Thermal modeling of the module integrated DC-DC converter

for flexible thin-film PV modules

Milo Aanski, Jelena Popovi-Gerber, Jan Abraham Ferreira
Delft University of Technology
Mekelweg 4
Delft, The Netherlands
Tel.: +31 / (0)15 2781898.
Fax: +31 / (0)15 2782968.
E-Mail: m.acanski@tudelft.nl
URL: http://ewi.tudelft.nl/epp/

Acknowledgements
This project is funded by the Innovatiegerichte Onderzoeksprogramma's (IOP). Authors would like to
thank Nuon Helianthos for their technical support.

Keywords
Modeling, Photovoltaic, Thermal design

Abstract
Recently there is a trend in PV systems toward a more granular processing of the PV array power by
means of distributed maximum power point tracking (DMPPT). This is achieved by connecting a DCDC converter to each PV module in a PV system and thus performing power processing on a module
level as opposed to centralized power processing in traditional PV system architectures. In order to
decrease the manufacturing costs and to improve the level of integration, the converter can be directly
integrated into the PV module. This on the other hand brings tight thermal coupling between the PV
module and the integrated converter and introduces additional heat to the PV module which may
deteriorate its performance. It is therefore important to model the thermal behavior of the system in
order to estimate reached temperatures and to test the effectiveness of different converter thermal
management strategies. This paper presents an investigation into the thermal behavior of a DC-DC
converter integrated into a flexible PV module. The goal of this work is to develop analytical thermal
models of the PV module and the integrated converter that will allow to quickly predict the reached
system temperatures for a set of PV module and converter specifications and boundary conditions.
Finally, the results obtained from the developed thermal models show satisfying accuracy when
compared to the experimental results and the results obtained using computational fluid dynamics
(CFD) simulations.

Introduction
The development of clean energy resources as alternatives to fossil fuels has become one of the most
important challenges for modern science. Among the variety of renewable sources, photovoltaics is
nowadays the fastest growing power generation technology in the world, and large investments and
R&D efforts are being made in order to decrease the cost and to improve the performance, efficiency
and reliability of PV modules and power electronic converters for PV systems. Today, crystalline
silicon technology dominates the PV cell market with around 80% share, but most of the current
research is focused on thin-film PV technology. Thin-film market share has grown steadily over the
past few years, especially after introduction of novel production lines that can produce low cost PV
modules [1]. Currently there are no substantial limits to the widespread deployment of PV. Solar
power will have to become an integral part of built environment, and when that happens, suitable
power electronic infrastructure will be needed.

Fig. 1a shows the basic PV system architecture in use today. In a typical PV installation, modules are
connected together in series and parallel combinations forming a PV array connected to a central or
string inverter which performs maximum power point tracking and DC-AC conversion. Despite
advances in PV cell technology and power electronics, this basic PV system architecture has remained
the same. Nowadays, as the PV energy penetrates into the electricity grid and built environment,
disadvantages of conventional PV system architecture with centralized maximum power point tracking
become more pronounced, especially in small residential and building integrated PV systems. In
systems with centralized power tracking whenever the performance of even one module is impaired,
there is a power loss along the entire PV array and shading, soiling and aging related mismatches
together with irregular maintenance may significantly reduce the system power output [2]. To
overcome this problem, new distributed architectures were proposed where each PV module has its
own maximum power point tracking (MPPT) unit [3]. In that way the power output of each module is
maximized regardless of the performance of other modules connected in the same PV array.

DC/DC

(a)

(b)

