PhotoVoltaic Cost Reduction, Reliability,

Operational performance, Prediction and

Simulation.

Title:

Modeling and sizing of large PV-diesel hybrid systems without energy

storage.

Authors:

J. Muoz, J.M. Carrillo.

MODELING AND SIZING OF LARGE PV-DIESEL HYBRID SYSTEMS WITHOUT ENERGY STORAGE
J. Muoz(1), J.M. Carrillo.
Grupo de Sistemas Fotovoltaicos. Instituto de Energa Solar Universidad Politcnica de Madrid (IES-UPM)
(1)
Address: ETS Ingeniera y Diseo Industrial (ETSIDI). Ronda de Valencia, 3. 28012 Madrid (Spain)
E-mail: [email protected]
Orcid ID: orcid.org/0000-0002-9301-9215
ABSTRACT: This paper describes the modelling of components and control strategies for PV-diesel hybrid systems
without energy storage, which have been implemented in an online and free-software simulator of PV systems called
SISIFO, which is publicly available at www.sisifo.info. Among other features, this software tool allows the
prediction of fuel savings for different sizes and types of PV generators, the comparison of selected control strategies
and the evaluation of the spinning reserve requirements in systems with multiple diesel generators operating in
parallel.
Keywords: PV System, Hybrid, Software, Modelling, Simulation

INTRODUCTION

The reduction of cost for PV modules caused by the

massive installation of grid-connected systems has
increased the economic competitiveness of conventional
decentralised PV applications. In particular, there is a
renewed interest in PV-diesel hybrid systems, which are
the object of this paper, and water pumping systems,
which are discussed in a concurrent paper of this
conference [1].
Today, largest energy systems for remote village
electrification and industrial loads typically rely on
conventional diesel generator sets (gensets) sized in the
MW range. The addition of PV generators to these standalone gensets allows the reduction of fuel consumption
and the improvement of genset operation, efficiency and
reliability of the electricity service.
There are different configurations for coupling PV
and diesel gensets [2]. The most common solution
includes energy storage, normally a battery bank, which
can be used for different purposes. For example, storing
the surplus of PV power or adding its power to that of the
gensets for meeting peak load demands, which may
increase the potential for fuel saving.
Nevertheless, the inclusion of batteries also increases
the cost of the hybrid system and the complexity of
control schemes. For these reasons, some authors have
presented experimental and economical studies of hybrid
systems without energy storage for rural village
electrification [3][4]. The simplicity of systems without
storage and the reduced investment, operation and
maintenance costs are advantages when the retrofitting of
already existing diesel systems is concerned. In
particular, for remote industrial applications in which the
load is nearly constant or peak load demand occurs
during daylight hours.
This paper describes the modelling of components
and control strategies for PV-diesel hybrid systems
without energy storage, which have been integrated in an
online and free-software simulator of PV systems called
SISIFO (available at www.sisifo.info).
This simulation tool allows the prediction for fuel
savings as function of PV size and type, which is required
to asses if the hybrid system is cost-effective, i.e., if fuel
saving at least compensates initial investment, operation
and maintenance costs. Besides, the tool allows the
comparison of selected control strategies in terms of their
energy performance and the evaluation of the spinning
reserve. Simulations are performed in steady-state

conditions, which assume that any system variable is

constant during each time step, which may be varied
from minutes to hours.
The presentation of this paper has been organised
with the following structure. Section 2 describes the
modeling of system components, focusing in diesel
gensets. Section 3 describes the implemented control
strategies for a single genset and multiple gensets
operating in parallel. Finally, Section 4 presents a
simulation example for a real case of study.

HYBRID SYSTEM MODELING

The configuration of the simulated PV-diesel hybrid

system without energy storage is displayed in Figure 1.
The diesel genset may be composed by a single or
multiple units (up to N in the figure), which create the
low-voltage AC grid that supplies the load. To this grid,
PV generators are connected in parallel through standard
DC/AC converters, or inverters, designed for
conventional grid-connected PV applications.

