International Journal of Power Electronics and Drive Systems (IJPEDS)

Vol. 13, No. 1, March 2022, pp. 298~308

ISSN: 2088-8694, DOI: 10.11591/ijpeds.v13.i1.pp298-308  298

Optimization of passive filter components through active

filtering of current ripple reduction in an inverter

Jeyaraman Ramakrishnan1, Chengalvarayan Natarajan Ravi2

1
Department of Electrical and Electronics Engineering, Sathyabama Institute of Science and Technology, Chennai, India
2
Department of Electrical and Electronics Engineering, Vidya Jothi Institute of Technology, Hyderabad, India

Article Info ABSTRACT

Article history: This paper employs the active input current filtering concept developed and
explained to optimize the conventional passive filter components in a two-
Received Aug 2, 2021 stage telecommunication inverter. A prototype inverter has been designed
Revised Jan 21, 2022 for a rated power of 800 W. Optimization of the passive filter components
Accepted Jan 28, 2022 consisting of inductance inductors tank (LTank) and capacitance capacitors
tank (CTank) was performed through simulations. The simulation model
used for this optimization was validated using the developed prototype. The
Keywords: focus of this work is to limit the input direct current (DC) current ripple
within the limits specified in European Telecommunications Standards
Direct current-direct current Institute (ETSI) 300132-2 for a battery supplied telecommunication inverter
converter and optimize the passive filter components. Using the active filtering
Passive filter technique, limit the input current ripple and simultaneously reducing the
Push-pull converter filter capacity in DC-link. This advantage of active filtering technology over
Ripple reduction the conventional passive inductors and capacitors (LC) filter concept is
Second-order ripple explored and demonstrated in this paper.
Two-stage inverter This is an open access article under the CC BY-SA license.

Corresponding Author:
Jeyaraman Ramakrishnan
Department of Electrical and Electronics Engineering, Sathyabama Institute of Science and Technology
SH 49A, Semmancheri, Chennai, Tamil Nadu 600119, India
Email: [email protected]

1. INTRODUCTION
In most telecommunication and industrial applications, the inverter is mainly used for power
backup. The single-phase two-stage inverter is widely used in telecommunication and data center
applications for critical backup solutions. This inverter contains a direct current-direct current (DC-DC)
converter stage and a direct current-alternating current (DC-AC) inverter stage. The battery DC supply is
boosted to a high voltage by using a DC-DC converter and this stores in the DC link capacitors [1]. Then this
high voltage is converted to an AC supply by using a DC-AC converter. The battery power is consumed as a
sinusoidal form of power and it causes a second-order current ripple in the battery. The product of the V and
I is equal to the power, which is in a sine pattern. Hence, this inverter intermediate stage DC link current has
a second-order sinusoidal ripple. It is also reflected in the DC input battery side. The DC input current has an
AC ripple signature and it resembles the AC output power of the inverter supply. The DC input has flat DC
and AC components of current. The battery is an electrochemical device. This AC ripple current reduces the
performance of the battery's backup time [2]–[5]. In the study of battery AC ripple frequency, the battery
manufacturers typically specify that the acceptable ripple current percentage is less than 10% and the high-
frequency ripple creates a significantly less effect in the battery chemical reaction and low-frequency ripples
are creating an overheating issue in the battery and reduce the inverter efficiency [6]–[9]. This affects the
overall life span of the telecommunication battery power system. The telecommunication system operates

Journal homepage: http://ijpeds.iaescore.com

Int J Pow Elec & Dri Syst ISSN: 2088-8694  299

with a -48 V DC supply. It is mainly derived from the battery bank and some equipment operates with an AC
supply. In this application, the industrial telecom inverter is employed. The inverter takes DC power from the
battery bank and in this case, the inverter injects a second-order noise into the telecommunication equipment.
These frequencies create a humming noise in telecommunication networks. The telecommunication European
Telecommunications Standards Institute (ETSI) standard specifies the psophometric low-frequency spectrum
range and limits. It should be satisfied by all telecommunication equipment. To meet the psophometric noise
emission limits, the inverter DC input current ripple percentage should be less than the 3% level [10]. This
paper primary focus is on determining the root cause of the second-order ripple generation phenomenon in an
inverter and optimizing inductors and capacitors (LC) filter components using passive and active approaches
ripple reduction techniques [11]. The amount of DC ripple current can be reduced by increasing the LC filter
value in the DC link LC filter or by implementing the ripple reduction technique in the DC-DC converter.
The passive method enhances the LC filter used in the DC-DC converter intermediate section DC-link
output. It needs a bigger value of inductance and capacitance to archive less than 3% acceptance level and
also it needs more room space to accommodate the filter. In the active method, the dual feedback and
proportional integral (PI) closed-loop control techniques help to reduce DC ripples [12]–[14]. This paper
explains about LC filter method and a very commonly used active filter method and optimizing the LC filter
size in comparison with the passive method. In the simulation and experiment, both methods are
implemented and the characteristics of ripple and percentage of the DC input ripples are compared.