Fig. 1: (a) Traditional PV system, (b) PV system with distributed MPP tracking
This paper considers a distributed architecture where each PV module has its own integrated DC-DC
MPPT which routes the optimized DC power to a central inverter over a common DC bus (Fig. 1b).
This effectively decouples the PV module from the rest of the system and eliminates mismatch losses.
A central inverter is retained and it still presents a single point of failure, although, a simplified cost
reduced version that performs only DC-AC conversion without the MPP tracking can be used. To
increase the performance and level of integration and to decrease manufacturing costs, the DC-DC
converter can be optimized for the particular PV module and integrated directly on its back side. This
will also introduce thermal interdependence in the system and change the temperature profiles of both
the PV module and converter. On the one side the module working temperature can go as high as 80C
in hot weather and under poor ventilation [4]. On the other side the integrated converter will introduce
additional heat losses to the module, changing its temperature profile. It is therefore the purpose of
converter thermal management to remove the excess heat without causing negative effects on the
reliability and performance of the PV module.
While it is possible to employ a CFD analysis to determine the system temperature profile, analytical
modeling can result in faster estimations while providing more insight into the system thermal
behavior. This paper presents the thermal analysis of a PV module with a DC-DC converter integrated
on its back side. The goal of this work is to develop analytical thermal models of the PV module and
integrated converter which will allow quick prediction of the system temperatures for a set of
operating conditions and reliable estimation of the effectiveness of different thermal management
strategies.

Module integrated converter concept

A concept of the DC-DC converter integrated into a flexible PV module is illustrated in Fig. 2a. The
back layer of the PV module is used as a substrate instead of a standard PCB in order to reduce the
cost and to improve the level of integration by reducing the number of packaging layers [5]. After the
component placement, the converter can be encapsulated for protection sharing the same back
encapsulation material with the PV module. The mechanical connection between the PV module and
the integrated converter will change the temperature profile of both the converter and PV module.

Since PV modules are low efficiency devices, large part of absorbed light energy has to be dissipated
into the environment. On the one side, with its presence, the converter may block one part of the heat
flow from the PV module to the back side and on the other hand it will add additional heat losses to
the PV module, which can create a high temperature hot-spot on the PV module surface. Since the
area available on the back side of the PV module is relatively large, the converter can use a large area
to spread the heat losses generated in the high power loss density components. Therefore, thermally
enhanced converter construction is desirable not only to allow the heat flow from the PV module to
the environment, but also to remove and spread the heat from the high power loss components.
Flexible PV
module

Integrated
converter

(c)
(a)
(b)
Fig. 2: (a) PV module integrated converter concept, (b) Flexible PV module, (c) Low profile converter
prototype
For the modeling and experimental validation, an amorphous silicon (a-Si) thin-film flexible PV
laminate shown in Fig. 1b was used [6]. The particular PV module measures 6m by 0.3m and has a
total thickness of approximately 1mm. With the efficiency of almost 7%, the maximum output power
of the PV module is 115W under standard test conditions. Since the flexible PV modules are
frameless, they require a supporting structure. In the following analysis different supporting structures
(in terms of thermal characteristics) will be addressed.
A previously designed prototype of a low profile buck-boost converter was considered in the
following analysis (Fig. 2c). The converter operates at relatively high switching frequency (512kHz)
owing to fast Gallium Nitride devices. The maximum efficiency is relatively low (92%) due to
suboptimal design, but nevertheless the converter can still be used to validate the thermal model. The
converter is constructed on a flexible copper clad laminate instead directly on the PV module, but
since the flexible material that carries the copper layer is very thin this does not change the thermal
behavior of the system significantly. The designed inductor (upper half in Fig. 2c) is made out of 5
flexible PCB layers sandwiched between the flexible magnetic layers in order to enhance the
inductance. The maximum thickness of the converter is determined by the height of the designed
planar inductor and measures approximately 2mm. Finally, the whole converter measures
approximately 6cm by 12cm.

Thermal modeling
To date, there have been many papers dealing with module integrated converters design [3, 7, 8], and
recently there is an increasing number of manufacturers offering commercially available solutions for
DMPPT systems. Although the converter is not physically constructed in the PV module, the term
integrated is used in the electrical sense of the word, meaning that each PV module has its own
converter to track the maximum power point. This also means that it is still necessary to connect the
PV module and the converter using additional wiring and also to provide a mounting place for the
converter, which increases the installation costs and can be a problem in tight places, for example in
BIPV applications. Furthermore, the commercially available solutions are designed to be more flexible
in terms of input voltages/currents in order to accommodate a range of different PV modules and this
have a negative impact on the converter electrical performance. Integrating the converter directly into
the PV module allows the design to be optimized for the particular PV module but also introduces
interdependence between the PV module and the converter thermal behavior. This has already been
addressed in previous work in case of a module integrated DC-DC converter [9], and to some extent in

literature in case of module integrated inverters [10, 11]. The aim of the thermal modeling is to
determine the critical system temperatures first by calculating the heat flow inputs and then by
carrying out a balance of heat flow inputs and heat flow outputs.

