1 tayangan

Diunggah oleh Đạt Nguyễn Thái

ennergies solar

- Grid Connected PV System with CSI
- FLXPRO17-DEN1425W
- Final Ashish ABSTRACT 1
- 200 Icrera2013 Spain
- 1KW DCC-AC Inverter
- A SOLAR POWER PRODUCTION SYSTEM USING MPPT WITH GRID SYNCHRONIZATION FOR DISTRIBUTED POWER PRODUCTION AND DISTRIBUTION
- VSC Current Control
- A Modified Adaptive Hill Climbing Maximum Power Point Tracking (MPPT) Control Method for Photovoltaic Power Systems
- Arduino Solar Charge Controller Version 30
- Brochure-Energy-Storage.pdf
- My Published MPPT Paper RSER
- 53-759
- 9 Basic DC to DC Converter Circuits
- Report
- Flx pro 15 - 17
- Final - Ieee 3013 Power Elec. Project Titles
- Proposal for 7 HP solar pumping system
- ic5_eng
- Inverter DC to AC
- Inverter Toshiba VF-A7 Manual

Anda di halaman 1dari 25

Tracking in Grid-Connected Photovoltaic Systems

Paula Andrea Ortiz Valencia 1, * and Carlos Andres Ramos-Paja 2

Received: 11 September 2015 ; Accepted: 19 October 2015 ; Published: 2 November 2015

Academic Editor: Tapas Mallick

1 Instituto Tecnológico Metropolitano (ITM), Calle 73 No 76A-354 Vía al Volador,

Medellín 05001000, Colombia

2 Universidad Nacional de Colombia, Carrera 80 No 65-223-Block M8-113, Medellín 050041,

Colombia; caramosp@unal.edu.co

* Correspondence: paulaortiz@itm.edu.co; Tel.: +57-4581-3957 or +57-4460-0727 (ext. 5587)

Abstract: The maximum power point tracking (MPPT) of photovoltaic systems must be as fast and

accurate as possible to increase the power production, which eventually increases the PV system

profitability. This paper proposes and mathematically analyses a sliding-mode controller to provide

a fast and accurate maximum power point tracking in grid-connected photovoltaic systems using a

single control stage. This approach avoids the circular dependency in the design of classical cascade

controllers used to optimize the photovoltaic system operation, and at the same time, it reduces

the number of controllers and avoids the use of linearized models to provide global stability in all

the operation range. Such a compact solution also reduces the system cost and implementation

complexity. To ensure the stability of the proposed solution, detailed mathematical analyses

are performed to demonstrate the fulfillment of the transversality, reachability and equivalent

control conditions. Finally, the performance of the proposed solution is validated using detailed

simulations, executed in the power electronics simulator PSIM, accounting for both environmental

and load perturbations.

Keywords: DC/DC converter; maximum power point tracking (MPPT); photovoltaic systems;

sliding mode control; grid-connection

1. Introduction

The world energy demand is projected to more than double by 2050, and more than triple by the

end of the century [1]. Incremental improvements in existing energy networks will not be adequate

to supply this demand in a sustainable way. Hence, it is necessary to find sources of clean energy

with a wide distribution around the world. The energy generation with photovoltaic (PV) systems is

inexhaustible, hence it is a suitable candidate for a long-term, reliable and environmentally friendly

source of electricity. However, PV systems require specialized control algorithms to guarantee the

extraction of the maximum power available, otherwise the system could be unsustainable.

The PV generator, also known as PV array, produces DC power that depends on the

environmental conditions and operating point imposed by the load. To provide a high power

production, the PV system includes a DC/DC converter to isolate the operating point of the generator

(voltage and current) from the load, where such a power converter is regulated by an algorithm

that searches, online, the maximum power point (MPP, i.e., the optimal operation condition) known

as Maximum Power Point Tracking (MPPT) algorithm. The classical structure of a grid-connected

photovoltaic system is presented in Figure 1, in which the PV generator interacts with a DC/DC

converter controlled by a MPPT algorithm [2,3]. Such a structure enables the PV system to modify

Tracking (MPPT) algorithm. The classical structure of a grid-connected photovoltaic system is presented

in Figure 1, in which the PV generator interacts with a DC/DC converter controlled by a MPPT

algorithm [2,3]. Such a structure enables the PV system to modify the operation conditions in

agreement with the environmental circumstances (mainly changed by the irradiance and temperature)

Energies 2015, 8, 12363–12387

so that a maximum power production is achieved [4,5]. Figure 1 also illustrates the gird-connection

side of the PV system, which is formed by a DC-link (capacitor ) and a DC/AC converter (inverter).

the operation conditions in agreement with the environmental circumstances (mainly changed by

The inverter is controlled

the irradiance to follow

and temperature) so athat

required

a maximum power factor,

power provideissynchronization

production and protect

achieved [4,5]. Figure 1

against

also illustrates the gird-connection side of the PV system, which is formed by a DC-link (capacitor bulk

islanding, among others. Moreover, the inverter must to regulate the DC-link voltage at the

capacitor

Cb ) and, where

a DC/AC twoconverter

cases are(inverter).

possible:The first, the inverter

inverter regulates

is controlled the DC

to follow component

a required power of factor,

voltage,

provide synchronization and protect against islanding, among others. Moreover, the inverter must

but due to the sinusoidal power injection into the grid, voltage experiments a sinusoidal perturbation at

to regulate the DC-link voltage at the bulk capacitor Cb , where two cases are possible: first, the

twice inverter

the grid regulates

frequencythe andDCwith a magnitude

component of Cbinversely proportional

voltage, but due to thetosinusoidal

the capacitance [3]. In the

power injection intosecond

case, the

the grid,

DC component of voltage

Cb voltage experiments is not properly

a sinusoidal regulated,

perturbation at twicewhich produces

the grid multiple

frequency and withharmonic

a

components with amplitude inversely proportional to the capacitance [3]. In both cases, the DC/DC

magnitude inversely proportional to the capacitance [3]. In the second case, the DC component of

Cb voltage is not properly regulated, which produces multiple harmonic components with amplitude

converter output terminals are exposed to voltage perturbations that could be transferred to the PV

inversely proportional to the capacitance [3]. In both cases, the DC/DC converter output terminals

generator terminals,

are exposed thus perturbations

to voltage degrading the thatMPP

could tracking process.

be transferred to theConcerning

PV generator the DC/DC

terminals, thuspower

converter, the boost

degrading the MPP topology

tracking is the most

process. widelythe

Concerning used

DC/DC due power

to theconverter,

low voltage levels

the boost exhibited

topology is by

commercial

the most PV modules

widely used[6].

due to the low voltage levels exhibited by commercial PV modules [6].

Figure 1. Typical structure of a photovoltaic (PV) system.

Multiple types of MPPT solutions are reported in the literature, which differ in complexity,

number of sensors needed for operation, convergence speed, cost-effective range, etc. [7]. Two of

the most commonly used MPPT techniques are Perturb and Observe (P & O) and Incremental

Conductance (IC); the reason for this popularity is its implementation simplicity and its relatively

good performance [4,8]. Other MPPT techniques are based on using fractional values of the open

circuit voltage and short circuit current, i.e., the Fractional Open Circuit Voltage and the Fractional

Short-Circuit Current. The main advantages of those solutions are the low cost and implementation

simplicity since they only require a single (voltage or current) sensor [9,10]; But their efficiency is

low compared with the P & O and IC algorithms. In contrasts, techniques based on computational

intelligence, such as neural networks and fuzzy logic, offer speed and efficiency in tracking the

MPP [11–13]; however its complexity and implementation costs are high compared with the P & O

and IC algorithms, which make them costly solutions.

The main problem of using traditional MPPT algorithms acting on the duty cycle of the DC/DC

converter associated to the PV generator, i.e., Figure 1, concerns the large disturbances caused by

irradiance transients in the system operating point, which generates a slow tracking of the MPP.

This condition is also present at the system start-up, in which the MPPT algorithm takes a large

amount of time to reach the MPP [3]. To mitigate such disturbances and speed-up the MPPT

procedure, a two-stage control structure is usually adopted: it generally involves an algorithmic

MPPT controller in cascade with a conventional voltage regulator (e.g., based on lineal or nonlinear

control) as depicted in Figure 2. Moreover, such a structure is also needed to increase the reliability

of double-stage grid-connected PV systems: the sinusoidal oscillation on the DC-link caused by the

inverter operation must be mitigated, otherwise the MPPT procedure could be inefficient as reported

in [3]. Such mitigation is traditionally performed by using large electrolytic capacitors for Cb , however

the electrolytic technology introduces reliability problems due to its high failure rate [14]. Then, the

12364

Figure 2. Moreover, such a structure is also needed to increase the reliability of double-stage

grid-connected PV systems: the sinusoidal oscillation on the DC-link caused by the inverter operation

must be mitigated, otherwise the MPPT procedure could be inefficient as reported in [3]. Such

mitigation

Energies is traditionally

2015, 8, 12363–12387 performed by using large electrolytic capacitors for Cb, however the

electrolytic technology introduces reliability problems due to its high failure rate [14]. Then, the

voltage regulator

voltage in Figure

regulator 2 enables

in Figure 2 enablestotomitigate

mitigate the voltageoscillations

the voltage oscillations in the

in the DC-link

DC-link produced

produced by by

small non-electrolytic capacitors, thus improving the system reliability [3,14].

small non-electrolytic capacitors, thus improving the system reliability [3,14].

Figure 2. Figure

Structure

2. of a PV system

Structure with

of a PV a classical

system with apulse width

classical modulation

pulse (PWM)-based

width modulation cascade control.

(PWM)-based

cascade control.

However, the design process of such a voltage regulator depends on the MPPT parameters, while

However, the design process of such a voltage regulator depends on the MPPT parameters, while

the MPPT controller stability also depends on the cascade controller performance [3]: for example, to

the MPPT controller stability also depends on the cascade controller performance [3]: for example, to

ensureensure

the stability of aofPa&

the stability P&OOalgorithm,

algorithm, its perturbationperiod

its perturbation periodmust

must be be larger

larger thanthan the settling

the settling time time

of theofPVthevoltage, which

PV voltage, depends

which depends ononthe

thesystem

system operating pointand

operating point andononthethe cascade

cascade voltage

voltage controller.

controller.

However, the voltage controller design requires performance criteria, e.g., the settling

However, the voltage controller design requires performance criteria, e.g., the settling time as time as in [3],in [3],

which are usually imposed in terms of the P & O parameters. Such a circular dependency in the two

controllers design makes difficult to guarantee both system stability and desired performance in all

the operating conditions.

When the voltage regulator is implemented using conventional linear control techniques it is

necessary to linearize the system model around a given operation point, which is usually the MPP at

some irradiance condition [3]. However, due to the nonlinear nature of the PV module and DC/DC

converter, the performance (and even stability) of the linear controller is limited to the neighborhood

around the MPP [15]. This constraint puts at risk the system performance since the operating

point changes with the unpredictable and unavoidable environmental perturbations. To address this

problem, the work in [16] uses a sliding mode controller (SMC) to regulate the inductor current of

a boost converter associated to the PV module, which enables to guarantee global system stability

at any operating point. The solution proposed in that work considers three controllers in cascade

as follows: the SMC that generates the activation signal for the MOSFET, a PI controller designed

to provide the SMC reference depending on the command provided by a P & O algorithm, which

is in charge of optimizing the power. However, the design of the PI controller requires a linearized

model of the system around the MPP, hence it cannot guarantee the same performance in all the range

of operation. In fact, a wrong design of such a PI controller could make the P & O unstable, hence

both PI and P & O controllers have a circular dependency on their parameters. Similarly, the work

in [17] uses a SMC to regulate the input capacitor current of the boost converter. This solution has

a major advantage over the work reported in [16]: the solution in [17] does not require a linearized

model since the transfer function between the capacitor current and voltage is linear and it does not

depend on the irradiance or temperature conditions. Therefore, such a solution is able to guarantee

the desired performance in all the operating range. However, as in the previous work, the three

controllers are designed separately, which makes difficult to perform the system design: again, a

wrong P & O perturbation period could lead to an unstable system operation.

To avoid the circular dependency between voltage controllers and MPPT algorithms, a single

controller in charge of both MPPT and voltage control operations is required. This problem has

been addressed in [18–21] by using the sliding mode control technique. However, those works

12365

Energies 2015, 8, 12363–12387

are based on two considerations difficult to apply to grid-connected PV systems: First, there is not

considered a capacitor linking the PV array and the DC/DC converter; second, the load impedance

is considered constant.