2. TOPOLOGY
The passive and active DC input current ripple reduction method has been experimented with on a
two-stage inverter. The DC-DC converter is a push-pull topology and the DC-AC inverter is a full-bridge
topology. The model circuit of the telecommunication inverter is depicted in Figure 1.

Figure 1. The basic circuit of the telecommunication inverter

The DC input current ripple can be reduced using the LC filters. There are two types of LC filters
used in this inverter they are DC-link Ltank and Ctank LC filter and Lload and Cload LC filter. The Lload and Cload
LC filter is mainly used to filter the AC output of the full-bridge inverter sinusoidal pulse width modulation
(PWM) spectrum. The Ltank and Ctank LC filters employ the DC input ripple reduction. In the active method,
the input ripple current shall be reduced by controlling the PWM of the DC-DC converter. Both ripple
reduction methods are implemented in a push-pull DC-DC converter and results are analysed [15]–[17]. The
push-pull converter has two switches metal–oxide–semiconductor field-effect transistor (MOSFETs) Q1 and
Q2 and at a time one switch is ON and the second switch is OFF and vice versa. The complementary PWM
signals are given to the metal–oxide–semiconductor (MOS) Q1, Q2, and it operates with a maximum of less
than 0.5 duty cycle. The push-pull converter operates with a switching frequency of 50 kHz. The transformer
provides galvanic isolation between high voltage circuits and low voltage circuits. The 48 V battery voltage
is step up to a 385 V DC by using this transformer. The transformer pulsated secondary voltage is rectified by
using rectifiers D1, D2, D3, and D4. The high-frequency switching diodes are used. This rectified DC link
voltage ripple is filtered by the inductor Ltank and capacitor Ctank. The DC link voltage is stored in the buffer
capacitor Ctank. This DC link voltage is converted to an AC supply with full-bridge insulated gate bipolar
transistors (IGBTs) by using sinusoidal PWM techniques. In (1) shows the push-pull converter voltage gain
function.

𝑁𝑠
𝑉𝑡𝑎𝑛𝑘 = ( ) 𝑉𝑖𝑛 𝑥 𝑑 (1)
𝑁𝑝

Optimization of inductors and capacitors filter components through active … (Jeyaraman Ramakrishnan)
300  ISSN: 2088-8694

Where,
𝑉𝑡𝑎𝑛𝑘 − 𝐷𝐶 𝑙𝑖𝑛𝑘 𝑣𝑜𝑙𝑡𝑎𝑔𝑒, 𝑉𝑖𝑛 − 𝐷𝑐 𝑖𝑛𝑝𝑢𝑡 𝑣𝑜𝑙𝑡𝑎𝑔𝑒 𝑓𝑟𝑜𝑚 𝐵𝑎𝑡𝑡𝑒𝑟𝑦 , 𝑁𝑝 −
𝑃𝑢𝑠ℎ 𝑝𝑢𝑙𝑙 𝑡𝑟𝑎𝑛𝑠𝑓𝑜𝑟𝑚𝑒𝑟 𝑃𝑟𝑖𝑚𝑎𝑟𝑦 𝑇𝑢𝑟𝑛𝑠, 𝑁s − 𝑃𝑢𝑠ℎ𝑝𝑢𝑙𝑙 𝑡𝑟𝑎𝑛𝑠𝑓𝑜𝑟𝑚𝑒𝑟 𝑆𝑒𝑐𝑜𝑛𝑑𝑎𝑟𝑦 𝑡𝑢𝑟𝑛𝑠, 𝑑 −
𝑃𝑊𝑀 𝐷𝑢𝑡𝑦 𝑐𝑦𝑐𝑙𝑒

3. ROOT CAUSE ANALYSIS OF THE RIPPLE GENERATION AND RIPPLE REDUCTION

METHODS
This sinusoidal waveform, which has a distinct fundamental frequency signature, powers the entire
power system. The power can be stored in the form of DC in a battery and converted to AC as needed to
power the critical backup loads. When the flat DC battery power is sliced and boosted to higher DC link
voltage by DC-DC converter and then it is again converted similar to an alternator output sine signature
waveform [18]–[20]. The source of the ripple waveform characteristics is theoretically analyzed and depicted
in Figure 2. In (4) and this product are in a sine pattern and have a frequency of double times of the
fundamental AC frequency.

Figure 2. Theoretical analysis of the ripple current waveform analysis

The DC to AC converter converts DC-link power to the instantaneous AC output power of the
inverter in a sinusoidal pattern and, due to the sinusoidal pattern, the DC link current has a second-order
ripple [21]. This is also injected through the DC-DC converter transformer and the DC input battery current
has the same sinusoidal ripple in the DC input current.