Energy balance
The generated electrical power of a PV module depends primarily on the solar radiation energy
(irradiation) being absorbed and transformed. To a lesser extent, it also depends on the PV module
operating temperature - as the operating temperature increases, the output power decreases, typically
by 0.2%/C in case of a-Si PV modules. The Suns irradiance is only partially converted into electrical
energy, one part of the module reaching irradiance is reflected back from the PV module surface and
the remaining part continues its way through the front layers of the module until it reaches the active
semiconductor layer. If the module is transparent, transmitted component of the radiation will be also
present. Under realistic conditions, reflective losses are around 10% of the incoming irradiance, in the
case of perpendicular incidence [12]. Furthermore, only one part of the absorbed radiation is converted
to electrical energy, while the rest is converted into heat. The output electrical power of the PV
module is determined by module efficiency and load profile. For typical single-junction a-Si thin-film
modules stabilized efficiency is in the range of 5%-8%. Therefore, the processes and parameters that
determine the energy balance are module reaching irradiance and module temperature, optical
properties of the PV module, photovoltaic conversion efficiency, balance of heat flows and electrical
characteristic of the load (Fig. 3).

PV module losses

Solar
irradiation

Converter losses

Reflected
radiation

Generated heat:
- Conduction heat flow
- Convection heat flow
- Radiation heat flow

Transmitted
Electrical
radiation output power

Fig. 3: Energy balance

Heat flows
The absorbed component of the solar irradiance that is not converted into electrical energy will
generate the heat flow. The PV module exchanges its heat with the surrounding area through three
modes of heat transfer: conduction, convection and radiation. Determining single components of the
heat flow is not a straightforward process since they depend significantly on the surrounding
environmental conditions and the way the module is mounted. Therefore, to predict the PV module
temperature profile, individual heat flows have to be identified and modeled.
The conduction heat flow takes place in the PV module between the parts with different temperatures
and also between the PV module and the converter. The heat generated in the module active layer will
be conducted through the front and back layers to the PV module surface. The conduction heat flow
within the PV module and the converter is represented with an equivalent thermal resistance network
[13, 15], replacing all distributed heat sources with point thermal current sources Q and defining
thermal resistances Rij between nodes i and j in the PV module and the converter thermal network. In
this way, a predefined path for the heat flow is assumed and represented with the thermal resistance:
Rij =

lij

kij ,eq Aij

(1)

Therefore, equivalent thermal resistances between two points are derived and based upon the geometry
(the length lij and cross section Aij of the heat path) and the thermal conductivities k of different
materials (equivalent thermal conductivity kij,eq). In the analyzed converter the thermal conductivities
between nodes in the thermal network are determined by the copper layers in the flexible PCB and the
additional heat spreading copper layer on top of the components.

The convection heat flow takes place on the PV module and converter surface whenever its
temperature is higher than the ambient temperature. This heat flow cannot be expressed explicitly and
therefore has to be computed by iterations or by approximate equations. In following simulations only
natural convection was considered, neglecting the heat dissipation by forced convection (i.e. no wind).
Natural convection equations for the inclined plate with constant heat flux were applied to the PV
module surfaces, using the average surface temperatures [14]. Since the heat transfer coefficient h is a
function of the surface temperature, which can be obtained only after the calculation of the heat
transfer coefficient, an iteration process has to be employed, assuming the initial value for the surface
temperature. In the equivalent thermal model, convection heat flow on the surface with the area A is
then given represented with:
Ri =

1
hi Ai

(2)