Regarding the first consideration, PV systems commonly consider a capacitor between the PV

source and the DC/DC converter to stabilize the PV voltage, which in turns stabilizes the power

produced, otherwise the current ripple at the DC/DC converter input will produce undesired

oscillations that will degrade the MPPT procedure. The worst case of that current ripple is exhibited

by DC/DC converters with discontinuous input current, e.g., buck or buck-boost topologies, while

DC/DC converters with continuous input current only inject the inductor current ripple, e.g., boost,

Sepic or Cuk topologies. In any case, a capacitor in parallel with the PV source is almost always

considered; in fact, the experimental scheme of [21] includes such a capacitor despite it is not taken

into account for the SMC analysis. It must be noted that this capacitor defines the dynamic behavior

of the PV voltage and power, therefore it must be considered in the PV controller design.

Regarding the second consideration, which assumes constant the load impedance, it is not

applicable to grid-connected inverters since, as reported in [22], such inverters are controlled to

provide a constant average value in the DC-link, with sinusoidal oscillations at twice the grid

frequency due to the injection of single-phase AC power. Therefore, the best representation of such a

load (from the DC/DC converter side) is a voltage source and not a constant impedance. This model

will be further justified in Section 2. Therefore, in the design of a PV controller for grid-connected

applications it is desirable to account for loads with non-constant impedance.

Other interesting sliding-mode controller for PV systems is presented in [23], it considering a

buck converter. Despite this work considers the capacitor linking the PV source and the power stage,

it is not applicable to grid-connected PV systems since the DC/DC converter provides an output

voltage lower than the one provided by the PV source. Such a condition is incompatible with the

high input-voltage required by classical buck inverters for grid-connection [22,24]. Moreover, this

work also assumes constant the load impedance, and it does not analyze the switching frequency,

which is an important parameter for implementation.

On the basis of those previous works and considerations, this paper presents the analysis and

design of a SMC aimed at performing a fast MPPT action on grid-connected PV systems using a single

control stage. This approach avoids the circular dependency among cascade controllers, reduces the

number of controllers and avoids the use of linearized models to provide global stability in all the

operation range. In such a way, even the use of a PWM is not needed. Such a compact design also

reduces the system cost and complexity.

The paper is organized as follows: Section 2 presents the non-linear mathematical model

representing the PV system, then Section 3 introduces the proposed sliding surface and the controller

structure. Section 4 presents the mathematical analysis of the transversality, reachability and

equivalent control conditions to demonstrate the global stability of the proposed controller. Finally,

Section 5 illustrates the performance of the proposed solution using detailed simulations executed in

a standard power electronics simulator. The conclusions close the paper.

2. System Model

A simplified circuital scheme of the PV system is presented in the Figure 3, which considers

a boost converter due to the widely adoption of such a step-up DC/DC structure in PV systems,

however the analysis presented in this paper can be extended to other DC/DC topologies.

The scheme includes a voltage source as the system load, this to model the DC-link of double-stage

structures in commercial PV inverters, in which the DC/AC stage regulates the DC-link voltage

(Cb capacitor voltage) [25]. This voltage source model is widely used to represent the closed-loop

grid-connected inverters due to its satisfactory relation between accuracy and simplicity, which is

confirmed in references [3,17,25–29].

12366

Energies 2015, 8, 12363–12387

The scheme considers the SMC acting directly on the MOSFET by means of the signal u, which

is a discontinuous variable. Hence, no linearization or PWM are needed. This condition reduces the

implementation cost, and circuit complexity, in comparison with SMC solutions implemented with

Energies

PWM 2015, 8 e.g., [21].

circuits, 6

FigureFigure

3. Circuital scheme of the sliding mode controller (SMC) loop.

The dynamic behavior of the DC/DC converter is modeled by the switched Equations (1) and

The dynamic behavior of the DC/DC converter is modeled by the switched Equations (1) and (2) [30],

(2) [30], where i L represents the inductor current, v pv is the PV voltage, i pv represents the PV module

wherecurrent,

represents the inductor current, is the PV voltage, represents the PV module current,

v is the load voltage and L and C represent the inductor and capacitor values:

b in

is the load voltage and and represent the inductor and capacitor values:

di L v pv ´ vb p1 ´ uq

dt

“

L

1 (1)

(1)

dv pv i pv ´ i L

“ (2)

dt Cin

(2)

The current of the PV module is modeled with the simplified single diode model [17] given in

TheEquation

current(3).

of In

thesuch model isc

PV amodule is represented the short-circuit

modeled with current

the simplified thatdiode

single is almost proportional

model [17] given in

to the irradiance [17], B is the diode saturation current and A represents the inverse of the

Equation (3). In such a model represented the short-circuit current that is almost proportional to the

thermal voltage that depends on the temperature [25]. Such a model parameters are calculated

irradiance

as A [17],

“ Bistcis.exp

thep´B

diode saturation current

Bstc and A represents the

stc .voc q, B “ 1`a . T ´T and Bstc “ pvinverse

lnp1´i MPP {istc qof the thermal voltage that

, where istc and Tstc

v p pv stc q MPP ´voc q

depends

are on

the the temperature

short-circuit [25].and

current Such a model of

temperature parameters

the module areunder

calculated

standard . exp (STC),.

as test conditions ,

respectively. vand ln 1 ⁄

oc represents the open-circuit voltage, while ,v MPP whereand i MPPandcorrespond are the short-circuit

to the PV

.

voltage and current, respectively, at the MPP for the given operating conditions. Finally, av is the

current and temperature

voltage temperature of the module

coefficient [31]: under standard test conditions (STC), respectively. represents

the open-circuit voltage, while and correspond

´ to the

¯ PV voltage and current, respectively, at

i pv “ isc ´ B e Av pv ´ 1 (3)

the MPP for the given operating conditions. Finally, is the voltage temperature coefficient [31]:

In conclusion, the non-linear equations system formed 1

by Equations (1)–(3) describes the PV (3)

system dynamic behavior in any operation condition.

In conclusion, the non-linear equations system formed by Equations (1)–(3) describes the PV

3. Sliding Surface and Controller Structure

system dynamic behavior in any operation condition.

The non-linear relation between the PV current and voltage, given in Equation (3), produces

a non-linear

3. Sliding Surface relation between theStructure

and Controller PV power and voltage (or current), which exhibits a maximum as

depicted in Figure 4. In such a maximum, i.e., the MPP, the derivative of the power with respect to the

voltage (or current) is zero as given in Equation (4). Therefore, the sliding-mode controllers proposed

The non-linear relation between the PV current and voltage, given in Equation (3), produces a

in [18–21,23] are based on Equation (4). In a similar way, this paper proposes a sliding surface based

non-linear relation

on Equation (4):between the PV power and voltage (or current), which exhibits a maximum as

depicted in Figure 4. In such a maximum,dpi.e., pv the MPP,

“ 0 and

dp pvthe derivative of the power with respect to the

“0 (4)

dv pv

voltage (or current) is zero as given in Equation di pv

(4). Therefore, the sliding-mode controllers proposed

in [18–21,23] are based on Equation (4). In a similar way, this paper proposes a sliding surface based

on Equation (4): 12367

0 0 (4)

Energies 2015, 8, 12363–12387

Energies 2015, 8 7

90 90

80 80

70 70

60 (d ppv/d vpv) = 0 60

(d ppv/d ipv) = 0

ppv [W]

p [W]

50 50

40

pv

40

30 30

20 20

10 10

0 0

0 5 10 15 20 0 1 2

ipv [A] 3 4 5

vpv [V]

Figure

Figure4.4.Power

Power curves ofaaPV

curves of PVmodule.

module.

Taking into

Taking account

into account . . i pv

thatthat p pv = v pv , the

, thefollowing

followingrelation holds

relation at at

holds thethe

MPP:

MPP:

dp pv d ` . ˘ dv pv 0 (5)

“ v pv .i pv “ v pv ` i pv “0 (5)

di pv di pv di pv

Then, the following relation defines the maximum power condition, i.e., at the MPP:

Then, the following relation defines the maximum power condition, i.e., at the MPP:

0 for and (6)

dv pv v pv

` “ 0 for v pv “ v MPP and i pv “ i MPP (6)

di pv Equation

To correlate the MPP condition i pv (4) with the time-varying signals of the DC/DC converter,

the condition in Equation

To correlate the MPP(6) condition

is expressed in terms(4)

Equation of with

the time

the derivatives of both

time-varying the of

signals PVthe

voltage

DC/DCand

current, obtaining expression Equation (7):

converter, the condition in Equation (6) is expressed in terms of the time derivatives of both the PV

voltage and current, obtaining⁄ expression Equation (7):

⁄

0 for and (7)

dv pv {dt v pv

From such an expression, “ 0 for

the `proposed v pv “ v MPP

switching function i pv and

and Ψ “ i MPP

sliding surface Φ , given (7)

in

di pv {dt i pv

Equation (8), are defined:

From such an expression, the proposed switching function Ψ and sliding surface Φ, given in

⁄

Equation (8), are defined: Ψ ∧ Φ Ψ 0 (8)

dv pv⁄{dt v pv

Ψ“ ` ^ Φ “ tΨ “ 0u (8)

di pv {dt i pv

Then, the state-space model of both the PV system and SMC is formed by the simultaneous

Equations

Then,(1)–(3) and (8). The

the state-space modelrelations

of bothwithin such

the PV a state-space

system and SMCmodel are illustrated

is formed in the block

by the simultaneous

diagram of Figure 5, which put in evidence the signals processing and variables exchangedthe

Equations (1)–(3) and (8). The relations within such a state-space model are illustrated in block

between

diagram of Figure 5, which put in evidence the signals processing and variables exchanged between

the PV system and the SMC: the PV system disturbances are the short-circuit current (defined by the

the PV system and the SMC: the PV system disturbances are the short-circuit current (defined by the

irradiance and temperature) and the load voltage; the control signal is generated by the SMC on the

irradiance and temperature) and the load voltage; the control signal u is generated by the SMC on the

basis

basisof

of the

the PV current and

PV current and voltage;

voltage; and

and the

the SMC derives both

SMC derives both the

the PV

PVvoltage

voltageand

and current

current toto construct

construct

the

theswitching

switchingfunction.

function. Finally,

Finally, the

the SMC includes aa comparator

SMC includes comparator to to implement

implement thethe sign

sign function

function that

that

triggers

triggersthethechanges

changes on

on the

the signal

signal u.. The

The design

design of such aa comparator

of such comparator is is described

described inin the

the following

following

section in terms of the transversality and reachability conditions.

section in terms of the transversality and reachability conditions.

12368

Energies 2015, 8, 12363–12387

Energies 2015, 8 8

Figure 5. Block diagram of the SMC and PV system.

4. Analysis of the Sliding-Mode Controller

4. Analysis of the Sliding-Mode Controller

The SMC must fulfill three conditions to guarantee stability and a satisfactory performance:

The SMC mustreachability

transversality, fulfill three

and conditions to guarantee

equivalent control stability andanalyses

[32]. The transversality a satisfactory performance:

the controllability

of the system, the reachability analyses the ability of the closed-loop system to reach the surface, and

transversality, reachability and equivalent control [32]. The transversality analyses the controllability

the equivalent control grants local stability.

of the system,

Those the reachability

conditions analyses

grant the theofability

existence of the closed-loop

the sliding-mode, which alsosystem

imposesto the

reach the surface,

conditions

and the equivalent

defined control

in Equation (9)grants

[33]: local stability.

dΨ

Those conditions grant the existence of the Ψ“ sliding-mode,

0 ^ “ 0 which also imposes the conditions (9) defined

dt

in Equation (9) [33]:

Such expressions provide information concerning the surface and its derivative. In that way, the

Ψ (10):

derivate of the switching function is given in Equation

Ψ 0Λ 0 (9)

d2 v d2 i

ˆ ˙

dΨ

pv pv pv pv dv di

“ 2 i pv ` 2 v pv ` 2 ˚ (10)

Such expressions provide information

dt d t concerning

d t the surface

dt dtand its derivative. In that way,

the derivate of the switching function is given in Equation (10):

To analyze expression Equation (10) it is required to also derive expressions Equations (2) and

Ψ Equations (11)–(13), which are components of Equation (10):

(3), it leading to expressions

2 ∗ (10)

d2 v pv

ˆ ˙

1 disc dv pv Av pv di L

“ ´ BA e ´ (11)

To analyze expression Equation

d2 t(10) it

Cinis required

dt todtalso derivedtexpressions

Equations (2) and (3),

it leading to expressions Equations (11)–(13),

d i pv which

di

aredvcomponents

pv

of Equation (10):

sc

“ ´ BA e Av pv (12)

1dt dt dt

d2 i pv d2 v pv

˙2 (11)

d2 isc

ˆ

dv pv

2

“ 2 ´ BA 2 e Av pv ´ BA2 e Av pv (13)

d t d t d t dt

Then, the small-signal admittance of the PV module y “ ´BAe Av pv is introduced into the(12)

previous equations to provide more compact expressions:

(13)

d2 v pv

ˆ ˙

1 disc dv pv di L

“ `y ´ (14)

d2 t Cin dt dt dt

Then, the small-signal admittance of the PV module y is introduced into the previous

equations to provide more compact expressions:

12369

Energies 2015, 8, 12363–12387

d i pv disc dv pv

“ `y (15)

dt dt dt

˙2

d2 i pv d2 isc d2 v pv

ˆ

dv pv

“ ` y ` Ay (16)

d2 t d2 t d2 t dt

Replacing Equations (1) and (14)–(16) into Equation (10), and performing some mathematical

manipulations, the following expression for the switching function derivate is obtained:

´ ¯ ´ ¯

dΨ disc i pv `y v pv v ´v p1´uq i pv `y v pv

dt “ dt C ´ pv Lb Cin

´in ´ ¯ ¯

dv pv i pv `y v pv dv pv dv pv

` dt y Cin ` yAv pv dt ` 2 didtsc ` 2y dt

(17)

2

` v pv dd2isct

This new expression is used to analyze the transversality, reachability and equivalent control

conditions in the following subsections.