𝑣(𝑡) = 𝑉𝑚 𝑠𝑖𝑛 𝜔𝑡 (2)

i (t) = 𝐼𝑚 𝑠𝑖𝑛 𝜔𝑡 (3)

𝑝𝑙𝑜𝑎𝑑 = 𝑉𝑚 𝑠𝑖𝑛 𝜔𝑡 𝑥 𝐼𝑚 𝑠𝑖𝑛 𝜔𝑡 (4)

𝑝𝑙𝑜𝑎𝑑 = 𝑉𝑚 𝐼𝑚 𝑠𝑖𝑛2 𝜔𝑡 (5)

𝑝𝑙𝑜𝑎𝑑 𝑉𝑚 𝐼𝑚 (1−𝑐𝑜𝑠2𝜔𝑡)
𝑝𝑖𝑛 = = (6)
 2

𝑉𝑚 𝐼𝑚 (1−𝑐𝑜𝑠2𝜔𝑡)
𝐼𝑖𝑛 = (7)
 𝑥2 𝑉𝑖𝑛

Where,
𝜔 = 2𝑓 ACoutput frequency,

Int J Pow Elec & Dri Syst, Vol. 13, No. 1, March 2022: 298-308
Int J Pow Elec & Dri Syst ISSN: 2088-8694  301

𝑉𝑚, 𝐼𝑚 – Peak voltage, current of the load.

𝑉𝑖𝑛 – input DC voltage, 𝐼𝑖𝑛 – Input DC current.
𝑃𝑙𝑜𝑎𝑑 – Load power, 𝑃𝑖𝑛 – Input DC power.
The instantaneous AC output power of the inverter is in a sinusoidal pattern and this power
consumes from the DC-link intermediate section. Due to the sinusoidal pattern, DC-link tank capacitor power
reflects the same sinusoidal in DC link current [22].

3.1. Inductors and capacitors filter passive method direct current input ripple reduction
This LC filter ripple reduction is done with a passive element like a capacitor and an inductor.
Normally, the fixed PWM duty cycle is preferred for the LC filter ripple reduction. The DC link voltage is
more sensitive to the PWM duty cycle of the DC-DC converter. If the DC-DC converter operates with a
closed loop of DC-link voltage feedback, it has a lot of perturbation in the DC input battery current during
the correction of the DC-link voltage using the PWM duty cycle it may need large size of DC input filter. To
optimize the cost and complexity of the DC-link control and filtering, it has been designed with open-loop
control and high voltage DC-link capacitor being used in industrial practice. Hence, this method is employed
with an open-loop control of the DC-DC converter. The Vref setpoint is fixed to a maximum duty cycle of
the PWM.
By enhancing either inductor Ltank or capacitor of the Ctank in the DC link stage of the filter helps to
block the injection of the second-order frequency to the DC input side. This method is called the LC filter
method. The LC filter ripple reduction control diagram is depicted in Figure 3. Incrementing the Ltank
inductor value will cause an increase in the loss of the Ltank filter. Hence, in the LC filter, the capacitor Ctank is
increased for this experiment to evaluate the DC input ripple. An increment in the Ctank capacitor value will
help to reduce the DC input ripple current. This can be implemented by adding a filter capacitor in the DC
link section. The LC ripple-reduction is simulated in a push-pull DC-DC converter. This is a conventional
method and is also called a passive method.

PWM DC-DC
Controller Converter
Vref
DC link Voltage

Figure 3. LC filter method open-loop control diagram

3.2. The active method of direct current input ripple reduction

In the active method, the push-pull converter triggering pulses are controlled by a closed-loop
operation with the feedback of the DC link voltage and current. The Active ripple-reduction technique
closed-loop control diagram is depicted in Figure 4. The Vref is compared with DC-link voltage feedback
and the difference is given to the PI stage1. This PI stage 1 output is the set point of the current reference.
This is compared with the push-pull converter transformer for secondary current feedback. This summer
output is given to the PI stage 2. The PI stage 2 output is directly given to the PWM controller to generate the
PWM pulses for the push-pull converter. This dual feedback control implementation helps to reduce the
second-order fundamental frequency ripple in the DC input [23], [24]. This compensation technique will help
to balance the DC-link capacitor energy and maintain the average level of energy consumption from the DC
input sources and the DC input current ripples are controlled. The percentage of ripple will be calculated with
a formula of:

𝐼𝑟𝑚𝑠 𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝑡ℎ𝑒 𝑟𝑖𝑝𝑝𝑙𝑒 𝑎𝑐 𝑐𝑜𝑚𝑝𝑜𝑛𝑒𝑛𝑡𝑠