The radiation heat flow also takes place on the PV module surface and represents the heat exchange
between the module and its surroundings. In the following analysis, radiation heat flow on a chosen
surface can be represented with the heat source:
4
Q = A s sky (Ts4 Tsky
) + A s gr (Ts4 Tgr4 ) , Tsky = 0.0552Ta1.5 , Tgr = Ta

(3)

where is Stefan-Boltzmann constant, A is the surface area of the considered surface, S is emissivity
of the PV module surface, and Ts, Tsky, Tgr, Ta are the surface, sky, ground and ambient temperatures
respectively. The sky temperature was determined using the approximate equation given by Swinbank
[8], while the ground temperature is assumed to be equal to the ambient temperature. So called view
factors sky and gr are equal to the cosine and sine of the PV module elevation angle and therefore
depend on the PV module orientation.

Thermal models
The tight thermal coupling between the PV module and the integrated converter brings a strong
interdependence between their thermal models and behavior. In normal situations, without the
integrated converter, a simple one-dimensional model may be sufficient to estimate the PV module
operating temperature [8]. Integrating the converter into the PV module introduces additional heat
sources on top of the existing heat losses from the PV module active area. This will create a hot-spot
on the module surface, and change the temperature profile of both the PV module and the converter.
Since there are limits in operating temperatures for the module and converter components it is
necessary to identify the main heat sources and to model the heat transfer processes in order to
estimate operating temperatures. Obviously, the critical part of the system is the location of the PV
module where the converter is positioned, since this is where the most of heat sources are located.
Since the area occupied by the converter is relatively large and, unlike the heat losses in the PV
module, the losses within the converter are not distributed evenly, the assumption of uniform heat flux
distribution may result in lower estimated maximum temperatures, especially if the heat source
components are distributed over a larger area or if the heat spreading is poor. This nonuniformity in
the generated heat flux is also the reason why the system thermal modeling should start from the
converter. Therefore it is important to first identify main individual sources of heat in the converter.
In the following analysis a more detailed thermal model based on the prototype of the converter shown
in Fig. 2c will be used. Fig. 4a shows the basic topology of the buck-boost converter designed for the
distributed MPPT PV system where the converters are connected in series to form strings which are
further connected in parallel to a central inverter over a common DC bus. Under the normal operating
conditions and the maximum power level the input voltage to the converter is approximately 17V
while the output voltage together with the DC bus voltage is controlled by the central inverter and set
to nominal 48V. Since the converter has a very low profile (compared with its width and length) the

equivalent thermal model is represented as a single layer body, but with different thermal properties in
different directions. Fig. 4b shows the position of the main heat loss sources in the circuit and the
amount of generated losses when the converter is operating under the maximum power level. As
expected, the majority of losses come from the boost stage and the inductor. The marked areas are
represented as nodes in the equivalent thermal resistance network. Adding the internal power losses
the complete thermal model of the converter is illustrated in Fig. 4c.
PC1 PC2
PG1

PL

PAUX

PG2

PCTRL

PL/2
2.05W

PL/2
2.05W

PC1
0.44W
PG1
0.06W
PAUX
0.13W

PC2
3.1W
PG2
0.11W
PCTRL
0.05W

PL/2

PL/2

PC1

PC2

PG1

PG2

PAUX

PCTRL

(a)
(b)
(c)
Fig. 4: Converter thermal modeling: (a) Schematic, (b) Main heat sources, (c) Thermal model
A typical PV module consists of many layers with different thicknesses and thermal properties, but
considering a very low profile of the PV module compared with its width and length, the whole
module can be represented as a single layer body with different directional thermal conductivities. The
obtained converter model is further used to model the thermal behavior of the PV module. For the area
of the module where the converter is mounted the same nodes that were chosen for the converter
thermal resistor network will determine the position of nodes in the PV module thermal network. As
for the rest of the module, remaining areas are represented with only one node each, since their
temperature will always be lower than the hot-spot temperature and therefore are not critical (Fig. 5a).
Since all PV module construction and material details can be entered as model parameters, it is also
easy to adapt the thermal model to different PV modules.
Finally, the supporting structure for the flexible PV module has to be taken into account as well. In the
following analysis two types of supporting surfaces are considered, one with a low thermal
conductivity that corresponds to a worst case mounting conditions and one that allows the convection
heat flow on the back side of the PV module. The supporting structure is modeled using the same
nodes arrangement as for the PV module. To obtain the complete thermal model, first the PV module,
converter and supporting structure thermal models are combined, and after that the heat sources that
represent radiation heat transfer together with thermal resistors that represent convection heat flow
have to be added. This is illustrated in Fig. 5b, with the cross section of the complete thermal model.
Location of the
converter
PV module