To ensure the ability of the controller to act on the system dynamics, the transversality condition

given in Equation (18) should be granted [33]:

ˆ ˙

d dΨ

‰0 (18)

du dt

Deriving Equation (17) with respect to the signal u leads to expression Equation (19):

ˆ ˙ ˆ ˙

d dΨ v i pv ` y v pv

“´ b (19)

du dt L Cin

The analysis of Equation (19) requires to review the condition in which the MPP occurs: taking

into account that p pv = v pv ¨ i pv , condition Equation (4) corresponds to expression Equation (20).

Therefore, expressions Equations (18) and (19) lead to the transversality condition Equation (21):

dp pv ´ ¯

“ v pv ¨ ´BAe Av pv ` i pv “ v pv .y ` i pv “ 0 (20)

dv pv

ˆ ˙

d dΨ

‰ 0 i f v pv ‰ v MPP (21)

du dt

Figure 6 shows the simulation of Equation (19) considering the following conditions: a BP585 PV

module with parameters B = 0.894 µF and A = 0.703 V´1 , and a DC/DC converter with L = 100 µH,

Cin = 44 µF and vb = 24 V. The simulation illustrates the transversality condition provided by

Equation (21):

‚ At the left of the MPP (voltages lower than v MPP ) the transversality condition is fulfilled, hence

the SMC is able to act on the PV system to drive it towards the MPP.

‚ At the right of the MPP (voltages higher than v MPP ) the transversality condition is also fulfilled.

‚ At the MPP (v pv = v MPP ) the transversality condition is not fulfilled. This is not a problem

because the PV system is already at the MPP. Moreover, if the system diverges from the MPP

the transversality condition is fulfilled.

12370

At the right of the MPP (voltages higher than ) the transversality condition is also fulfilled.

At the MPP ( ) the transversality condition is not fulfilled. This is not a problem

because

Energies 2015, the PV system

is already at the MPP. Moreover, if the system diverges from the MPP

8, 12363–12387

the transversality condition is fulfilled.

80

60

ppv [W]

MPP

40

20

0

0 2 4 6 8 10 12 14 16 18 20 22

40

Transversality

No transversality

20

0

Positive

Negative transversality transversality

-20

0 2 4 6 8 10 12 14 16 18 20 22

Voltage [V]

Figure 6. Simulation of the transversality condition.

The previous conditions impose a behavior similar to the classical hysteresis implementation

Theofprevious conditions impose a behavior similar to the classical hysteresis implementation of the

the SMC for DC/DC converters: the system oscillates around the surface forming a hysteretic

SMC trajectory.

for DC/DC converters:

In the followingthe system oscillates

subsections around the surface

it will be demonstrated that suchforming a hysteretic

hysteresis trajectory.

band is imposed

In the following subsections it will be demonstrated that such hysteresis band is imposed by the

by the peak-to-peak amplitude of the voltage ripple at the PV module terminals.

peak-to-peak amplitude

4.2. Equivalent ofCondition

Control the voltage ripple at the PV module terminals.

The equivalent

4.2. Equivalent control condition analyzes the ability of the system to remain trapped inside

Control Condition

the surface [31,32]. This condition is analyzed in terms of the equivalent analog value ueq of the

discontinuous control signal u: if such an equivalent value is constrained within the operational

The equivalent control condition analyzes the ability of the system to remain trapped inside the

limits the system will not be saturated and its operation inside the surface is possible. Since the

surface [31,32].

control signalThis condition tois the

u corresponds analyzed

MOSFETinactivation

terms ofsignal,

the equivalent

its operational analog

limits value

are 0 and 1.of the

discontinuous

Moreover,control signal

since the : ifconsiders

analysis such an the

equivalent valuethe

system inside is constrained within thedtoperational

surface, the condition dΨ

“ 0 holds.limits

Then, matching expression Equation (17) to zero, and replacing u = u

the system will not be saturated and its operation inside the surface is possible. eq , enables to obtain the ueqsignal

Since the control

value given in Equation (22):

u corresponds to the MOSFET activation signal, its operational limits are 0 and 1. Moreover, since the

C

analysis considers the system inside

ueq “ 1the

v

surface,

´ vpv ` v1 didtscthe

` vcondition

y dv pv

dt `

in 0 holds.

dv pv

dt

Then, matching expression

b b b vb pi pv

¯`y v pv q

Equation (17) to zero, and replacing ˚ yAv pv, enables tosc obtain the value given in Equation (22):

´

dv pv di dv pv

dt ` 2 dt ` 2y dt ˙ (22)

1 `

Cin v pv d2 isc

1 vb pi pv `yv pv q d2 t

Such an expression can be rewritten as Equation (23), where u1 and u2 are given in Equations(22)

∗

(24) and (25), respectively: 2 2

ueq “ u1 ` u2 (23)

v pv

Such an expression can be rewritten as Equation (23),

u1 “ 1 ´ where and are given in Equations

(24) (24)

vb

and (25), respectively:

Cin v pv d2 isc

„ ˆ ˙

1 disc dv pv Cin dv pv dv pv disc dv pv

u2 “ `y ` yAv pv `2 ` 2y ` (25)

vb dt dt i pv ` y v pv dt dt dt dt i pv ` y v pv d2 t (23)

1 (24)

12371

Energies 2015, 8, 12363–12387

The analysis of Equation (23) must be addressed by writing the derivative of Equation (8) in a

different way as given in Equation (26):

ˆ ˙ ¯2 ´

d2 v pv di pv d2 i pv dv pv

´ ¯

2 di pv dv pv di pv

i pv d2 t dt ´ d2 t dt

` dt dt i pv ´ dt .v pv

dΨ

“ ¯2 (26)

dt

´

di pv

i pv 2 ˚ dt

Since the solar irradiance does not exhibit fast changes in comparison with the switching

frequency of the converter, it is assumed constant ( didtsc “ 0) to simplify the analysis of Equations

(17), (25) and (26), obtaining the simplified expressions Equations (27) and (28), respectively:

˜ ` ˘ ¸

dΨ dv pv i pv y ´ i pv A ´ y2 v pv

“ (27)

dt dt y ˚ i pv

ˆ ˙

y dv pv Cin dv pv ` ˘

u2 “ 1` Av pv ` 2 (28)

vb dt i pv ` y v pv dt

Taking into account that i pv “ 0 is a realistic condition since i pv = 0 does not occur at any MPP,

dΨ dv pv

the sliding-mode condition dt “ 0 is reached in Equation (27) for dt “ 0. Then, replacing those

dv pv

conditions dt“ 0 and i pv “ 0 into Equation (28) leads to u2 = 0.

Therefore, at low irradiance derivatives, expression Equation (23) for ueq is approximately

equal to u1 . To evaluate the validity of such an approximation, ueq from Equation (23) was

simulated at different irradiance derivatives and contrasted with u1 from Equation (24): the first test

considers a change from the highest irradiance possible on earth to a complete shade in 1 second,

i.e., 1 sun per second or dS/dt = 1 kW/(m2 ¨ s). The second test considers a derivative of

10 suns per s (dS/dt = 10 kW/(m2 ¨ s)), and the third test considers a derivative of 100 suns per s

(dS/dt = 100 kW/(m2 ¨ s)), which is very large. The tests results are presented in the Table 1, where it

is observed that the error generated by assuming ueq « u1 is less than 1% for the case with the largest

derivative, while the errors for the other cases are less than 0.1% and 0.01%.

1 kW /(m2 . s) 0.0075% < e < 0.008%

10 kW /(m2 . s) ´0.0912% < e < 0.04%

100 kW /(m2 . s) 0.989% < e < 0.6883%

Thus, the equivalent control analysis is based on expression Equation (29), which does not

introduce a significant error:

v pv

ueq « 1 ´ (29)

vb

Finally, since the equivalent control condition for a DC/DC converter application is 0 < ueq < 1,

such an expression is analyzed using the ueq value given in Equation (29), which leads to expression

Equation (30):

v pv

0 ă 1´ ă 1 Ñ 0 ă v pv ă vb (30)

vb

The condition in Equation (30) is always fulfilled since the analyzed PV system considers a boost

converter. Therefore, the equivalent control is fulfilled.

12372

Energies 2015, 8, 12363–12387

The reachability conditions analyze the ability of the system to reach the desired condition Ψ = 0.

The work in [32] demonstrated that a system that fulfills the equivalent control condition also fulfills

the reachability conditions. That work also shows that the sign of the transversality condition imposes

the value of u for each reachability condition. A positive value of the transversality condition imposes

the following reachability conditions:

ˇ ˇ

dΨ ˇˇ dΨ ˇˇ

lim “ ą0 (31)

ΨÑ0´ dt u“1 dt ˇu“1,Ψ“0

ˇ

ˇ ˇ

dΨ ˇˇ dΨ ˇˇ

lim “ ă0 (32)

Ψ Ñ0` dt ˇu“0 dt ˇu“0,Ψ“0

Instead, a negative value of the transversality condition imposes the following

reachability conditions: ˇ ˇ

dΨ ˇˇ dΨ ˇˇ

lim “ ą0 (33)

ΨÑ0´ dt u“0 dt ˇu“0,Ψ“0

ˇ

ˇ ˇ

dΨ ˇˇ dΨ ˇˇ

lim “ ă0 (34)

ΨÑ0` dt u“1 dt ˇu“1,Ψ“0

ˇ

Moreover, the reachability conditions define the implementation of the switching law [32].

However, Section 4.1 shows that the transversality condition in this SMC is both positive and negative

depending on the operation condition. In such a way, from Figure 6 it is observed that the voltage

range for negative transversality (v pv < v MPP ) is much larger than the voltage range for positive

transversality (v pv > v MPP ); therefore the implementation and analysis of the proposed SMC is

performed in such a negative transversality condition, which imposes the following control law:

{Ψ < 0 Ñ u = 0 ^ Ψ > 0 Ñ u = 1}.

Then, the reachability of the surface is fulfilled in the following conditions: dΨ

dt ă 0 with u = 1

and dt ą 0 with u = 0. Moreover, from Equation (27) it is observed that the sign of dΨ

dΨ

dt is the same

dv pv

sign of dt . Therefore, the analyzed reachability conditions are fulfilled in the following states:

dv pv

‚ The MOSFET is ON (u = 1) and the voltage is decreasing, i.e., dt ă 0.

dv pv

‚ The MOSFET is OFF (u = 0) and the voltage is increasing, i.e., dt ą 0.

It must be noted that a DC/DC converter always exhibits a periodic voltage ripple around an

average value [34]. In fact, the PV voltage corresponds to the voltage at the input capacitor Cin in

Figure 3, which voltage ripple is defined by the second order filter formed by Cin and the inductor L.