𝑝𝑒𝑟𝑐𝑒𝑛𝑡𝑎𝑔𝑒 𝑜𝑓 𝐼 𝑟𝑖𝑝𝑝𝑙𝑒 = 𝑥 100 (8)
𝑎𝑣𝑎𝑟𝑎𝑔𝑒 𝑑𝑐 𝑐𝑢𝑟𝑟𝑒𝑛𝑡

It is elaborated further,

√𝐼𝑟𝑚𝑠 2 −𝐼𝑖𝑛2
𝑃𝑒𝑟𝑐𝑒𝑡𝑎𝑔𝑒 𝑜𝑓 𝐼 𝑟𝑖𝑝𝑝𝑙𝑒 = 𝑥 100 (9)
𝐼𝑖𝑛

By using the active ripple current reduction method, with an optimal value of DC-link capacitor, the DC
input current ripple-reduction can be achieved. The DC input power consumption is maintained based on the
AC output power consumption and the DC input current ripple is controlled [25].

Optimization of inductors and capacitors filter components through active … (Jeyaraman Ramakrishnan)
302  ISSN: 2088-8694

PI PI PWM DC-DC
+ Stage 1 Stage 2 Controller Converter
Vref DC link
Voltage

Iint

Vint

Figure 4. Active method ripple-reduction closed-loop control diagram

4. SIMULATION ANALYSIS
4.1. Inductors and capacitors filter method direct current input ripple reduction
The LC filter method has been verified in a dual-stage inverter simulated model with a help of power
simulator (PSIM) simulation tool software. The LC filter PSIM simulation model is depicted in the Figure 5.
It is a combination of a push-pull converter and a full-bridge DC-AC converter. The DC input is a 48 V
battery source and it is connected to the push-pull converter. The PSIM simulation components parameter
details are shown in Table 1.

Figure 5. Simulation model of LC filter method

Table 1. Simulation model PSIM parameter

Parameters Values
Input DC voltage (𝑽𝒊𝒏) 48 V
Transformer ratio (𝑵𝒑: 𝑵𝒔) 1:10
Forward inductor (𝑳𝒕𝒂𝒏𝒌) 1.5 mH
DC tank voltage (𝑽𝒕𝒂𝒏𝒌) 385 V
Switch frequency of DC-DC (𝒇𝒔) 50 kHz
Switch frequency of DC-AC (𝒇𝒔) 19.5 kHz
Power rating (𝑾) 0.8 kW
R Load 66 ohm

The push-pull converter output is rectified by the diode full bridge and the rectified DC link voltage
is 385 V. It is connected to the DC-AC IGBT full-bridge inverter section and it converts 385 V to 230 V AC
50 Hz supply. The AC output is connected to the 800 W Load. The inverter ripple analysis is started with the
minimum value of the Ctank capacitor with 180 µF capacitor at 48 V DC input. Figure 6 waveform shows the
response of the 180 µF capacitor. The DC input average current of 17.4 A and the ripple current of 10.6 A is
observed and the percentage of ripple is 60.8%. Figure 7 waveform shows the response of the Ctank with a
720 µF capacitor. In the DC input average current of 17.7 A, the ripple current of 3.51 A is observed and the
percentage of ripple is 19.8%.

Int J Pow Elec & Dri Syst, Vol. 13, No. 1, March 2022: 298-308
Int J Pow Elec & Dri Syst ISSN: 2088-8694  303

Figure 6. LC filter method DC input current ripple Figure 7. LC filter method DC input current ripple
with a 180 µF DC link capacitor with 720 µF DC link capacitor

DC link capacitor at 4,140 µF ripple current reduces to 0.66 A. The percentage of ripple reduces to
less than 3.7% and it meets the standard psophometric noise test passing limits. In the simulation, the DC link
Ctank capacitor value varied from 180 µF to 4,140 µF and the DC input ripple current and percentage of ripple
are tabulated. By increasing the Ctank capacitor value, the DC input current ripple amplitude is gradually
reduced. The graph with filter capacitance vs percentage of input ripple current and DC input current ripple
amplitude is depicted in Figure 8. This passive method is used in the conventional model industrial telecom
inverters to meet the telecom standard requirement.

70.0
60.0
180 ; 60.8
50.0
% of ripple

40.0 % of Ripple Current vs Ctank

30.0
20.0 4,140 ; 3.7
10.0
0.0
- 1,000 2,000 3,000 4,000 5,000
Capacitance in µF

Figure 8. LC filter passive method: DC-link capacitance vs % of DC input current ripple