Convection

PV module
losses
Converter
losses

Radiation

PV module
Converter

Supporting
surface

(b)
(a)
Fig. 5: (a) PV module thermal model, (b) Cross-section of the complete thermal model

The complete thermal model that includes the PV module and converter thermal network, supporting
structure and ambient conditions defines a set of linear equations with one equation for each node.
This set of equations is formed and can be solved in the same way as the electrical circuits in nodal
analysis. Translated into the linear matrix equation form this system can be represented with:
G1k
k
Q1
T1
G
Q
T
21
2
2
Q = G T , Q = . , T = . , G = .

.
.
.

Qn
Tn
Gn1

G12

. .

2k

. .

. .

.
Gn 2

. .
. .

G1n

G2 n

.
.

k Gnk

(4)

where Q is the vector of heat sources, T is the temperature difference vector and G is the matrix of
thermal conductivities. The previous matrix equation can be implemented as a spreadsheet in
computational software programs like Mathcad or represented as an electrical circuit, transforming the
thermal values into their analog electrical values and using circuit solvers such as SPICE. In either
case the resulting node temperatures can be obtained performing the matrix inversion:
(5)

T = G 1 Q

In the end, several iterations have to be performed until the node temperatures start to converge, since
the convection heat transfer coefficients, radiation heat exchange and the PV module efficiency all
depend on temperature.

Experimental and simulation results

To validate the developed thermal models, the experimental setup shown in Fig. 6a was used. In order
to reach the full level of irradiance (1000W/m2), the tilt angle of the setup was set to approximately
45. This is also a typical mounting angle for roof mounted PV modules at the tested location. For
practical reasons only a part of the PV module was tested (1m by 0.3m), but since the area of interest
is around the converter and thus small in comparison to the PV module surface, it is not necessary to
analyze the whole PV module. The thermal behavior of the system depends significantly on the type
of supporting structure, therefore two different scenarios were considered. In the first scenario the PV
module and the converter were mounted on a surface with a very low thermal conductance (styrofoam
k=0.06 W/mK) which corresponds to the worst case mounting conditions such as well insulated walls.
In this way almost all generated heat is transferred to the front side of the PV module. In the second
scenario, the PV module and the converter are mounted on a thin plastic plate (2mm, k=0.2W/mK),
which allows the heat to flow and convection to take place on the back side of the plate. This scenario
could correspond to the mounting conditions where the PV module is attached to the glass, for
example window. Respective thermal models for both scenarios are shown in Fig. 6b and c.

PV module
Converter
Supporting
surface
(a)
(b)
(c)
Fig. 6: (a) Experimental setup and thermal models of the system with (b) Low and (c) High thermal
conductivity surface

The goal of the thermal models is to predict reached temperatures in the system for different mounting
conditions. Furthermore, it is important to determine where the temperature will first reach its limits.
In the following analysis the PV module operating temperature was limited to 100C while the
maximum operating temperature for the converter components (used integrated circuits, switches,
passives and magnetic material for the inductor) is limited to 125C. As already mentioned, the
converter is operated at the full PV module power with heat losses shown in Fig. 4b. Since only a part
of the PV module is used, the converter is powered via external power supply, while the maximum
power point tracking of the mounted PV module is performed by another DC-DC converter.
The results of thermal measurements for both types of supporting structures are shown in Fig. 7.
Obtained measurement results were also compared to the results obtained using CFD analysis (Ansys).
The ambient temperature was approximately 21C while the irradiation level was slightly above
1000W/m2, which is very close to standard test conditions for PV modules. Under these conditions,
the power losses generated in the PV module are in the range of 80-85mW/cm2. On top of these losses
the PV module has to deal with the additional 110mW/cm2 heat flux coming from the integrated
converter when operated at full power level. It can be seen that in case of the low thermal conductivity
supporting surface the maximum operating temperature for the PV module was already reached before
reaching the maximum power level in the converter components (Fig. 7a). The reason for this is
relatively low converter efficiency of approximately 92% and thus high losses concentrated on a small
area. In the case of a high thermal conductivity surface, the converter can still be operated under the
maximum power level, since the maximum reached temperature is bellow the operating limits of the
PV module (Fig. 7b), but the reached temperature could be even higher if the ambient temperature
increases. Meantime, the component temperatures are below their operating limits with the calculated
maximum temperature of 112C reached in the boost stage, which makes the hot-spot temperature on
the PV module surface the limiting factor in the system design.

