Jaya Christa, K. S. Krishna Veni

Power factor, THD (%)

2.5
2 power
factor
1.5
1 THD (%)
0.5
10 20 30 40 50 60 70 80 90 100
Output Power (W )

Fig. 15 c. Experimental result for ANFIS tuned PI – hysteresis

Fig.15a. Experimental input voltage and input current controller of proposed LED driver
waveforms at rated power with ANFIS tuned PI voltage
controller and hysteresis current controller for LED
driver
Fig. 15 c depicts the practical readings of the
proposed high power factor LED driver with the variation
Fig. 15a shows the experimental waveforms of input
in load.
voltage and input current for rated lamp load power with
the ANFIS tuned PI voltage controller and the hysteresis
current controller for LED driver. The input power factor
8. Performance analysis of conventional
and input current THD are measured with the power
boost converter with proposed LED driver
quality analyzer shown in Fig. 15a. From Fig. 15 a, the
input power factor is 0.9990 and the input current THD is
The performance of the conventional boost LED
1.240% for the proposed LED driver.
driver and proposed LED drive in terms of the obtained
input current THD and input PF at the supply mains with
output power is evaluated as shown in Fig.16 a and 16 b
respectively. ANFIS tuned PI voltage controller and
hysteresis current controller are implemented for both
LED drive. As shown theses figures, a very low source
current harmonics and high input power factor are
achieved for the proposed LED driver.

8
Source current THD(%)

6 Conventional
boost
Fig.15b. Experimental input voltage and input current 4 converter
waveforms at 20% of load power with ANFIS tuned PI
voltage controller and hysteresis current controller for 2 Proposed
LED driver Converter
0
Fig. 15b shows the experimental waveforms of input 10 20 30 40 50 60 70 80 90 100
voltage and input current for 20% of lamp load power with Output Power (W )
the ANFIS tuned PI voltage controller and the hysteresis Fig. 16 a. Comparison of source current harmonics for
current controller for LED driver. The input power factor conventional boost and proposed topology LED driver
and input current THD are measured with the power
quality analyzer shown in Fig. 15b. From Fig. 15 b, the
input power factor is 0.9981 and the THD is 2.606% for 1
Power factor

the proposed LED driver. Conventional

boost
0.8 converter

Proposed
0.6 Converter
10 20 30 40 50 60 70 80 90 100

Output Power (W )

Fig. 16 b. Comparison of source side power factor for

conventional boost and proposed topology LED driver
 Single phase positive output super-lift luo converter fed high power LED lamp with unity power factor … 1017