4.2. The active method of direct current input ripple reduction

The active method of the closed-loop DC ripple reduction method has been simulated in the PSIM
tool and simulation model is depicted in the Figure 9. The active DC ripple reduction has a dual closed-loop
operation, the DC link voltage and current are taken as feedback and given to the closed-loop PI system. The
first PI stage closed-loop has compared with the fixed Vref of 2.5 V reference and the feedback of DC link
voltage and then the error voltage is processed in PI stage1 and this output has compared with the feedback
of transformer secondary current and then the error is processed in the second PI stage is a current closed
loop.
By this active method of ripple reduction techniques, the less than 5% of DC ripple is achieved with
720 µF filter capacitors. At 48 V DC input, the DC input ripple current has simulated and the AC output
voltage, current and DC input voltage, current waveform response is shown in Figure 10. Due to the narrow
bandwidth of the 100 Hz ripple reduction of active filter method at steady-state condition, the input current of
inverter begets 3.2% DC ripple current and 0.62 A of ripple current with 720 µF capacitor and same
operating condition of the passive method at 48 Vdc input and 800 W load. From the summary of the
simulation study and comparison analysis results, the active filter DC ripple reduction is more effective than
the LC filter method. The test result analysis and comparisons are depicted in Figure 11.

Optimization of inductors and capacitors filter components through active … (Jeyaraman Ramakrishnan)
304  ISSN: 2088-8694

Figure 9. Simulation model of active filter method

70.0 60.8
60.0 LC filter method (%)
% of current ripples

50.0 Active filter method (%)

40.0
30.0
19.8
20.0
10.0 3.2 3.7
0.0
180uF 720uF 4,140uF
Filter capacitance (µF)

Figure 10. Active method: DC input current ripple Figure 11. Simulation results of % ripple current
with a 720 µF DC link capacitor comparison between LC filter method and active filter
method

Based on the comparison between the passive and active methods, with the same value of the DC
link capacitor, the current ripple is greater with the passive method and less with active methods. The passive
method is 19.8% (3.51 A), but it is reduced to 3.2% (0.62 A) by the active method. For the battery, the
accepted ripple current limit is less than 5% and also meets the psophometric norm.

5. EXPERIMENTAL RESULTS
The LC filter and active filter ripple control methods have been experimented with and verified in
the dual-stage 48 V/230 V, 1 kVA telecom industrial inverter. The experimental test setup of industrial
inverter and ripple waveform capture set up has been shown in the Figure 12. The first stage converter is a
push-pull-forward converter is designed to operate with the nominal voltage of 48 V and the second stage is a
full-bridge inverter stage. The telecom site system is set up with a 48 Vdc operational voltage, and it draws
power from the battery, which is connected to a battery charger in parallel. The standard working range of a
telecom inverter is 40 to 60 Vdc. The push-pull-forward converter is taking power from the battery and it is
switched to 50 kHz. The push-pull converter PWM generation has designed with a UC 2525 PWM controller

Int J Pow Elec & Dri Syst, Vol. 13, No. 1, March 2022: 298-308
Int J Pow Elec & Dri Syst ISSN: 2088-8694  305

and operational amplifier LM2903. These PWM pulses are given to the MOSFET drivers and it is triggering
the MOSFET Q1 and Q2. The DC-DC converter transformer is designed with ETD 54 N87 core and the
transformation ratio of 1:10 and the transformer secondary pulsated output voltage is rectified with D1 to D4
fast diode full-bridge rectifier. The rectified DC link voltage is filtered with an LC filter. The DC link voltage
of 385 V is converted to the AC sinewave by sine PWM techniques. The Atmel 328 microcontroller is used
to control the second stage of the full-bridge inverter. The AC sine wave output is generated by using the
level 3 sinusoidal PWM technique. Its switching frequency is 19.5 kHz. The second stage full-bridge inverter
designed with IGBT T1 to T4. The output of the full-bridge pulsated sine output is filtered by the differential
LC filter and its inductor values are 1.2 mH and capacitance is 4.7 µF/275 V. By this above process, the
DC-link 385 V voltage is converter to AC 230 V/50 Hz supply and it supplies to the critical telecom load
backup application. In the experimental prototype hardware components, part number details are shown in
Table 2.

Figure 12. The experimental test set of dual-stage industrial telecom inverter

Table 2. Prototype major component parts list

Parameters Part Numbers
Microcontroller ATMEGA328
PWM Controller U1 TI UC2525
IGBT T1 to T4 ST STGP19N60KD -600 V, 20 A
MOSFET Q1, Q2 IR IRFP4127PBF -200 V, 70 A
Operational Amplifier TI LM2903
Diode D1 to D4 Fairchild MUR880 – 800 V, 8 A