Analytical
103 C

Measured
97C

CFD
105C

Analytical
92C

Measured
96C

CFD
99C

(b)
(a)
Fig. 7: Analytical, experimental and CFD simulation results for the setup with supporting surface with
(a) Low and (b) High thermal conductance
In both cases, the results obtained using the analytical model coincide well with the measurement
results and the results obtained using CFD simulations. From the previous results it can be concluded
that the considered converter cannot be safely used for PV module integration without taking the
additional measures. One obvious solution is to improve the converter efficiency which would
decrease the heat flux density added to the PV module. Considering the case from Fig. 7a, the
efficiency should be improved from 92% to 97% in order to keep the hot-spot temperature in safe area.
It can be shown that this would also further decrease the hot spot temperature for the scenario in Fig.
7b by 12C.
Another possibility to decrease the hot-spot temperature is to use the relatively large available surface
area on the PV module back side in order to spread the heat and reduce the heat flux density coming
from the converter. This could be achieved using for example multiphase topologies to distribute the
processed power and spread the converter losses. Alternative simple solution which will be shown in
the following analysis is to use an additional heat spreading layer.

In the following analysis copper layers with two different thicknesses were attached between the PV
module and the converter using the double sided adhesive thermal tape. By changing the in-plane
thermal resistances in the analytical model as the consequence of the introduced conductive layers, it
is possible to estimate the size of the heat spreading layer required to decrease the hot-spot
temperature into the safer area. Fig. 8a shows the calculated maximum reached hot-spot temperature
versus the size of the heat spreading layer for two different heat spreading copper layer thicknesses
(70m and 140m). Fig. 8b illustrates the attached heat spreading layer.
130

Tmax [C]

120
110
100

90
80
70
0

10

11

12

x [cm]
70um,low k

70um, high k

140um, low k

140um, high k

(b)
(a)
Fig. 8: (a) Calculating the required heat spreading layer size, (b) Application of the heat spreading
layer
As it can be seen, for the given heat spreading layer thickness there is an optimum size of the layer
after which further improvements are negligible. Increasing the thickness of the heat spreading layer
will further improve the system thermal behavior but will also add to the total converter thickness and
will reduce the flexibility of the PV module. In the analyzed example, even after applying the thicker
(140 m) heat spreading layer, the overall design still stays fairly flexible.
To test the effectiveness of the heat spreading layer, as predicted by the previous analysis, the
experimental setup was once again tested under the same conditions and for the both surface
conductivity scenarios using the 16x16 cm2, 140m copper layer. Fig. 9 shows the obtained
measurement results and comparison with the analytical model and CFD simulations.

Analytical
98C

Measured
93C

CFD
95C

Analytical
84C

Measured
80C

CFD
81C

(a)
(b)
Fig. 9: Measuring the effectiveness of the heat spreading layer: (a) Low and (b) High thermal
conductance supporting structure
As it can be seen, the achieved maximum hot-spot temperature was decreased by more than 18C in
case of high conductivity supporting surface (Fig. 9b). As for the scenario with low conductivity
surface, the converter is now able to operate under the maximum power level (Fig. 9a), but the reached
hot-spot temperature is very close to the maximum allowed temperature due to the still high converter
heat losses transferred to the ambient through the PV module. Nevertheless, the results obtained using
the analytical thermal model still prove the effectiveness of the additional heat spreading layers. Once
again, the result predicted by the model and Fig. 8a matches well with the measurement and
simulation results.