From the first differential equation of the DC/DC converter, i.e., Equation (1), it is observed that

the inductor current exhibits an almost triangular waveform as in any other second-order filter;

hence in steady-state the capacitor current has a triangular waveform centered in zero, again as in

any other second order filter [30]. Therefore, the capacitor current increases with the MOSFET OFF

(u = 0). This behavior is observed in the simulation of Figure 7, which includes the proposed SMC

with the control action implemented as u = 1 for Ψ < 0 and u = 0 for Ψ > 0, and adopting the same

electrical parameters previously described in Section 4.1. The implementation of the SMC is described

in Section 5. The simulation also confirms that the capacitor current decreases with the MOSFET ON

(u = 1). Moreover, it is noted that with a constant value for u (u = 0 or u = 1) the capacitor current

exhibits positive, negative and zero values.

12373

This behavior is observed in the simulation of Figure 7, which includes the proposed SMC with the

control action implemented as 1 for Ψ 0 and 0 for Ψ 0, and adopting the same electrical

parameters previously described in Subsection 4.1. The implementation of the SMC is described in

Section 5. The simulation also confirms that the capacitor current decreases with the MOSFET ON

1).2015,

(Energies Moreover, it is noted that with a constant value for (

8, 12363–12387 0 or 1) the capacitor current

exhibits positive, negative and zero values.

18.6

d vpv /dt > 0 d vpv /dt = 0 d vpv /dt < 0

vpv [V]

18.4

18.2

10 0.0203 0.0204 0.0204 0.0204 0.0204 0.0204

iL [A] 5

0

2 0.0203 0.0204 0.0204 0.0204 0.0204 0.0204

iC [A]

-2

0.0203 0.0204 0.0204 0.0204 0.0204 0.0204

1

u [-]

0.5

0

0.0203 0.0204 0.0204 0.0204 0.0204 0.0204

Time [s]

Figure

Figure7.7.PV

PV system dynamicbehavior.

system dynamic behavior.

This

Thismeans

means that,

that,asasreported

reportedbybythe thesecond

seconddifferential

differential equation

equation of of the DC/DC converter,

the DC/DC i.e.,

converter, i.e.,

Equation (2), the voltage ripple is a series of quadratic waveforms generated from the integral of the

Equation (2), the voltage ripple is a series of quadratic waveforms generated from the integral of

the capacitor

capacitor current

current iCin =iCin

since since = iipv

ipv − L. In i L . Ina way,

´ such such the

a way,PV the PV voltage

voltage exhibits exhibits

positive positive and

and negative

negative derivatives for both states of u as reported in Figure 7. Therefore, since

derivatives for both states of as reported in Figure 7. Therefore, since the switching function derivative the switching

function derivative has the same sign of the PV voltage derivative, as reported in Equation (27), the

has the same sign of the PV voltage derivative, as reported in Equation (27), the reachability conditions

reachability conditions are fulfilled in a fraction of the time intervals in which the MOSFET is in ON

are fulfilled in a fraction of the time intervals in which the MOSFET is in ON and OFF states, which

and OFF states, which means that in the other fractions of the time intervals the system diverges from

means that inSuch

the surface. the conditions

other fractions

createofa the time behavior

periodic intervals in

thewhich

systemthediverges from the surface.

system converges Such

and diverges

conditions

to/from the create a periodic behavior in which the system converges and diverges to/from the surface.

surface.

ToToillustrate

illustratethe the

periodic behavior

periodic of theofreachability

behavior conditions,

the reachability and theand

conditions, resulting periodic periodic

the resulting behavior

of the SMC,

behavior Figure

of the SMC, 8 shows

Figuretwo diagrams:

8 shows A state flow

two diagrams: describing

A state the convergence

flow describing and divergence

the convergence and

divergence states in Figure 8(a), and the associated time-depending signal of the

states in Figure 8(a), and the associated time-depending signal of the switching function in Figure 8(b). switching function

in Figure 8(b). For the sake of illustration, the system is considered at the beginning of the analysis

For the sake of illustration, the system is considered at the beginning of the analysis underdΨ the surface, i.e.,

under the surface, i.e., Ψ < 0, and with a negative switching function derivative, i.e., dt ă 0, which

Ψ 0 , and with a negative switching function dvderivative,

pv

i.e., 0 , which corresponds to a

corresponds to a negative PV voltage derivative dt ă 0 as reported in Equation (27). Due to the

dv

differential equation governing the capacitor “ iCin , the condition dΨ

Cin dtpv dt ă 0 also corresponds

to a negative input capacitor current, i.e., iCin ă 0. Those initial conditions are observed at the top of

Figure 8(a). Then, due to the control action implemented for the SMC, the control signal is set to u = 0

and the system enters in the first state of operation:

1. Diverge from the MPP: u = 0, Ψ < 0, dΨdt ă 0 (iCin ă 0). In this first stage (block 1 in Figure 8(a))

the reachability conditions are not fulfilled, hence the system diverges from the MPP, which

is observed in Figure 8(b). Such a condition is caused by the negative value of the switching

function derivative, which reduces even more the value of Ψ, hence the desired condition Ψ = 0

is not achievable. However, the control action u = 0 forces the increment of iCin , it driving the

system to iCin “ 0 ( dΨ dΨ

dt “ 0) and eventually to iCin ą 0 ( dt ą 0), which corresponds to the

second stage of the SMC operation.

2. Converge to the MPP: u = 0, Ψ < 0, dΨ dt ą 0 (iCin ą 0). In this second stage the reachability

condition Equation (33) is fulfilled, hence the system converges to the MPP, which is also

observed in Figure 8(b). In this case, the positive value of the switching function derivative

increases the value of Ψ, it driving the system towards the desired condition Ψ = 0. However,

when the condition Ψ = 0 is reached the switching function derivative is still positive, i.e.,

dt ą 0 (iCin ą 0), hence Ψ becomes positive. Moreover, when Ψ ą 0 the SMC imposes u “ 1,

dΨ

12374

Energies 2015, 8, 12363–12387

3. Diverge from the MPP: u = 1, Ψ ą 0, dΨ dt ą 0 (iCin ą 0). In this third stage the reachability

conditions are not fulfilled, hence the system diverges from the MPP. Such a condition is caused

by the positive value of the switching function derivative, which increases even more the value

of Ψ, hence the desired condition Ψ “ 0 is not achievable. However, the control action u “ 1

forces the decrement of iCin , it driving the system to iCin “ 0 ( dΨ

dt “ 0) and eventually to iCin ă 0

dΨ

( dt ă 0), which corresponds to the fourth stage of the SMC operation.

4. Converge to the MPP: u “ 1, Ψ ą 0, dΨ dt ă 0 iCin ă 0). In this fourth stage the reachability

condition Equation (34) is fulfilled, hence the system converges to the MPP. In this case, the

negative value of the switching function derivative decreases the value of Ψ, it driving the

system towards the desired condition Ψ “ 0. However, when the condition Ψ “ 0 is reached

the switching function derivative is still negative, i.e., dΨ dt ă 0 (iCin ă 0), hence Ψ becomes

negative. Moreover, when Ψ ă 0 the SMC imposes u = 0, which corresponds to the first stage

of the SMC operation.

Energies 2015, 8 15

Limit cycle: Repeated periodically generating the hysterical behavior

(a)

(b)

Periodic Periodic behavior of the reachability

of the reachability conditions.

conditions. (a) Convergence

(a) Convergence and and divergence

divergence conditions;

conditions;

(b) Hysteretic (b) Hysteretic behavior.

behavior.

The previous four stages of the system operation are continuously repeated, they form a

limit-cycle [35] that imposes a hysteretic behavior to the SMC switching function around the

surface Ψ 0. Moreover, from Figure 8 it is noted

12375that the hysteresis band is defined by the condition

0, which corresponds to 0 and 0 as reported in Equation (27). This dependency is

Energies 2015, 8, 12363–12387

The previous four stages of the system operation are continuously repeated, they form a

Energies 2015, 8[35] that imposes a hysteretic behavior to the SMC switching function around the

limit-cycle 16

surface Ψ “ 0. Moreover, from Figure 8 it is noted that the hysteresis band is defined by the condition

dv pv

dΨ

depicted in0,Figure

dt “ which 7, and confirmed

corresponds to dt by

“ 0Figure

and icin9,“which presentsinΨ,

0 as reported and(27).

Equation u generated by the same

This dependency

is depicted in Figure 7, and confirmed by Figure 9, which presents Ψ, dΨ

and u generated by the

simulation producing producing

same simulation the signalsthe

presented in Figurein7.Figure

signals presented In addition, Figure dt9Figure

7. In addition, also confirms the limit-cycle

9 also confirms the

of thelimit-cycle

switchingof function.

the switching function. Finally, it is noted that the hysteresis band of the SMC is definedby the

Finally, it is noted that the hysteresis band of the SMC is defined

voltage

by ripple at theripple

the voltage inputatcapacitor, which can

the input capacitor, be modified

which by changing

can be modified the input

by changing capacitance.

the input capacitance.

[-] 0

Reachabality

-0.2

Condition

-0.4

4

x 10

5

d /dt [-] 0

-5

0.0203 0.0203 0.0203 0.0203 0.0203 0.0204

1

u [-] 0.5

0

0.0203 0.0203 0.0203 0.0203 0.0203 0.0204

Time [s]

Figure 9. SMC hysteretic behavior.

4.4. Switching Frequency

4.4. Switching Frequency

The switching frequency is an important parameter in the operation of DC/DC converters, hence

it must be calculated

The switching frequencyin order

is antoimportant

provide practical guidelines

parameter in thefor the SMCof

operation implementation.

DC/DC converters, hence it

The switching frequency is calculated from the DC/DC converter differential equations and

must be calculated in order to provide practical guidelines for the SMC implementation.

ripple magnitudes as reported in [30]. In such a way, in a boost converter the ripple magnitude for

Thetheswitching frequency

inductor current is calculated

and capacitor from

voltage the DC/DC

are given converter

in Equations differential

(35) and equations

(36), respectively, andTsripple

where

magnitudes as the

represents reported in period,

switching [30]. Ind represents

such a way, in acycle,

the duty boostandconverter the

Fsw =1/Ts is theripple magnitude

switching frequency:for the

inductor current and capacitor voltage are given

v pv in Equations v pv (35) and (36), respectively, where Ts

∆i L “ dTs , d “ 1 ´ (35)

represents the switching period, d represents the

2L duty cycle, andvb Fsw =1/Ts is the switching frequency:

∆ ∆i (35)

∆v pvd“ , L Ts1 (36)

2 8Cin

∆ SMC system highlighting the time intervals (36)

Figure 10 presents a simulation of the closed-loop

in which the MOSFET is turned on (Ton ) ∆and turned 8 off (Toff ). In such a figure it is observed that

the magnitude of the voltage ripple ∆v pv is mesured from the average PV voltage v pv to the peak

Figure 10 presents

(maximum) voltagea vsimulation ofpvthe

pv´ max as ∆v “ vclosed-loop

pv´max ´ v pv . SMC

Then, system highlighting

approximating the time

the average intervals in

PV voltage

to the MPP voltage, the voltage ripple magnitude becomes ∆v

which the MOSFET is turned on (Ton) and turned off (Toff). In such a figure it is observed

pv « v pv ´ max ´ v MPP , where the that the

MPP

LambertW pexpp1qp B`isc qq´1

magnitude

voltageofis the voltage

calculated ripple

from ∆ (20)

Equation is mesured

as v MPP “from the average

A PV using

voltage to the peak

the LambertW

function [36].

(maximum) voltage as ∆ . Then, approximating the average PV voltage

to the MPP voltage, the voltage ripple magnitude becomes ∆ , where the MPP

voltage is calculated from Equation (20) as using the

LambertW function [36].