5.1. Inductors and capacitors filter direct current ripple reduction method
The LC filter method and active filter DC ripple reduction methods are implemented in the industrial
inverter 48 V/230 V, 0.8 kW telecom inverter and the experimental test results are analyzed. The LC filter
method has experiment, the push-pull converter control circuit is modified to generate a full pulse width
constant PWM pulses for the UC2525 controller. Since it is operated with an open-loop system, the DC-DC
converter is operated up to 43 V input supply and the DC link voltage is maintained less than the maximum
operation voltage of the DC link capacitors (450 V). The LC filter method experimental analysis is done at
43 V input supply and waveforms are captures and analyzed. In the LC filter method, the DC input current
has a ripple peak to peak of 16 A peak to peak and it has 40 A overall peak current at 800 W resistive load
and average current of 26.2 A and the percentage of ripple 21.6%. With an 180 µF capacitor, the DC input
current waveform is shown in Figure 13.
With a 720 µF (4x180 µF) C1 DC link capacitor, at 43 V voltage operation, the DC input current has
a ripple peak to peak of 10 A peak to peak and it has 25 A overall peak current at 800 W resistive load and
average current of 21.2 A and the percentage of ripple 16.7%. With a 720 µF capacitor, the DC input current
waveform is shown in Figure 14. The DC input current has a ripple peak to peak of 1.4 A peak to peak and it
has 22 A overall peak current at 800 W resistive load and average current of 21.6 A and the percentage of
ripple 2.3%. With a 4,140 µF capacitor, the DC input current waveform is shown in Figure 15.

Optimization of inductors and capacitors filter components through active … (Jeyaraman Ramakrishnan)
306  ISSN: 2088-8694

Figure 13. LC filter method: DC input current ripple Figure 14. LC filter method: DC input current ripple
with a 180 µF DC link capacitor with a 720 µF DC link capacitor

Figure 15. LC filter method: DC input current ripple with a 4,140 µF DC link capacitor

5.2. The active method of direct current ripple reduction control

The active method is operating with closed-loop control of operating DC-DC converter system. It
gives dual benefits one helps to maintain the DC link voltage at 385 V and the second help to average power
consumption for the DC link section. With the active method, the DC input current waveform is shown in
Figure 16.

Figure 16. Active filter method: DC input current ripple with a 720 µF DC link capacitor

With 720 µF DC-link capacitance, the input DC current has a ripple of 1.2 A peak to peak and 22.0
A overall peak current at 800 W resistive load. The average current of 21.5 A and the percentage of ripple
1.97%. It features a low amount of AC components and a flat DC input current. The inverter size is reduced,
the weight is reduced, and the cost is reduced using this technology.
It contributes to the battery's improved performance and longevity. This ripple reduction also aids
efficient power consumption and enhances the inverter's backup duration. From the summary of the test
results, the active filter DC ripple-reduction is more effective than the LC filter method and the test results

Int J Pow Elec & Dri Syst, Vol. 13, No. 1, March 2022: 298-308
Int J Pow Elec & Dri Syst ISSN: 2088-8694  307

comparisons are depicted in Figure 17. The psophometric noise test (ETSI300132-2) measures battery ripple
voltage with a standard test procedure and the battery ripple should be lower the 2 mV. The emission of noise
is measured with decibel. The passing limit is less than -51 db. This test was conducted in the operating range
of 40 to 60 V DC input. With the active method, the noise emission reduces and it passes the limits of the
standard.

50.0
43.2
% of current ripples
LC filter method (%)
40.0

30.0
Active filter method (%)

20.0 16.7

10.0
1.97 2.3
0.0
180uF 720uF 4,140uF
Filter capacitance (µF)

Figure 17. Experimental results of % ripple current comparison between LC filter method and active filter
method

6. CONCLUSION
The low-frequency ripple-reduction active and passive methods has been tested within the
telecommunication working with 48 V DC supply and the output of AC 230 V with a power rating of 1
kVA/800 W. The root cause of ripple generation is being investigated. In this paper, the push-pull DC-DC
converter DC input ripple-reduction characteristics are investigated by the LC filter method and the active
dual PI closed-loop control method. The two different ripple-reduction methods' results are compared. In the
LC method, which is evaluated by increasing the Ctank capacitance value, the DC input current ripple-
reduction is observed and, with a 4,140 µF capacitance value, the ripple current is reduced to less than 5%
and the DC input supply ripple voltage is less than 2 mV (-51 dB). It also complies with the psophometric
norm ETSI300132-2. The DC input ripple is reduced in the active ripple-reduction method to a small value of
Ctank capacitance of 720 µF, and with this value, the 5% DC input ripple psophometric norm requirement is
met. The LC method requires more space to accommodate the large capacitance and is more expensive than
the active dual PI ripple-reduction method.

ACKNOWLEDGEMENTS
This material is based upon work supported by the R&D Department of CET Power Solutions India
Pvt Ltd, Chennai.