PV
generator

Inverter

Load

Genset 1 Genset 2

G
Genset N

Figure 1: Configuration of a PV-diesel hybrid system

without energy storage.
The detailed modeling of the grid-connected PV
system is beyond the scope of this paper, but is described
elsewhere by the authors [5][6], where the interested
reader is referred. It is just worth mentioning that SISIFO
allows the simulation of three static and six sun-tracking
PV generators. For example, static ground-mounted or
building-integrated PV generators in roofs or faades,
and one-axis horizontal or two-axes sun-trackers. Suntrackers generate more energy at higher equipment cost
and some of them may be cost-effective in particular
energy scenarios at present PV modules prices.
We focus here on the modeling and characterisation
of gensets and load profiles. For simplicity, it is assumed
that all gensets of the hybrid system are identical. Each

225

140

220

120

215

100

210

80

205

BFSC

60

200

FC

40

195

20

190

Fuelconsumption [l/h]

BFSC[g/kWh]

2.1 Engine model

Modeling of engines, as well other system
components, is based on information provided by
manufacturers in their datasheets.
Fuel consumption is calculated from the Brake Fuel
Specific Consumption (BFSC), which indicates the
amount of fuel consumed by the engine per unit of
energy as a function of the brake output power. The
BSFC is given in [gr/kWh], or [l/kWh], and it is normally
provided by the manufacturer by means of figures or
tables (see Figure 2). Normally, BFSC increases at low
brake power levels. Hence, as the available PV power
increases, the brake power of the engine decreases but at
a higher BSFC, i.e., engines consume less fuel but they
are some less efficient.

0
0

100

200
300
400
Brake power [kW]

500

600

Figure 2: Variation of the the Brake Fuel Specific

Consumption (BFSC) and fuel consumption (FC) as a
function of the brake output power of a rated 575kW,
constant speed, commercial engine.
Fuel consumption is calculated as the product of
BSFC and brake power, PB, whose result is a quasi-linear
relation (see Figure 2) that can be approximated by
Skarstein and Uhlen model [7]:

FC = aPB2 + bPB + cPBR

(1)

Where FC is the fuel consumption, in [l/h], PBR is the

rated brake power, in [kW], and a, b and c are constants
to be fitted for each particular engine. The quadratic term
for PB was not in the original model and has been added
to improve the fitting of fuel consumption for variable
speed diesel gensets [8].
In the example of Figure 2, the rated brake power is
PBR=575kW and fuel consumption may be fitted with
Equation 1 to obtain: a=0, b=0,217 l/kWh, c=0,012
l/kWh and R2=0,9995. Also from Figure 2, it can be
observed that the fuel consumption is not zero at the
origin, i.e., with the engine operating at no load. For the
previous example, this no-load fuel consumption is 6,9
l/h.
2.2 Alternator model
The alternator is characterized by its rated output
power (PAR) and three experimental parameters (k0, k1
and k2), which are used to calculate the power conversion
efficiency, A, by means the following equation:

A =

P0
p0
=
PB
p 0 + ( k 0 + k1 p 0 + k 2 p 02 )

(2)

Where p0=P0/PAR being P0 the output AC power of

the alternator and PB the brake engine power. Parameters
k0, k1 and k2 are fitted from the power efficiency curve
provided by the manufacturer.
Starting from output alternator power, P0, the
program calculates the brake power, which, together
BSFC, allows the calculation of the instantaneous fuel
consumption using Equation 1.
2.3 Power rating
The power rating of the genset is indicated by the
prime power (PRP), which is defined as the maximum
output power that can deliver to a variable load and for an
unlimited number of hours [9]. Besides this rated power,
two other parameters, overload and minimum powers, are
required as input data.
Overload, or maximum power, PMAX, indicates the
engine capacity to exceed the prime power, usually
expressed as percentage of the latter and together with the
overload duration. For example, a 110% of the prime
power for a period of 1h is usually stated [10]. For some
gensets, this overload power is only used for control
purposes and the maximum power is limited to the PRP.
In these cases, manufacturers discourage to use this
overload capacity to supply the load and simulations
should be performed with PMAX =PRP.
Finally, for efficiency and reliability reasons, the
engine should operate over the minimum power, PMIN,
which typically ranges 30-40% of rated capacity.
Operating the engine below this level not only decreases
the conversion efficiency, but also this may short the
engines lifetime and increase maintenance costs.
Loadpower [kW]

genset is composed by an engine, or motor, and an

alternating current generator, or alternator, which are
modelled as follows.

500
400
300
200
100
0
0

12
16
Hour

20

24

Figure 3: Example of a real load profile from a remote

petrol exploitation field.
2.4 Load profile
The load profile may be specified by a times series of
power consumption, with any time step, during the period
of analysis (typically a year). It is worth mentioning that,
despite time step can be varied from seconds to hours,
simulations are performed assuming steady-state
conditions, in which the above described models can be
used.
If a yearly time series of load demand is not
available, which is usually the case, the user may specify
either a daily load profile that repeats everyday or the
simplest case of constant power consumption. Figure 3
shows an example a real load profile from a petrol
exploitation field, which will be used for the simulation
example described in Section 4.
Since the hybrid systems discussed here have not
energy storage, PV power is directly used to supply the
load during sunshine hours. Hence, the matching between
load consumption and solar generation profiles has a