12376

Energies 2015, 8, 12363–12387

Energies 2015, 8 17

vpv-max

18.5

18.4 v

vpv [V] vpv

18.3

18.2

1

u [-]

0

0.0197 0.0197 0.0197 0.0197 0.0197 0.0198 0.0198 0.0198 0.0198

Time [s]

Figure 10. Principle for calculating the switching frequency.

dv

The maximum PV voltage v pv´ max is achieved when dtpv “ 0, which also corresponds to dΨ “0

Theand

maximum PV voltage is achieved when 0, which also corresponds to dt

u = 1 as demonstrated in the previous subsection. Then, replacing those values in Equation (17),

0 and

u = 1and

as assuming

demonstrated in theEquation

isc constant, previous(37)subsection.

is obtained: Then, replacing those values in Equation (17),

and assuming constant, Equation (37) is

v pvobtained:

´max

ˆ

i pv ` y v pv´max

˙

0“´ (37)

L Cin

0 (37)

Finally, replacing expressions Equations (35) and (37) into Equation (36) with v pv « v MPP and

∆v pv «replacing

Finally, v pv´max ´ expressions

v MPP enables Equations

to calculate Ton as:and (37) into Equation (36) with

(35) and

∆ enables to calculate

g

f ˆ as: ˇ ˙

i pv ˇ

f 16LCin ´ v MPP

f

´ y ˇv “ MPP

e pv

Ton “ (38)

16 v MPP d

(38)

dv pv

Similarly, taking into account that the minimum PV voltage v pv´ min is achieved when “ 0, dt

dΨ

dt “ 0 and u = 0, the voltage ripple measured from v pv´min to the average PV voltage can be

Similarly, takingas into

approximated ∆v pv account

« v pv´maxthat the . minimum

´ v MPP PV

Then, To f f is voltageas in Equation

calculated is achieved when

(39), where d1 = 1´d 0,

represents the complementary duty cycle:

0 and u = 0, the voltage ripple measured from to the average PV voltage can be

approximated as ∆

g

f . Then,ˆ

i ˇ is calculated as in Equation (39), where d′ = 1−d

ˇ ˙

f 16LCin ´ pv v

f

y ˇv “ MPP ´ MPP

represents the complementary duty

To f fcycle:

“

e pv

(39)

v MPP d1

Finally, the switching frequency16 imposed by the SMC is given in Equation (40). In such an

expression the duty cycle is evaluated as reported in Equation (35): (39)

′

1 1

Fsw “ “d (40)

Finally, the switchingTonfrequency

` To f f imposed the SMC ´ by

´v MPP is given in Equation

¯ (40). In such an

ˆ ˇ ˙

i pv ˇ

16LCin y ˇ

´b b

v pv “MPP vb vb

expression the duty cycle is evaluated as reportedv MPP

in Equation (35):vb ´v MPP ` v MPP

The accuracy 1 1 switching frequency of the PV system

of expression is illustrated by calculating the

simulated in Figure 10: Equation (40) predicts a switching frequency equal to 27.4 kHz, while the

simulation reports a switching16 frequency equal to 28.41 kHz, which correspond to an acceptable (40)

prediction error of 3.5%.

12377

Energies 2015, 8, 12363–12387

Moreover, from expression Equation (40) it is evident that the switching frequency depends

on the inductance and capacitance values, PV module characteristics and system operating point.

Therefore, if Cin or L change due to aging or by other effects such as inductor saturation, the switching

frequency will also change. However, the previous subsections demonstrate the robustness of the

proposed SMC to changes in Cin and L: the system will remain stable for any positive values of Cin

and L. In conclusion, tolerances in the converter parameters due to aging or practical saturations will

not compromise the system stability.

The sliding-mode dynamics provide information concerning the averaged voltage and current

behavior, which is useful to predict important time-based performance criteria such as settling

time. The sliding-mode dynamics consider the system within the sliding surface, hence expressions

Equation (7) hold, which leads to Equation (41):

dv pv v pv di pv

“´ . (41)

dt i pv dt

Similarly, from the second differential equation of the DC/DC converter, i.e., Equation (2), the

dv pv

relation i pv “ Cin dt ` i L is obtained, which derivative is given in Equation (42):

di pv d2 v pv di

“ Cin 2 ` L (42)

dt d t dt

Then, substituting Equations (1) and (41) into Equation (42), and naming the instantaneous PV

v pv

module impedance R pv “ i pv , the relation Equation (43) is obtained:

d2 v pv dv pv

R pv Cin L `L ` R pv v pv “ R pv vb d1 (43)

d2 t dt

Expressing Equation (43) in Laplace domain leads to the model reported in Equation (44):

v psq R pv vb

“ (44)

d1 psq R pv Cin L s2 ` L s ` R pv

Such a second-order model describes the closed-loop system a behavior around a given operation

condition, in a which the natural frequency is wn “ 1{ pCin Lq and the damping ratio is

ξ “ 1{p2R pv C 1{ pCin Lqq. From the classical second order system analysis [37], the settling time of

the PV voltage is, approximately, ts “ 6.4R pv Cin . To test such an estimation, the PV system controlled

by the SMC, considering the parameters previously described in Section 4.1, was simulated in both

step-up and step-down irradiance transients. The first test considers a step change in the irradiance

from 600 W/m2 to 1000 W/m2 , which corresponds to a change in isc from 3 A to 5 A, obtaining

the results presented in Figure 11(a): in such conditions the average PV module impedance is 5.1 Ω,

which leads to ts = 1.4 ms. To provide a more clear calculation of the settling time from the DC/DC

converter waveforms, the PV voltage is also filtered to remove the switching ripple, which enables to

verify the accuracy of the proposed approximation. Similarly, the second test considers a step change

in the irradiance from 1000 W/m2 to 600 W/m2 , obtaining the results presented in Figure 11b, where

the settling time estimation is the same, i.e., ts = 1.4 ms. Both simulations confirm the validity of the

approximation. It must be noted that the settling time will change if Cin or L change, e.g., due to

aging or practical saturations.

12378

enables to verify the accuracy of the proposed approximation. Similarly, the second test considers a

step change in the irradiance from 1000 W/m2 to 600 W/m2, obtaining the results presented in Figure 11b,

where the settling time estimation is the same, i.e., ts = 1.4 ms. Both simulations confirm the validity of

the approximation. It must be noted that the settling time will change if or change, e.g., due to aging

Energies 2015, 8, 12363–12387

or practical saturations.

19

Unfiltered signal

18.5 Filtered signal

vpv [V]

18

Predicted ts

17.5 ts = 1.4 ms

17

0.028 0.0285 0.029 0.0295 0.03 0.0305 0.031 0.0315 0.032

2 2

(a) Transient from 600 W/m to 1000 W/m

19

18

vpv [V]

17

Predicted ts

16 ts = 1.4 ms

0.028 0.0285 0.029 0.0295 0.03 0.0305 0.031 0.0315 0.032

Time [s]

2 2

(b) Transient from 1000 W/m to 600 W/m

Figure 11.11.Settling-time

Figure estimation.

Settling-time estimation.

5. Numerical Results and Performance Evaluation

The proposed SMC was implemented in the power electronics simulator PSIM using C language.

TheThis

proposed

procedureSMCenables

was implemented in the power

to tests, realistically, electronics simulator

the performance of the SMC PSIM using C language.

by emulating the

implementation of the controller in a Digital Signal Processor (DSP). In addition,

This procedure enables to tests, realistically, the performance of the SMC by emulating the implementation the ANSI C code

used in the PSIM simulation (within a C block) can be directly executed in any DSP without

of the major

controller in a Digital Signal Processor (DSP). In addition, the ANSI C code used in the PSIM

modifications.

simulation Figure

(within12a presents

C block)the cansimulation

be directly executed

scheme and in anyimplementation,

SMC DSP without major which modifications.

uses four devices to

Figure 12 presents

implement the simulation

the controller: a C block scheme

emulating andtheSMC implementation,

DSP, two classical comparators whichand uses four

a S-R devices to

flip-flop

to store the u signal. Such a structure is widely adopted in literature

implement the controller: a C block emulating the DSP, two classical comparators and a S-R flip-flop to implement the switching

function due to its simplicity and reliability [32]. However, since in this work the SMC performs

to store the u signal. Such a structure is widely adopted in literature to implement the switching

both MPPT and voltage regulation actions, there is no need of cascade (or any other) controllers or

function due to itsFigure

modulators. simplicity

12 alsoand reliability

presents the ANSI[32].C However,

code used to since in this

calculate thework the SMC

switching performs

function, which both

MPPTis and voltage

provided regulation

by using an emulatedactions, there is no Converter

Digital-to-Analog need of (DAC) cascade (orswitching

to the any other) controllers

circuit driving or

the MOSFET. To test the SMC in the same conditions used to illustrate the mathematical analysis, this

modulators. Figure 12 also presents the ANSI C code used to calculate the switching function, which is

simulation scheme considers a DC-link voltage near to 30 V imposed by the standard voltage source

provided

modelby representing

using an emulated Digital-to-Analog

the inverter. Converterscheme

An additional simulation (DAC)will to be

theused

switching

afterwardscircuit driving

to test the the

MOSFET.SMC, To testhigh

under theboosting

SMC inconditions,

the sameand conditions

interacting used

withtoa illustrate the mathematical

detailed grid-connected analysis, this

inverter.

simulationFigure

scheme 13 considers

shows twoatests DC-link voltage

performed near

in the to simulation

first 30 V imposed scheme: by Figure

the standard voltagethe

13(a) presents source

dynamic response of the system for a change in the irradiance from 600 W/m2 to 1000 W/m2 ,

model representing the inverter. An additional simulation scheme will be used afterwards to test the

while Figure 13(b) presents the dynamic response of the system for a change in the irradiance from

SMC, 1000

underW/m high2 to

boosting

600 W/m conditions,

2 . In bothand casesinteracting with a the

the SMC drives detailed grid-connected

PV system to the optimalinverter.

operation

Figure 13 shows

condition within twothetests performed

estimated in time.

settling the first simulation

Moreover, suchscheme:

simulations Figurealso13(a)

show presents

the pointsthein dynamic

the

2 2 condition; then

response of the system for a change in the irradiance from 600 W/m to 1000 W/m , while Figure 13(b)

I-V curve in which the PV system operates: (1) corresponds to the initial steady-state

condition (2) is triggered by an irradiance perturbation; and (3) is the new MPP detected by2 the SMC. 2

presents the dynamic response of the system for a change in the irradiance from 1000 W/m to 600 W/m .

12379

In

In both

both cases

cases the

the SMCSMC drives

drives the

the PV

PV system

system to

to the

the optimal

optimal operation

operation condition

condition within

within the

the estimated

estimated

settling

settling time.

time. Moreover,

Moreover, such such simulations

simulations also

also show

show the

the points

points in

in the

the I-V

I-V curve

curve in

in which

which the

the PV

PV system

system

operates:

operates: (1)

(1) corresponds

corresponds to to the

the initial

initial steady-state

steady-state condition;

condition; then

then condition

condition (2)

(2) is

is triggered

triggered by

by an

an

Energies 2015, 8, 12363–12387

irradiance perturbation; and (3) is the new MPP detected

irradiance perturbation; and (3) is the new MPP detected by the SMC. by the SMC.

Figure 12.

12. Simulation

Figure

Figure 12. Simulationscheme

Simulation scheme and

scheme and SMC

andSMC

SMC implementation.

implementation.

implementation.

19

19

18

18 (1)

(1) (3)

[V]

(3)

Voltage[V]

17

17 18.5 (2)

18.5 (2)

Voltage

18

16

16

18

17.5

17.5

15

17

17

tts=1.4ms

=1.4ms

15 16.5

16.50.0595

s

0.06 0.0605 0.061 0.0615

14 0.0595 0.06 0.0605 0.061 0.0615

14

0.02 0.03 0.04 0.05 0.06 0.07 0.08

0.02 0.03 0.04 0.05 0.06 0.07 0.08

100

100

[W]

Power[W]

Power

50

50

0.02

0.02 0.03

0.03 0.04

0.04 0.05

0.05 0.06

0.06 0.07

0.07 0.08

0.08

Time

Time [s]

[s]

100

100

[W]

Power[W]

(2)

(2) (3)

(3)

50

50

Power

(1)

(1)

00

22 44 66 88 10

10 12

12 14

14 16

16 18

18 20

20 22

22

Voltage

Voltage [V]

[V]

(a)

(a)

Figure 13. Cont.

Figure

Figure 13.

13. Cont.

Cont.

12380

Energies 2015, 8 21

Energies 2015, 8, 12363–12387

Energies 2015, 8 21

19

18

Voltage[V]

19

19

18

(1)

17 (3)

18 17

(2)

Voltage[V]

19

16 16

18

(1)

17 (3)

15 15

17 ts = 1.4 ms

16 (2)

16 0.06 0.0605 0.061 0.0615

14

150.02 15 0.03 ts = 1.4 0.04

ms 0.05 0.06 0.07 0.08

100 0.06 0.0605 0.061 0.0615

[W] [W]

14

0.02 0.03 0.04 0.05 0.06 0.07 0.08

50

100

PowerPower

500

0.02 0.03 0.04 0.05 0.06 0.07 0.08

0 Time [s]

100

0.02 0.03 0.04 0.05 0.06 (1)

0.07 0.08

[W] [W]

100

50 (1)

PowerPower

(2)

50

(3)

0

0 2 4 6 8 10 12 14 16 (2)

18 20 22

0

Voltage[V]

0 2 4 6 8 10 12 14 16 18 20 22

(b)

Voltage[V]

(b)

Figure 13. System Response to irradiance perturbations. (a) Transient from 600 W/m2 to

2 2 2

1000Figure

FigureW/m ; (b)

13.13.SystemTransient

System from

Response

Response to1000 W/m

irradiance

to irradiance to 600 (a)

W/m

perturbations.

perturbations. . (a) from

Transient Transient

2 to 1000 W/m2 ; 2

from

600 W/m 600 W/m to

(b) Transient from 1000 W/m2 to 600 W/m2 .