REFERENCES
[1] L. Zhang, X. Ruan, and X. Ren, “Second-Harmonic Current Reduction for Two-Stage Inverter with Boost-Derived Front-End
Converter: Control Schemes and Design Considerations,” in IEEE Transactions on Power Electronics, vol. 33, no. 7, pp. 6361–
6378, July 2018, doi: 10.1109/TPEL.2017.2746878.
[2] N. Lu, S. Yang, and Y. Tang, “Ripple Current Reduction for Fuel-Cell-Powered Single-Phase Uninterruptible Power Supplies,” in
IEEE Transactions on Industrial Electronics, vol. 64, no. 8, pp. 6607–6617, Aug. 2017, doi: 10.1109/TIE.2017.2677329.
[3] Z. Wei, X. Deng, C. Gong, J. Chen, and F. Zhang, “A novel technique of low frequency input current ripple reduction in two-
stage DC-AC inverter,” IECON 2012-38th Annual Conference on IEEE Industrial Electronics Society, 2012, pp. 139–143, doi:
10.1109/IECON.2012.6388818.
[4] A. H. El Khateb, N. A. Rahim, J. Selvaraj, and B. W. Williams, “DC-to-DC Converter with Low Input Current Ripple for
Maximum Photovoltaic Power Extraction,” in IEEE Transactions on Industrial Electronics, vol. 62, no. 4, pp. 2246–2256, April
2015, doi: 10.1109/TIE.2014.2383999.
[5] Hong-Ju Jung and Rae-Young Kim, “A Second-order Harmonic Current Reduction with a Fast Dynamic Response for a Two-
Stage Single-phase Grid–connected Inverter,” Journal Electrical Engineering and Technology, vol. 9, no. 6, pp. 1988–1994,
2014, doi10.5370/JEET.2014.9.6.1988.
[6] Y. Shi, B. Liu, and S. Duan, “Low-Frequency Input Current Ripple Reduction Based on Load Current Feedforward in a Two-Stage
Single-Phase Inverter,” in IEEE Trans. on Power Elec., vol. 31, no. 11, pp. 7972–7985, 2016, doi: 10.1109/TPEL.2015.2513065.
[7] H. Deng, Q. Li, Z. Liu, L. Li, and W. Chen, “Low frequency current ripple mitigation of two stage three–phase PEMFC
generation systems,” Journal of Power Electronics, vol. 16, no. 6, pp. 2243–2257, 2016, doi: 10.6113/JPE.2016.16.6.2243.
[8] H. G. Jeong, G. S. Kim, and K. B. Le, “Second order harmonic reduction technique for photovoltaic power conditioning systems
using a proportional-resonant controller,” Energies, vol. 6, no. 1, pp. 79–96, 201, 2013, doi: 10.3390/en6010079.
Optimization of inductors and capacitors filter components through active … (Jeyaraman Ramakrishnan)
308  ISSN: 2088-8694