9. Conclusion [16] J. M. Kwon, W. Y. Choi, J. J. Lee, E. H. Kim, B. H.

Kwon, IEEE proceedings on Electronics Power
An improved power quality positive output super-lift Applications 5, 673 (2006).
Luo converter is proposed for high power LED industrial [17] Vashist Bist, Bhim Singh, IEEE Transactions on
lighting applications. The proposed LED driver has shown Industrial Informatics 4, 2064 (2014).
high level performance such as low total harmonic [18] J. Gnanavadivel, N. Senthil Kumar, P. Yogalakshmi,
distortion and unity power factor for variation of the LED J. Optoelectron. Adv. M. 18, 459 (2106).
lamp load power and variation of the supply voltage. From [19] Hung-Liang Chang, Ying-Nong Chang, Chun-An
the simulation results, it is noted that the proposed driver Chang, Chien-Hsuan Chang, Yu-Hung Lin, IET
with ANFIS tuned PI voltage and hysteresis current Power Electronics 9, 2139(2016).
controller offers unsurpassed results with a very low THD [20] H. Esam, Ismail, IEEE Transactions on Industrial
of 0.44% with unity power factor at source side and on Electronics 4, 1147 (2009).
load side, the voltage is regulated with no peak overshoot, [21] Bhim Singh, Ambrish Chandra. IEEE Transactions on
less settling time and rise time at load side under rated Industry Applications 51, 1179 (2015).
LED lamp load power. The ANFIS tuned PI voltage [22] Y. He, F. L. Luo, IEEE Proceeding on Electric Power
controller and the hysteresis current controller is Application 152, 1239 (2005).
implemented in a FPGA based hardware platform for [23] F. L. Luo, Electric power Applications, IEEE
proposed LED driver, with reduced input current THD of Proceedings on Electric power Applications146, 4415
1.24% in the experiment, along with the input power (1999).
factor close to unity without any source side filter. [24] F.L. Luo, IEEE Proceedings of the IEEE-PESC’98 /
Fukuoka; Japan 22, 1783 (1999).
[25] F. L. Luo, Proceedings of the second World Energy
References System international conference WES’98 Tornoto,
Canada, 254 (1999).
[1] H. J. Chiu and S. J. Cheng, IEEE Transaction on [26] F. L. Luo, H. Ye, IEEE Transactions on Power
Industrial Electronics 5, 2751(2007). Electronics 18, 105 (2003).
[2] H. Cho and O. K. Kwon, IEEE Transaction on [27] M. Zhu, F. L. Luo, Journal of Power Electronics9, 854
Consumer Electronics 4, 2054 (2010). (2009).
[3] C. S. Moo, Y. J. Chen, and W. C. Yang, IEEE [28] F. L. Luo, H, Ye, IEEE Transaction on Power
Transaction on Power Electronics 11, 4613 (2012). Electronics, 18,105 (2003).
[4] T. Hsieh, B. D. Liu, J. F. Wu, C. L. Fang, H. H. Tsai, [29] P. Comines, N. Munro, IEEE Proceeding Control
Y. Z. Juang , IEEE Transactions on Power Electronics Theory Application 149, 46 (2002).
12, 4562 (2012). [30] G. L, Diego, F Arturo, A, Manuel, IEEE
[5] B. Lee, H. Kim, and C. Rim, IEEE Transactions on Transactionon Power Electronics 2, 635(2008).
Power Electronics 26, 3694 (2011). [31] K. H. C. Martin, H. L. C. Martin, K. T. Chi, IEEE
[6] N. Chen, H. S. H. Chung, IEEE Transactions on Transaction on Industrial Electronics 2, 665 (2008).
Power Electronics 5, 2551(2013). [32] R. Wu, S. B. Dewan, G. R. Slemon, IEEE
[7] O. Faruk Farsakoglu, I. Atik, J. Optoelectron. Adv. M. Transactions on Industry Applications 26, 880 (1990).
17, 277(2015). [33] W. C. So, C.K. Tse, Y. S, Lee, IEEE Transactions on
[8] An Tang, Yongtao Zhao, Haoming Zhang, Tao Ma, Power Electronics 11, 24 (1996).
Fengxiang Shao, Hongsong Zhang, Optoelectron. [34] Y. P. Pan, J. E. Meng, D.P. Huang, Q.R. Wang, IEEE
Adv. Mat. 9, 20 (2015). Transactions on Fuzzy Systems 5, 807 (2011).
[9] J. Liu, H. Yang, P. Jiang, M. Yu, Y. Huang, J. Yang, [35] R. K. Mudi, N. R, Pal, IEEE Transactions on Fuzzy
J. Optoelectron. Adv. M. 17, 62 (2015). Systems 1, 2(1999).
[10] Wentao Zhang, Peicong Zhang, Junfeng Li, Kehui [36] A. Al-Hmouz, J. Shen, R. Al-Hmouz, J. Yan, IEEE
Qiu, Optoelectron. Adv. Mat. 8, 37 (2014). Transaction on Learning Technologies 5, 226 (2012).
[11] Luqiao Yin, Jinlong Zhang, Yang Bai, Weiqiao Yang, [37] K Premkumar, B. V. Manikandan,, Neurocomputing,
Jianhua Zhang, Optoelectron. Adv. Mat. 8, 427(2014). 138, 260 (2014).
[12] O. F. Farsakoglu, H. Y. Hasirci, J. Optoelectron. [38] K. Premkumar, B. V, Manikandan, Neurocomputing,
Adv. M.17, 816(2015). 157, 76 (2015).
[13] Jingquan Chen, Dragan Maksimovic, Robert W.
Erickson, IEEE Transactions on Power Electronics 2,
320 (2006).
[14] Bhim Singh, Ganesh Dutt Chaturvedi, Journal of
Power Electronics, 4, 318 (2007).
[15] Vashist Bist, Bhim Singh. IEEE Transaction on _______________________________
*
Industry Applications 2, 871 (2015). Corresponding author: [email protected],
[email protected]