1000 W/m2; (b) Transient from 1000 W/m2 to 600 W/m2.

Taking into account that grid-connected PV systems exhibit perturbations at the DC-link, the SMC

must be also

Taking intotested

Taking into in

account presence

account

that of sinusoidal

that grid-connected

grid-connected voltage

PV systems

PV systems oscillations at Cb. Such

exhibit perturbations

exhibit perturbations

atathe

at the test is presented

DC-link,

DC-link,

the

the SMC in

SMC must be also tested in presence of sinusoidal voltage oscillations at Cb . Such a test is presented

Figurebe14,

must where

also testeda large 20 V peak-to-peak

in apresence sinusoidalvoltage

voltageoscillation hasatbeen superimposed topresented

a 29 V DC

in Figure 14, where large 20 Vofpeak-to-peak voltage oscillations

oscillation has beenCsuperimposed

b. Such a testtois a 29 V DC in

component, this

Figurecomponent, corresponding

14, where this

a large to a 69%

20 V peak-to-peak

corresponding perturbation. The

voltage oscillation

to a 69% perturbation. simulation

has been

The simulation alsoalso considers

superimposed an irradiance

to a 29 V DC

considers an irradiance

2 2 22

perturbation from

perturbation 1000

from W/m

1000 W/m to 600

to 600 W/m

W/m atattt =

= 0.03

0.03 s,s,

andand even

even under

under

component, this corresponding to a 69% perturbation. The simulation also considers an the the perturbation

perturbation in thein the load

load

irradiance

voltage

voltage the SMCthe SMC is able to impose the expected behavior to reach the new MPP.

perturbation from is1000

ableW/mto impose

2

to 600theW/m

expected

2

at t behavior

= 0.03 s,toand reach

eventheunder

new MPP.

the perturbation in the load

voltage the SMC is able

20to impose the expected behavior to reach the new MPP.

vpv [V]

20

18

vpv [V]

18

16

90

p0.05 ppv

800 0.01 0.02 0.03 0.04 max 0.06

90 49.5

ppv [W]

70 49 pmax ppv

80 85.5

60 85 49.5

48.5

ppv [W]

70 85.5

84.5 49

0.04 0.045 0.05 0.055

50 0.01 0.015 0.02

60 85 48.5

40 84.5 0.04 0.045 0.05 0.055

50 0 0.01

0.01 0.015 0.02

0.02 0.03 0.04 0.05 0.06

40

40

0 0.01 0.02 0.03 0.04 0.05 0.06

vb [V]v [V]

40

30

b

30

20

0 0.01 0.02 0.03 0.04 0.05 0.06

20 Time [s]

0 0.01 0.02 0.03 0.04 0.05 0.06

Time [s]

Figure 14. System response with load perturbation.

Figure 14. System response with load perturbation.

12381

Energies 2015,

Energies 2015, 88 22

22

As anticipated

As anticipated before,

before, aa secondsecond simulation

simulation scheme scheme is is used

used to to test

test the

the SMC

SMC underunder high high boosting

boosting

Energies 2015, 8, 12363–12387

conditions and

conditions and interacting

interacting with with aa detailed

detailed grid-connected

grid-connected inverter. inverter. Such Such aa scheme

scheme is is presented

presented in in

Figure 15

Figure 15 [24],

[24], which

which includes

includes aa closed-loop

closed-loop inverterinverter withwith two control

control objectives: provide provide a givengiven power

power

As anticipated before, a second simulation scheme istwo used to test objectives:

the SMC under highaboosting

factor and

factor and regulate

regulate the the DC-link

DC-link voltage.

voltage. The The inverter

inverter isis designed

designed to to operate

operate at at 110

110 VAC@60

VAC@60 Hz Hz with an an

conditions and interacting with a detailed grid-connected inverter. Such a scheme is presented with in

average

average DC-link

DC-link

Figure 15 [24],voltage

voltage equal

whichequal

includes to 220

to 220 V. The

V. The DC-link

a closed-loop DC-link

inverter is is formed

formed

with by

two control 14 F,

by objectives:

14 F, which

which

provideat ata1000

1000 W/m22

givenW/m

experiment aa large

power factor

experiment large

and 120

regulate

120 V V peak-to-peak

peak-to-peak

the DC-link voltage. voltage

voltage

The oscillation superimposed

inverter is designed

oscillation superimposed

to operatetoat the DC component,

110 VAC@60

to the DC component,

Hz

with an

this corresponding

this

average

corresponding2 to DC-link

to aa 55% voltage

55% perturbation. equal

perturbation. Moreover,to 220 V.

Moreover, the The DC-link

the input is formed

input capacitor was by C

was changed

changed

b “ 14 µF,

to which150 FF

150

at 1000 W/m experiment a large 120 V peak-to-peak voltagecapacitor

oscillation superimposed to to the DC

to avoid

to avoid large PV

large

component,

PVthis

voltage

voltage oscillations

oscillations

corresponding

due perturbation.

due

to a 55%

to the

to the high

high voltage

voltage

Moreover,

conversion

conversion ratio imposed

the input ratio

imposed

capacitor

to the

to

was changed

the DC/DC

DC/DC

to

converter.

converter. Therefore,

Cin “Therefore,

150 µF to avoid the SMC

the SMCPV

large switching

switching frequency,due

frequency,

voltage oscillations calculated

calculated

to the high from

from Equation

Equation

voltage (40),ratio

(40),

conversion is approximately

is approximately

imposed

100 kHz.

to theIn addition,

DC/DC the

converter. simulation

Therefore, the also SMCconsiders

switching an irradiance

frequency, perturbation

calculated from

from Equation 1000 is 22 to

(40),W/m

100 kHz. In addition, the simulation also considers an irradiance perturbation from 1000 W/m to

300 approximately

W/m 2 100 kHz. In addition, the simulation also considers an irradiance perturbation from

2 at t = 0.65 s to illustrate the MPPT operation of the SMC. The results are presented in

300 W/m 1000 W/m at t 2=to0.65 s to illustrate

300 W/m 2 at t = 0.65 thes toMPPT operation

illustrate the MPPT of operation

the SMC.of The results

the SMC. Theare presented

results are in

Figure 16,

Figurepresented which

16, which confirms

confirms

in Figure

that even

that even

16, which

under

under

confirms that

high

high DC-link

evenDC-link

under high

perturbations

perturbations the SMC

the SMC the

DC-link perturbations

is able

is able

SMCto

to reach

is reach

able the

the

new MPP.

to reach It must

the new be

MPP.point

It out

must that

be the

point outinverter

that the also imposes

new MPP. It must be point out that the inverter also imposes a transitory behavior in the the

inverter also a transitory

imposes a behavior

transitory in

behavior the

in DC-link

DC-link

voltage

voltage due to

DC-link

due tovoltage

the change

the change

due toin the

in PV PV power

change

power

in at

at tt == 0.65

PV power 0.65ats,s, which

t which

= 0.65 s,iswhich

also mitigated

is also mitigated by the

is also mitigated

by the bySMC.

SMC.

the SMC.

FigureFigure

Figure 15. Circuital

15. Circuital scheme

15. Circuital schemeof

scheme

of the detailed

the detailed

of the detailed grid-connected

grid-connected PV system.

system.

PV system.

grid-connected PV

20

20

[V]

18

v [V]

18

pv

vpv

16

16

0.4 0.5 0.6 0.7 0.8 0.9 1

0.4 0.5 0.6 0.7 0.8 0.9 1

100

100 Pmax Ppv

Pmax Ppv

[W]

p [W]

50

50

pv

ppv

0

00.4 0.5 0.6 0.7 0.8 0.9 1

0.4 0.5 0.6 0.7 0.8 0.9 1

300

300

[V]

vbv [V]

200

b

200

0.4 0.5 0.6 0.7 0.8 0.9 1

Time [s]

Time [s]

FigureFigure

Figure 16. SMC

16. SMC behavior

16. SMC behaviorinteracting

behavior with

interacting with

interacting with aa grid-connected

grid-connected

a grid-connected inverter.

inverter.

inverter.

12382

Energies 2015, 8 23

Energies 2015, 8 23

Energies 2015, 8, 12363–12387

With the aim of further evaluate the proposed SMC, the performance of this controller is also

contrasted

With thewith aimtwo of classical solutionsthe

further evaluate based on the SMC,

proposed P & O thealgorithm. The P of

performance & this

O method is based

controller on

is also

contrasted

perturb With

thewith the

inputtwo aim of further

classical

variable evaluate

solutions

of the the

system,based proposed

e.g., on

thethe SMC,

P&

duty the performance

O algorithm.

cycle of this controller

The P &observe

of the converter, O method is also

is based on

the variation in

contrasted with two classical solutions based on the P & O algorithm. The P & O method is

the output

perturb thepower, and increase

input variable of theorsystem,

decrease thethe

e.g., perturbed variable

duty cycle to converter,

of the increase the power.the

observe

based on perturb the input variable of the system, e.g., the duty cycle of the converter, observe the

The flowchart

variation in

of

thethe P & power,

output O algorithm

variation in theand is shown

increase

output power,orin

andFigure

decrease 17.

increasetheorperturbed

decrease thevariable to increase

perturbed theincrease

variable to power.theThe flowchart

power.

of the P & O algorithm is shown in Figure 17.

The flowchart of the P & O algorithm is shown in Figure 17.

FigureFigure

17. Flowchart

17. Flowchart ofofthe

thePerturb andObserve

Perturb and Observe (P &(P

O)& O) algorithm.

algorithm.

The structure of the two classical solutions based on the P & O algorithm are presented in Figure 18:

(a) The &The

a P structurestructure ofdefining

of the two

O algorithm the two the

classical

classical duty solutions

solutions of based

based

cycle on theonP the

a PWM &(P O&P algorithm

&O O+ algorithm

PWM) are presented

are presented

driving in 18:

in MOSFET;

the Figure

Figure 18: (a) a P & O algorithm defining the duty cycle of a PWM (P & O + PWM) driving the

and

(a) a(b)P a&P O& algorithm

O algorithm defining

defining thetheduty

voltage

cyclereference

of a PWM of a (PPID&controller,

O + PWM) which in turns

driving the defines

MOSFET;the

MOSFET; and (b) a P & O algorithm defining the voltage reference of a PID controller, which in turns

duty

and (b)cycle

a P of

defines a duty

&theO PWM (P of&

algorithm

cycle aO + PID

defining

PWM the&+voltage

(P OPWM). Such aSuch

+ PID reference

+ PWM). figure

of a aPID put in

putevidence

controller,

figure thethe

which

in evidence inadvantages

turns of the

defines

advantages

duty cycle

proposed of a PWM

of thesolution:

proposed the (P & the

SMC

solution: O +SMC

only PIDonly

+ PWM).

requires a singleSuch

requires a control

a control

single figure put

block,block, ininstead

instead evidence the advantages

the classical solutions

the classical of the

require

solutions

two require

and three

proposed two and

control

solution: three control

theblocks. blocks.

SMC only requires a single control block, instead the classical solutions require

two and three control blocks.

Closed-loop

Cin Cb grid-

PV dc/dc Closed-loop

connected

Cin Cb grid-

inverter

PV dc/dc connected

inverter

P&O PWM

cascade control

(a) It requires the use of

P&O PWM

(b) cascade control

(a) P&O + PWM

(b) PID

Control (a) P&O

(b) P&O ++ PID

PWM + PWM

PID

Control (b) P&O + PID + PWM

Figure 18. Structure of a conventional Maximum Power Point Tracking (MPPT) architecture.

Figure 18. Structure of a conventional Maximum Power Point Tracking (MPPT) architecture.

Figure 18. Structure of a conventional Maximum Power Point Tracking (MPPT) architecture.

The P The

& OP algorithms

& O algorithmswerewere designed

designed followingthe

following the guidelines

guidelines given

giveninin

[3,38], which

[3,38], are aimed

which to

are aimed to

guarantee a perturbation period larger than the PV voltage settling time. Concerning the stand-alone

guarantee

The P &a Operturbation

algorithmsperiod larger than

were designed the PV the

following voltage settlinggiven

guidelines time.inConcerning

[3,38], whichtheare

stand-alone

aimed to

guarantee a perturbation period larger than the PV voltage settling time. Concerning the stand-alone

12383

Energies 2015, 8 24

P & OEnergies

controller (a), the parameters are: perturbation period Ta = 0.5 ms and perturbation amplitude

2015, 8, 12363–12387

Δd = 0.2.