[9] A. A. Ahmad and A. Abbrishamafar, “Low Frequency current ripple reduction in Two stage power converter with a PI Controller
and Non Linear Gain,” Iranian Journal of electrical and computer engineering, vol. 11, no. 2, pp. 102–108, 2012.
[10] R. Jeyaraman and C. N. Ravi, “Low-frequency DC Input current ripple Reduction in a Two-stage Single-Phase Inverter using DC
Input Voltage control,” 2020 6th International Conference on Advanced Computing and Communication Systems (ICACCS),
2020, pp. 342–347, doi: 10.1109/ICACCS48705.2020.9074204.
[11] J. Wang, K. Zou, C. Chen, and L. Chen, “A high frequency battery model for current ripple analysis,” 2010 Twenty-Fifth Annual
IEEE Applied Power Electronics Conference and Exposition (APEC), 2010, pp. 676–680, doi: 10.1109/APEC.2010.5433598.
[12] A. D. Kupchinov, N. N. Smotrov, and Y. P. Gusev, “The compatibility verification of batteries and battery chargers to ripples in
battery float charge mode,” 2018 IEEE Conference of Russian Young Researchers in Electrical and Electronic Engineering
(EIConRus), 2018, pp. 696–700, doi: 10.1109/EIConRus.2018.8317189.
[13] A. D. Kupchinov, Y. P. Gusev, and Y. V. Monakov, “Batteries current ripples in float charge mode,” 2017 IEEE Conference of
Russian Young Researchers in Electrical and Electronic Engineering, 2017, pp. 913–915, doi: 10.1109/EIConRus.2017.7910704.
[14] K. Gupta, M. Sharma, and D. Verma, “Design and Fabrication of Two Stage 2 KVA Inverter using Push-Pull Topology,” 2020
International Conf. on Communication and Signal Processing, 2020, pp. 1027–1031, doi: 10.1109/ICCSP48568.2020.9182234.
[15] B. G. Mayuri and U. S. Swati, “Design of residential photovoltaic power system using current-fed push-pull front end converter,”
2017 IEEE Inter. Conf. on Power, Control, Signals and Instr. Engin, 2017, pp. 354–357, doi: 10.1109/ICPCSI.2017.8392314.
[16] W. Meesrisuk, N. Sarasiri, and A. Jangwanitlert, “Ripple current reduction using interleaving technique for three-level ZVZCS
DC-DC converter,” 2016 IEEE Region 10 Conference (TENCON), 2016, pp. 837–841, doi: 10.1109/TENCON.2016.7848122.
[17] L. An and D. D. Lu, “Analysis of DC Bus Capacitor Current Ripple Reduction in Basic DC/DC Cascaded Two-Stage Power
Converters,” in IEEE Trans. on Industrial Electr., vol. 63, no. 12, pp. 7467–7477, Dec. 2016, doi: 10.1109/TIE.2016.2594162.
[18] M. Schuck and R. C. N. Pilawa-Podgurski, “Input current ripple reduction through interleaving in single-supply multiple-output dc-dc
converters,” 2013 IEEE 14th Workshop on Contr. & Model for Power Electr., 2013, pp. 1–5, doi: 10.1109/COMPEL.2013.6626436.
[19] J. Lee, Y. Jeong, and B. Han, “A Two-Stage Isolated/Bidirectional DC/DC Converter with Current Ripple Reduction Technique,”
in IEEE Transactions on Industrial Electronics, vol. 59, no. 1, pp. 644–646, Jan. 2012, doi: 10.1109/TIE.2011.2109343.
[20] M. N. Musarrat, M. R. Islam, K. M. Muttaqi and D. Sutanto, “Shunt Active DC Filter to Reduce the DC-Link Ripple Current
Caused by Power Converters to Improve the Lifetime of Aluminum Electrolytic Capacitors,” in IEEE Transactions on Industry
Applications, vol. 57, no. 4, pp. 4306–4315, July–Aug. 2021, doi: 10.1109/TIA.2021.3083636.
[21] W. Lee, B. Han, and H. Cha, “Battery ripple current reduction in a three-phase interleaved dc-dc converter for 5 kW battery
charger,” 2011 IEEE Energy Conversion Congress and Exposition, 2011, pp. 3535–3540, doi: 10.1109/ECCE.2011.6064247.
[22] A. K. Singh, M. K. Pathak, and Y. S. Rao, “A new two-stage converter with reduction of DC-link capacitor for plug-in electric vehicle
battery charger,” 2017 3rd Inter. Conf. on Comp. Intell. & Commun. Tech., 2017, pp. 1–6, doi: 10.1109/CIACT.2017.7977338.
[23] J.-H. Park, H.-G. Jeong, and K.-B. Lee, “Output Current Ripple Reduction Algorithms for Home Energy Storage Systems,”
Energies, vol. 6, no. 10, pp. 5552–5569, 2013, doi: 10.3390/en6105552.
[24] H. Wu, V. Pickert, S. Lambert, P. Allanf, X. Deng and H. Zhan, “A ripple reduction method for a two stages battery charger with
multi-winding transformer using notch filter,” 2017 IEEE 12th International Conference on Power Electronics and Drive Systems
(PEDS), 2017, pp. 1,101-1,106, doi: 10.1109/PEDS.2017.8289265.
[25] C. Bai et al., “Novel input current ripple compensation technique for capacitor-less dual-half-bridge converter,” 2017 IEEE
Applied Power Electronics Conference and Exposition (APEC), 2017, pp. 2015–2021, doi: 10.1109/APEC.2017.7930976.

BIOGRAPHIES OF AUTHORS

Jeyaraman Ramakrishnan was born in Aruppukottai Tamil Nadu in 1982. He

received his B.E degree in Electrical and Electronics Engineering from Madurai Kamaraj
University, Madurai, India, in 2004 and his Master degree in Power electronics from P.S.G
College of Technology Coimbatore, India, in 2007. Currently he is working as a senior
manager-R&D in CET Power Solutions India Pvt Ltd. Chennai and he is purchasing towards
his Ph.D degree in Sathyabama Institute of Science and Technology, Chennai, India. He has 15
years of industrial experience. His current research interests include power electronics and its
applications such as optimization of the converter and inverter, renewable energy system,
reliability and telecommunication inverter safety compliance with safety, EMC standards. He
can be contacted at email: [email protected].

Chengalvarayan Natarajan Ravi was born in Chennai, Tamil Nadu, in 1978.

He received Bachelor of Engineering degree in Electrical and Electronics Engineering in the
year 1999 from Crescent Engineering College, Chennai. Master of Engineering degree in
Power Systems in the year 2006 from B.S.A.R Crescent Engineering College, and Ph.D in
power system optimization techniques from Sathyabama University, Chennai, Tamilnadu,
India at present he is working as professor in Vidya Jyothi Institute Technology, Hyderabad,
Telangana State, India. He has 15 years of teaching and 5 years of industrial experience. He
received best teacher award in the year 2019 from ESN awards. He has guided several projects
in the areas Power Electronics, Power Systems, and Electric Drives. His area of interest is
power system optimization, FACTS, power electronics and renewable energy system. He can
be contacted at email: [email protected].

Int J Pow Elec & Dri Syst, Vol. 13, No. 1, March 2022: 298-308