CONTROL STRATEGIES

The design of the hybrid system requires the selection

of a control strategy to guarantee the reliability and
quality of the electricity service, minimizing the fuel
consumption and maximizing the lifetime of system
components.
SISIFO includes typical control strategies for
regulating the contribution of PV power to the hybrid
system, which are described in this section. But, before
introducing them, let us describe the implemented control
when only the gensets are in operation. SISIFO offers the
simulation of systems composed of a single genset and
also multiple gensets operating in parallel.
Table I summarises previously defined symbols and
other introduced below, which are intensively used
hereinafter in discussion.

starts/stops the gensets when the load power crosses a

certain percentage point of the total rated power, which is
usually called set-point.
The number of set-points is equal to the number of
gensets in the system less one. For example, for a system
with two gensets there is only a set-point at 50%. For
three gensets, two set-points at 33% (1/3) and 67% (2/3)
etc, and so on.
The starting/stopping of gensets are not performed
exactly at set-points but within a hysteresis band around
them. Figure 4 displays and example of system operation
for three gensets with ON/OFF control and hysteresis
band.
Hysteresis band
ON

3
Genset

great impact on fuel savings. For example, PV power

should be sometimes limited to avoid the operation of
diesel gensets below their minimum power.

OFF
ON

OFF
ON

Table I: Summary of symbols.

Symbol
Definition
NGEN
Required number of gensets in operation
NGEN. total Total number of gensets in the system
Load power
PL
PMAX
Maximum genset power
Minimum genset power
PMIN
PNET
Net load (PL- PPV)
Optimal genset power
POPT
PPV
PV power
Nominal PV power
PPV,STC
PRP
Prime power
3.1 Diesel only (single genset)
The single genset is sized to match the peak load
demand and it is permitted to operate from PMIN to PMAX.
The value for PMIN may be selected within a range of 0%
to 50% of PRP.
If PMIN is non-zero and the load power is lower than
this minimum, some corrective actions should be applied
following the recommendations of genset manufacturer.
Unless otherwise stated, the actions considered in the
simulation tool are either permitting genset operation
below PMIN or connecting a dump load to ensure the
genset operates at PMIN. Hence, a variable dump power
equal to (PMIN-PL) is connected whenever PL< PMIN.
3.2 Diesel only (multiple gensets)
The operation of multiple gensets in parallel are
controlled using two common strategies called here
ON/OFF and Optimal. Both strategies allow the
disconnection/reconnection of individual gensets to meet
the load demand and taking into consideration the
required spinning reserve.
The spinning reserve, which is specified as a
percentage of the present load demand, indicates the
surplus of power that could be supplied immediately by
the gensets in case of a sudden increase of load demand.
This way, the disconnection of gensets is only permitted
if the aggregated rated power of those that remain in
operation is equal of higher than the required spinning
reserve.
In the ON/OFF strategy, the gensets in operation
share the same portion of the load. The system controller

OFF

33

67

100

Load [%]

Figure 4: Operation of a system with three gensets with

ON/OFF control as a function of load demand. It is
assumed that the disconnection of any genset guarantees
the spinning reserve.
The Optimal strategy tries to ensure that all the
gensets, with the exception of one, operate at much as
possible at their optimum power operating point, POPT,
where the fuel consumption is minimal. This optimal
point is around 80% of the rated power but any valued
can be chosen. The optimal control algorithm has the
following steps:
1. The first calculation is to determine the required
number of gensets in operation, NGEN, which is equal to:
N GEN = Ceiling ( PL / POPT )

(3)

2. If only one genset is required in operation

(NGEN=1 or PLPOPT), its operation is that described in
section 3.1.
3. If several gensets are required in operation
(NGEN>1), there are two possible cases:
Case 3-a: POPT < PL N GEN ,total POPT
In this case (NGEN-1) gensets may operate at POPT,
and one unbalanced genset operates at PL-POPT (NGEN-1).
Case 3-b: N GEN ,total POPT < PL N GEN ,total PMAX
The load power is higher that the addition of optimal
powers but lower than the maximum power of the
system. Here, we can distinguish two sub-cases.
First, if the surplus (PL-POPTNGEN,total) does not
exceed the maximum power of a single genset,
(NGEN,total-1) gensets may still operate at POPT, and the
unbalanced genset at PL-POPT(NGEN-1), which is within its
range from POPT to PMAX.
Second, if the unbalanced genset reaches its
maximum power, PMAX, and the load increases in this
situation, the balanced gensets must increase its power
over POPT and sharing the same power, which is equal to
(PL-PMAX)/(NGEN-1).