TheP design of the(a),

& O controller PIDthecontroller

parametersfor

are:the cascade period

perturbation P & OTasolution (b) perturbation

= 0.5 ms and was based on the system

amplitude

2

Equations

∆d =(1)–(3)

0.2. linearized at the lower irradiance condition (100 W/m ) as suggested in [3]. Then,

the transferThe

function

design between

of the PIDthe PV voltage

controller and

for the the duty

cascade P &cycle is (b)

O solution calculated as on

was based in the

Equation

system (45).

Equations (1)–(3) linearized at the lower irradiance condition (100 W/m2 ) as suggested in [3]. Then,

It is important to remark that changes with the solar irradiance and v temperature, hence the settling

the transfer function between the PV voltage and the duty cycle dpv is calculated as in Equation

time ts(45).

of the

It isPV voltage to

important constantly changes

remark that [39]: with the solar irradiance and temperature, hence the

R pv changes

settling time ts of the PV voltage constantly changes [39]:

(45)

v pv psq R pv vb

“´ (45)

d psq L Cin R pv s2 ` L s ` R pv

Therefore, conventional linear controllers for PV systems must be designed at the longest settling

2

time value,Therefore,

which is conventional

obtained forlinear

the larger valuefor

controllers ofPV systems

(lowermust

irradiance, 100atW/m

be designed in thissettling

the longest example).

time value, which is obtained for the larger value of R (lower irradiance, 100 W/m2 in this example).

Then, using the root-locus placement technique, the pv

controller in Equation (46) was designed.

Then, using the root-locus placement technique, the controller in Equation (46) was designed.

The perturbation period

The perturbation of the

period of Pthe&PO& algorithm

O algorithm is is

designed

designedfor

forthe

theworst

worst case scenario(longest

case scenario (longest )

following the guidelines

ts ) following given given

the guidelines in [3],in it[3],leading to to

it leading TaTa== 0.5 ms and

0.5 ms and∆v

Δv

refref= = 0.25

0.25 V (perturbation

V (perturbation

amplitude applied to the PID reference):

amplitude applied to the PID reference):

10.1302

10.1302 ps `2.25 4ps ` 1423q

2.25e4q 1423

Gc psq “ ´ (46) (46)

s ps `8.453

8.453e5q5

Finally, Finally,

to provideto provide a fair comparison,

a fair comparison, both P &both O +PPWM & O + PWM

and P &andO +PPID & O + PID were

+ PWM + PWM were

implemented

implemented with a switching frequency equal to 100 kHz, which is the same switching frequency

with a switching frequency equal to 100 kHz, which is the same switching frequency exhibited by the SMC.

exhibited by the SMC.

A first test was

A first testperformed considering

was performed a much

considering a much larger

largerDC-link capacitanceCb “ 200200

DC-link capacitance µF to F to reduce

reduce

the voltage oscillations

the voltage generated

oscillations by the

generated by inverter. This This

the inverter. test also includes

test also the the

includes start-up of the

start-up PVPV

of the system

and ansystem and an

irradiance irradiance perturbation

perturbation at t = 0.65ats.t =The

0.65simulation

s. The simulation

resultsresults are presented

are presented in Figure

in Figure 19,19,where

where the fast tracking of the MPP provided by the proposed SMC is observed. The simulation also

the fastreports

tracking of the MPP provided by the proposed SMC is observed. The simulation also reports

the small voltage oscillations exhibited due to the large value of Cb by the three systems,

the small

which voltage oscillations

enables the correctexhibited

operation due to Pthe

of the & large

O + PWM valuesolution.

of byFinally,

the three

the systems, which

fast response enables

of the

the correct

SMCoperation of the

is translated intoPa& O + PWM

higher energysolution.

harvestedFinally,

from the thePVfast response

array, whichof the SMCincreases

eventually is translated

the into

system profitability.

a higher energy harvested from the PV array, which eventually increases the system profitability.

20

vpv[ V]

18.5 17

10 18 16

17.5 15

0.4024 0.4026 0.4028 0.403 0.4032 0.7 0.702 0.704 0.706 0.708 0.71

0

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

100

18

85.5

ppv [W]

16

85 14

50 12

84.5 0.65 0.652 0.654 0.656

0.41 0.411 0.412 0.413 0.414 0.415

0

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

200

vb [V]

150

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Time [s]

Figure 19. Performance comparison between the SMC and conventional P & O-based solutions

Figure 19. Performance comparison between the SMC and conventional P & O-based

considering small DC-link voltage oscillations.

solutions considering small DC-link voltage oscillations.

12384

Energies 2015, 8

Energies 2015, 8, 12363–12387

25

A second test was performed accounting for large voltage oscillations at the DC-link, i.e., using a

A second test was performed accounting for large voltage oscillations at the DC-link, i.e., using

smallera smaller14

Cb F.

“ The simulation

14 µF. results are

The simulation reported

results in Figure

are reported 20, where

in Figure again again

20, where the SMC exhibits a

the SMC

faster exhibits

MPPT aprocedure

faster MPPTwith higherwith

procedure energy

higherproduction. Moreover,

energy production. the simplest

Moreover, P & PO&solution

the simplest O

solution (P & O + PWM) is unstable since the DC-link voltage oscillations are transferred

(P & O + PWM) is unstable since the DC-link voltage oscillations are transferred to the PV array to the PV

array terminals, which confuses the P & O algorithm. In contrast, the more complex P & O solution

terminals, which confuses the P & O algorithm. In contrast, the more complex P & O solution

(P & O + PID + PWM) is stable, but it requires a long time to reach the MPP in comparison with

(P & O + SMC.

the PID + PWM) is stable, but it requires a long time to reach the MPP in comparison with the SMC.

25

vpv [V]

20

15

SMC P&O+PWM P&O+PID+PWM MPP

10

0.635 0.64 0.645 0.65 0.655 0.66

80 14.92

ppv [W]

60 14.91

40

20 0.6591 0.6592 0.6593 0.6594 0.6595

0

0.635 0.64 0.645 0.65 0.655 0.66

300

vb [V]

200

Time [s]

Figure 20. Performance comparison between the SMC and conventional P & O-based solutions

Figure 20. Performance comparison between the SMC and conventional P & O-based

considering large DC-link voltage oscillations.

solutions considering large DC-link voltage oscillations.

6. Conclusions

6. Conclusions

A detailed analysis and design of a MPPT solution based on sliding-mode control theory has

been introduced in this paper. This solution is aimed at performing a fast MPPT action on PV systems

A detailed analysis and design of a MPPT solution based on sliding-mode control theory has been

using a single control stage. Moreover, since it is based on a non-linear model, no linearization process

introduced in thisTherefore,

is needed. paper. This thissolution

approachis aimed

avoids at theperforming a fast MPPT

circular dependency action

among on PVcontrollers,

cascade systems using

a single control

reduces the stage.

number Moreover,

of controlsince it isand

devices based on a global

provides non-linear model,

stability in allnothelinearization process is

operation range.

needed. Therefore, this approach avoids the circular dependency among cascade controllers,and

Such characteristics are major improvements over classical MPPT solutions based on cascade reduces

linear controllers.

the numberThe of performance

control devices and provides global stability in all the operation range. Such characteristics

of the proposed solution was tested using realistic simulations accounting for

are major improvements

perturbations in bothover classical MPPT

the irradiance and thesolutions based

load voltage, on cascade

obtaining and linear

satisfactory controllers.

results. Moreover, the

The performance of the proposed solution was tested using realistic simulations

sliding-mode controller was contrasted with classical MPPT solutions based on the P & accounting for

O algorithm,

where the SMC provides a much faster tracking of the MPP, it increasing the produced energy.

perturbations in both the irradiance and the load voltage, obtaining satisfactory results. Moreover, the

Therefore, the proposed SMC enables to increment the profitability of PV installations; hence the

sliding-mode controller was

return-of-investment timecontrasted

is reduced. with classical MPPT solutions based on the P & O algorithm,

where theFinally,

SMC aprovides a much faster

further improvement to thetracking

SMC could ofbethe MPP, itin increasing

performed a future work—a the produced energy.

different, but

Therefore, the proposed

more complex, switchingSMC enables

circuit could tobe increment

designed to the

avoidprofitability of PV

the sign change of the installations; hence the

transversality—to

remove the divergence stages in the SMC operation. This new development will enable to constraint

return-of-investment time is reduced.

the hysteresis band of the SMC, which will eventually enable to reduce, even more, the settling time

Finally, a further

of the PV power improvement

to achieve a faster to the

MPPT SMC could be performed in a future work—a different, but

procedure.

more complex, switching circuit could be designed to avoid the sign change of the transversality—to

Acknowledgments: This work was supported by the Instituto Tecnológico Metropolitano, the Universidad

remove the divergence

Nacional de Colombiastages in the SMC

and Colciencias operation.

(Fondo nacional This new development

de financiamiento will laenable

para ciencia, to constraint

tecnología y la

the hysteresis

innovaciónband of the SMC, which will eventually enable to reduce, even more, the settlingand

Francisco José de Caldas) under the projects MicroRENIZ-25439 (Code 1118-669-46197), P14215

P14220, and under the study commission-agreement number 25 from Instituto Tecnológico Metropolitano.

time of

the PVAuthor

power to achieve a faster MPPT procedure.

Contributions: All the authors contributed equally to the work.

12385

Energies 2015, 8, 12363–12387

References

1. Lewis, N.S.; Crabtree, G.; Nozik, A.J.; Wasielewski, M.R.; Alivisatos, P. Basic Research Needs for Solar

Energy Utilization. In Basic Research Needs for Solar Energy Utilization; Department of Energy: Washington,

DC, USA, 2005.

2. Kish, G.J.; Lee, J.J.; Lehn, P.W. Modelling and control of photovoltaic panels utilising the incremental

conductance method for maximum power point tracking. IET Renew. Power Gener. 2012, 6, 259–266.

[CrossRef]

3. Femia, N.; Petrone, G.; Spagnuolo, G.; Vitelli, M. A Technique for Improving P & O MPPT Performances of

Double-Stage Grid-Connected Photovoltaic Systems. IEEE Trans. Ind. Electron. 2009, 56, 4473–4482.

4. Esram, T.; Chapman, P.L. Comparison of Photovoltaic Array Maximum Power Point Tracking Techniques.

IEEE Trans. Energy Convers. 2007, 22, 439–449. [CrossRef]

5. Gonzalez, D.; Ramos-Paja, C.A.; Petrone, G. Automated Procedure for Calculating the Controller

Parameters in Photovoltaic DC/DC Converters. Int. Rev. Electr. Eng. 2011, 6, 3027–3040.

6. Adinolfi, G.; Femia, N.; Petrone, G.; Spagnuolo, G.; Vitelli, M. Design of DC/DC Converters for DMPPT

PV Applications Based on the Concept of Energetic Efficiency. J. Sol. Energy Eng. Trans. ASME 2010, 132.

[CrossRef]

7. Qi, J.; Zhang, Y.; Chen, Y. Modelling and maximum power point tracking (MPPT) method for PV array

under partial shade conditions. Renew. Energy 2014, 66, 337–345. [CrossRef]

8. Petrone, G.; Ramos-Paja, C.A.; Spagnuolo, G.; Vitelli, M. Granular control of photovoltaic arrays by means

of a multi-output Maximum Power Point Tracking algorithm. Prog. Photovolt. Appl. 2012, 21, 918–932.

[CrossRef]

9. Ahmad, J. A fractional open circuit voltage based maximum power point tracker for photovoltaic arrays.

In Proceedings of 2nd International Conference on Software Technology and Engineering (ICSTE), San

Juan, Puerto Rico, 3–5 October 2010; pp. 247–250.

10. Ahmed, J.; Salam, Z. A Maximum Power Point Tracking (MPPT) for PV system using Cuckoo Search with

partial shading capability. Appl. Energy 2014, 119, 118–130. [CrossRef]

11. Khaehintung, N.; Kunakorn, A.; Sirisuk, P. A novel fuzzy logic control technique tuned by particle swarm

optimization for maximum power point tracking for a photovoltaic system using a current-mode boost

converter with bifurcation control. Int. J. Control Autom. Syst. 2010, 8, 289–300. [CrossRef]

12. Lin, W.-M.; Hong, C.-M.; Chen, C.-H. Neural-Network-Based MPPT Control of a Stand-Alone Hybrid

Power Generation System. IEEE Trans. Power Electron. 2011, 26, 3571–3581. [CrossRef]

13. Chiu, C.-S.; Ouyang, Y.-L. Robust Maximum Power Tracking Control of Uncertain Photovoltaic Systems:

A Unified T-S Fuzzy Model-Based Approach. IEEE Trans. Control Syst. Technol. 2011, 19, 1516–1526.