Conclusion
This paper investigates the thermal behavior of a DC-DC converter integrated into a flexible PV
module used in distributed maximum power point tracking PV systems. Integrating the optimized
converter into the PV module increases the level of integration and may lead to a better system
performance and cost reduction. On the other hand, this impose tight thermal coupling on the system,
putting the converter into harsh environmental conditions and introducing additional heat to the PV
module which may deteriorate its performance. Therefore it is important to model the thermal
behavior of the system in order to estimate reached temperatures and also to test the effectiveness of
different thermal management strategies which will remove the excess heat without causing negative
effects on the reliability and performance of the PV module and the converter.
The goal of this work is to develop thermal models of the PV module and integrated converter which
will allow quick prediction of the system temperatures for a set of PV module and converter
specifications and boundary conditions, for the system within a given environment. Critical system
temperatures are estimated modeling the system by means of equivalent thermal resistance network,
calculating the heat flow inputs and carrying out a balance of heat flow inputs and outputs.
The experimental results for different scenarios are presented and they coincide well with the result
obtained using the analytical model. The results of the analysis show the importance of high converter
efficiency and efficient heat spreading strategies in order to decrease the total introduced heat losses
and heat flux density. The model is also sufficiently flexible, and by changing the model parameters it
can be adapted to different operating conditions. Therefore, the introduced thermal model intend to
presents a step toward reliable integration of the power electronic converter into the PV module

References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]

[11]
[12]
[13]
[14]
[15]

Solar generation 6: Solar photovoltaic electricity empowering the world, European photovoltaic industry
association, 2011
Chouder, S. Silvestre: Analysis model of Mismatch Power Losses in PV Systems, Journal of Solar
Energy Engineering, vol. 131, May 2009
G.R. Walker, P. Sernia: Cascaded DCDC Converter Connection of Photovoltaic Modules, IEEE
Transactions on Power Electronics, Vol. 19, No. 4, July 2004
T. Nordmann, L Clavadetscher: Understanding temperature effects on PV system performance,
Proceedings of 3rd World Conference on Photovoltaic Energy Conversion, 2003
J. Popovic, J.A. Ferreira: An Approach to Deal with Packaging in Power Electronics, IEEE Transactions
on Power Electronics, Vol. 20, No. 3, May 2005
E.A.G. Hamers, M.N. van den Donker, B. Stannowski, R. Schlatmann, G.J. Jongerden: Helianthos: Rollto-Roll Deposition of Flexible Solar Cell Modules, Plasma Processes and Polymers, Vol. 4, p275, 2007
L. Linares, R.W. Erickson: Improved Energy Capture in Series String Photovoltaic via Smart Distributed
Photovoltaics, APEC99, February 2009
G.R. Walker, J.C. Pierce: PhotoVoltaic DC-DC Module Integrated Converter for Novel Cascaded and
Bypass Grid Connection Topologies - Design and Optimisation, PESC, 2006
M. Acanski, J. Popovic-Gerber, J.A. Ferreira: Thermal modeling of the module integrated DC-DC
converter for thin-film PV modules, EPE-PEMC, September 2010
M. Meinhardt,T. O'Donnell, H. Schneider, J. Flannery, C.O. Mathuna, P. Zacharias, T. Krieger:
Miniaturised Low Profile module integrated converter for photovoltaic applications with integrated
magnetic components, APEC99, Vol.1, p. 305, 1999
J. Liu, N. Henze: Reliability consideration of low-power grid-tied inverter for photovoltaic application,
24th European Photovoltaic Solar Energy Conference and Exhibition, Germany, September 2009
S. Krauter: Solar Electric Power Generation, Springer, 2006
D.S. Steinberg: Cooling Techniques for Electronic Equipment, Wiley, 1991
L.M. Jiji: Heat Convection, Second edition, Springer, 2006
J.P. Holman: Heat Transfer, 10th edition, McGraw-Hill, 2010