[CrossRef]

14. Petrone, G.; Spagnuolo, G.; Teodorescu, R.; Veerachary, M.; Vitelli, M. Reliability Issues in Photovoltaic

Power Processing Systems. IEEE Trans. Ind. Electron. 2008, 55, 2569–2580. [CrossRef]

15. Zitouni, N.; Khiari, B.; Andoulsi, R.; Sellami, A.; Mami, A.; Hssen, A. Modelling and non linear control of a

photovoltaic system with storage batteries: A bond graph approach. IJCSNS Int. J. Comput. Sci. Netw. Secur.

2001, 11, 105–114.

16. Bianconi, E.; Calvente, J.; Giral, R.; Mamarelis, E.; Petrone, G.; Ramos-Paja, C.A.; Spagnuolo, G.; Vitelli, M.

Perturb and Observe MPPT algorithm with a current controller based on the sliding mode. Int. J. Electr.

Power Energy Syst. 2013, 44, 346–356. [CrossRef]

17. Bianconi, E.; Calvente, J.; Giral, R.; Mamarelis, E.; Petrone, G.; Ramos-paja, C.A.; Spagnuolo, G.; Vitelli, M.

A Fast Current-Based MPPT Technique Employing Sliding Mode Control. IEEE Trans. Ind. Electron. 2013,

60, 1168–1178. [CrossRef]

18. Ghazanfari, J.; Farsangi, M.M. Maximum Power Point Tracking using Sliding Mode Control for

Photovoltaic Array. Iran. J. Electr. Electron. Eng. 2013, 9, 189–196.

19. Khemiri, N.; Khedher, A.; Mimouni, M.F. A sliding mode control approach applied to a photovoltaic

system operated in MPPT. In Proceedings of Systems, Signals & Devices (SSD), 2013 10th International

Multi-Conference, Hammamet, Tunisia, 18–21 March 2013; pp. 1–6.

12386

Energies 2015, 8, 12363–12387

20. Afghoul, H.; Chikouche, D.; Krim, F.; Beddar, A. A novel implementation of MPPT sliding mode controller

for PV generation systems. In Proceedings of EUROCON, 2013 IEEE, Zagreb, Croatia, 1–4 July 2013;

pp. 789–794.

21. Chu, C.-C.; Chen, H.-L. Robust maximum power point tracking method for photovoltaic cells: A sliding

mode control approach. Sol. Energy 2009, 83, 1370–1378. [CrossRef]

22. Romero-Cadaval, E.; Spagnuolo, G.; Garcia Franquelo, L.; Ramos-Paja, C.A.; Suntio, T.; Xiao, W.M.

Grid-Connected Photovoltaic Generation Plants: Components and Operation. IEEE Ind. Electron. Mag.

2013, 7, 6–20. [CrossRef]

23. Zhang, F.; Maddy, J.; Premier, G.; Guwy, A. Novel current sensing photovoltaic maximum power point

tracking based on sliding mode control strategy. Sol. Energy 2014, 118, 80–86. [CrossRef]

24. Petrone, G.; Spagnuolo, G.; Vitelli, M. A Multivariable Perturb-and-Observe Maximum Power Point

Tracking Technique Applied to a Single-Stage Photovoltaic Inverter. IEEE Trans. Ind. Electron. 2011, 58,

76–84. [CrossRef]

25. Trejos, A.; Gonzalez, D.; Ramos-Paja, C.A. Modeling of Step-up Grid-Connected Photovoltaic Systems for

Control Purposes. Energies 2012, 5, 1900–1926. [CrossRef]

26. Manganiello, P.; Ricco, M.; Petrone, G.; Monmasson, E.; Spagnuolo, G. Optimization of Perturbative PV

MPPT Methods Through Online System Identification. IEEE Trans. Ind. Electron. 2014, 61, 6812–6821.

[CrossRef]

27. Renaudineau, H.; Donatantonio, F.; Fontchastagner, J.; Petrone, G.; Spagnuolo, G.; Martin, J.-P.;

Pierfederici, S. A PSO-Based Global MPPT Technique for Distributed PV Power Generation. IEEE Trans.

Ind. Electron. 2015, 62, 1047–1058. [CrossRef]

28. Mamarelis, E.; Petrone, G.; Spagnuolo, G. An Hybrid Digital-Analog Sliding Mode Controller for

Photovoltaic Applications. IEEE Trans. Ind. Inf. 2103, 9, 1094–1103. [CrossRef]

29. Haroun, R.; Aroudi, A.E.; Cid-Pastor, A.; Garcia, G.; Olalla, C.; Martinez-Salamero, L. Impedance Matching

in Photovoltaic Systems Using Cascaded Boost Converters and Sliding-Mode Control. IEEE Trans.

Power Electron. 2015, 30, 3185–3199. [CrossRef]

30. Erickson, R.W.; Maksimovic, D. Fundamentals of Power Electronics, 2nd ed.; Springer: Berlin, Germany, 2007.

31. Sira-Ramirez, H.T. Sliding motions in bilinear switched networks. IEEE Trans. Circuits Syst. 1987, 38,

919–933. [CrossRef]

32. Gonzalez Montoya, D.; Ramos-Paja, C.A.; Giral, R. Improved design of sliding mode controllers based on

the requirements of MPPT techniques. IEEE Trans. Power Electron. 2015, 31, 235–247. [CrossRef]

33. Tan, S.-C.; Lai, Y.M.; Tse, C.K. General Design Issues of Sliding-Mode Controllers in DC/DC Converters.

IEEE Trans. Ind. Electron. 2008, 55, 1160–1174.

34. Park, J.-H.; Jeong, H.-G.; Lee, K.-B. Output Current Ripple Reduction Algorithms for Home Energy Storage

Systems. Energies 2013, 6, 5552–5569. [CrossRef]

35. Venturini, R.P.; Scarpa, V.V.R.; Spiazzi, G.; Buso, S. Analysis of limit cycle oscillations in maximum power

point tracking algorithms. In Proceedings of Power Electronics Specialists Conference (PESC 2008), Rhodes,

Greece, 15–19 June 2008; pp. 378–384.

36. Veberič, D. Lambert W function for applications in physics. Comput. Phys. Commun. 2012, 183, 2622–2628.

[CrossRef]

37. Kuo, B.; Golnaraghi, F. Automatic Control Systems, 7th ed.; Prentice-Hall: New Jersy, NJ, USA, 2002;

pp. 398–401.

38. Femia, N.; Petrone, G.; Spagnuolo, G.; Vitelli, M. Optimization of Perturb and Observe Maximum Power

Point Tracking Method. IEEE Trans. Power Electron. 2005, 20, 963–973. [CrossRef]

39. Ortiz-Valencia, P.A.; Trejos-Grisales, L.A.; Ramos-Paja, C.A. Maximum power point tracking in PV systems

based on adaptive control and sliding mode control. Rev. Fac. Ing. Univ. Antioq. 2015, 75, 67–79.

© 2015 by the authors; licensee MDPI, Basel, Switzerland. This article is an open

access article distributed under the terms and conditions of the Creative Commons by

Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).

12387

- Grid Connected PV System with CSIDiunggah olehijeteeditor
- FLXPRO17-DEN1425WDiunggah olehmuhammad.hafiz
- Final Ashish ABSTRACT 1Diunggah olehSamuel Davis
- 200 Icrera2013 SpainDiunggah olehMarian Enachescu
- 1KW DCC-AC InverterDiunggah olehLixo4618
- A SOLAR POWER PRODUCTION SYSTEM USING MPPT WITH GRID SYNCHRONIZATION FOR DISTRIBUTED POWER PRODUCTION AND DISTRIBUTIONDiunggah olehTJPRC Publications
- VSC Current ControlDiunggah olehElectricEngineering
- A Modified Adaptive Hill Climbing Maximum Power Point Tracking (MPPT) Control Method for Photovoltaic Power SystemsDiunggah olehCesar Augusto Corazao Peralta
- Arduino Solar Charge Controller Version 30Diunggah olehMarius Danila
- Brochure-Energy-Storage.pdfDiunggah olehaderitodomingos1531
- My Published MPPT Paper RSERDiunggah olehSwati Kar
- 53-759Diunggah olehMk Pandiyan
- 9 Basic DC to DC Converter CircuitsDiunggah olehuser123phc
- ReportDiunggah oleh13121977
- Flx pro 15 - 17Diunggah olehSam Al-Mimar
- Final - Ieee 3013 Power Elec. Project TitlesDiunggah olehAnonymous QIuAGIadXm
- Proposal for 7 HP solar pumping systemDiunggah olehVamsi
- ic5_engDiunggah oleheric hernandez
- Inverter DC to ACDiunggah olehZponthanna Made
- Inverter Toshiba VF-A7 ManualDiunggah olehsunhuynh
- 6_1_5_12Diunggah olehyong_fan
- an66f.pdfDiunggah olehMarco Perez
- IVT_Solartechnik_Prospekt2017-ENG-MAIL.pdfDiunggah olehBogdan Ilie
- Simulación fotovoltaicaDiunggah olehJediKnight012
- Efficient Single Phase Transformerless Inverter ForGrid-Tied PVG System With Reactive Power Control - 2016Diunggah olehmtechprojects
- CBC915Diunggah olehachew911
- 2015 - 2016 Ieee Power Electronics TitlesDiunggah olehI3E Technologies
- REFF1Diunggah olehMahendiran Cr
- PBL Report CompleteDiunggah olehShahfizan Nasriq Ali
- Digital Control of a Single-Phase UPS Inverter for Robust AC-Voltage TrackingDiunggah olehPhạm Văn Tưởng

- Daniel Golani, Brenda Appelbaum-Golani-Fish Invasions of the Mediterranean Sea(2010)Diunggah olehCorina Taraș-Lungu
- Intracellular Whole-Cell Patch-Clamp Recordings of Cortical Neurons in Awake Head-Restrained MiceDiunggah olehcorestore
- User Manual ADU-555CDiunggah olehromaditya
- Archaeology Anthropology and Interstellar Communication TAGGEDDiunggah olehJames Quinn
- Pacto por la Cultura Thesis.pdfDiunggah olehKDelpaso
- peebles2003.pdfDiunggah olehRuskin Patel
- Hovercraft GuidebookDiunggah olehgitaanjali
- The Effect of Market Orientation to Achieve Competitive AdvantageDiunggah olehThe Ijbmt
- Proline XC Technical SpecificationsDiunggah olehLuis Alberto Díaz Olmedo
- Mantak ChiaDiunggah olehMohammedFaizaanFaraz
- Percona Toolkit 2.2.12Diunggah oleholragon
- MELJUN IT102 Syllabus Computer Fundamental and ProgrammingDiunggah olehMELJUN CORTES, MBA,MPA
- Linear ProgrammingDiunggah olehHenra
- IRJET-Fuzzy Logic based Expert System for Detecting Colorectal CancerDiunggah olehIRJET Journal
- 7VK512x Catalog SIP2004 EnDiunggah olehHaj Omid
- Crisis Management -Crisis CommunicationsDiunggah olehFarah Bulsara
- Sustainable Concrete PavementDiunggah olehGustavo Tonelli
- Graphomotor Activities.20140727.120415Diunggah olehpoint3hook
- IndexDiunggah olehAUGUSTUS11
- Board Of Directors Annual EvaluationDiunggah olehJohn doe
- bkS Std X B FDiunggah olehdeepashaji
- Depliant - Solution _ Air End [GB]Diunggah olehgeorge
- A Bridge Not Too Far - From Design Thinking to Integral DesignDiunggah olehSunil Malhotra
- Bump Hunting AlgorithmDiunggah olehvarun
- g8Diunggah olehFemmy Andrifianie
- README.txtDiunggah olehJack Sim
- Function of the Human BrainDiunggah olehErdős Alpár
- Losh HackingaristotleDiunggah olehTatenda Gatawa
- Implemenration of an Eulerian Multi Phase Model in OpenFOAM and Its Application to Polydisperse Two Phase FlowsDiunggah olehAghajani
- Contract Bursa Privata OMV PetromDiunggah olehregno7