3.3 Diesel (single genset) plus PV

The diesel generator must supply the net load, which
is defined as the difference between the load demand and
the generated PV power:
PNET = PL PPV
(4)
Obviously, the generated PV power reduces the fuel
consumption and the power levels at which the genset
operates. Here, we can distinguish two cases depending
on the load demand. First, if PL is lower than PMIN, the
PV power is cut-off to avoid genset operation below this
minimum. Second, if PL is higher than PMIN the previous
equation applies provided that PNETPMIN. If PNET<PMIN,
the PV power is reduced to ensure that at least PNET=PMIN
to avoid genset operation below this minimum.
3.4 Diesel (multiple gensets) plus PV
The gensets share the net load, PNET, as described in
section 3.2 substituting PL by PNET. Besides, if the
generated PV power, PPV, cause a genset operation below
PMIN the PV power is limited to ensure that such genset
operates with a minimum power equal to PMIN.
SIMULATION EXAMPLE
This section presents a simulation example of a real
industrial application powered at present by a single
diesel genset. Genset characteristics and load demand are
summarised in Table II. Daily load demand is displayed
in Figure 3, where the peak load is 402kW, the average
value is 327kW, and the minimum load is 193kW. The
integration of this profile gives a daily energy
consumption of 7840kW. In yearly figures, this gives an
electricity consumption of 2861MWh that requires a
around 705m3 of diesel. Figure 5 shows the Sankey
diagram for a yearly period, which includes the previous
figures together with the brake energy (2995MWh) and
alternator losses (4,5%).

This section illustrates the simpler application of the

simulation tool, which is the estimation of fuel savings as
a function of the PV power penetration level. The
operation of the system is described in section 3.3
considering that PMIN may vary from 0% to 30% of the
rated power.
For simplicity, a ground-mounted static PV generator
tilted the latitude has been considered, whose nominal
power under Standard Test Conditions (STC), PPV,STC,
may vary from 100kW to 500kW. The nominal power of
the inverter has been assumed to be equal to PPV,STC.
Figure 6 shows the variation of fuel saving as a
function of PPV,STC for different values of PMIN. It can be
observed that for PMIN=0% fuel saving increases nearly
linear with the PV power in the simulated range. And
also that, for a non-zero PMIN, fuel savings deviate
towards to the horizontal from a given PV power, which
limits the PV penetration level. For example, at PMIN
=30%, the fuel saving increases linearly with PV power
up to around 200kW. Beyond this point, the extra power
capacity has a lower impact on the increase of fuel
savings.

Table II: Characteristics of a real application powered by

a single diesel genset.
System characteristic
Value
Engine rated power
PBR=575kW
Fuel consumption
See Figure 2.
Alternator
PAR =525kW
k0=0,02; k1=1e-4; k2=0,023
Prime power
PRP=525kW
Daily load consumption 7840 kWh (see Figure 3).

Figure 5: Sankey diagram of the system with only a

single diesel genset for a yearly period.

40
PMIN=0%

35

30
25

MIN

MIN

=15%
=30%

20
15
10
5
100

150

200

250
300
350
PV power [kWp]

400

450

500

Figure 5: Yearly fuel savings as a function of the

installed PV power, PPV,STC, using the minimum genset
power, PMIN, as parameter.
Besides, the simulator allows a detailed analysis of
the hybrid system behaviour. For example, Figure 6
shows the yearly Sankey diagram (PPV,STC =300kW and
PMIN=30%) and Figure 7, the diurnal variation of PV,
diesel and load powers, during a summer day.

Figure 6: Sankey diagram of the hybrid system, with

PPV,STC =300kW and PMIN=30% , for a yearly period.

Figure 7: Diurnal variation of PV, diesel and load

powers, during a summer day (PPV,STC =300kW and
PMIN=30%).
4

CONCLUSIONS
This paper has presented the modelling of
components and control strategies for PV-diesel hybrid
systems without energy storage, which have been
implemented in SISIFO, an online and free-software
simulator of PV systems that is publicly available at
www.sisifo.info.
Besides, a simulation example of real case study is
described, which illustrates some of the capacities of this
simulation tool.
ACKNOWLEDGEMENT
This work has been possible thanks to the funding of
the FP7 European Programme (Energy) in the project
PhotoVoltaic Cost reduction, Reliability, Operational
performance, Prediction and Simulation (PVCROPS),
Project reference 308468. (www.pvcrops.eu).
REFERENCES
[1]

[2]

[3]

[4]

[5]

J. Muoz, J.M. Carrillo, F. Martnez-Moreno, L.M.

Carrasco, L. Narvarte. Modeling and simulation of
large PV pumping systems. EU PVSEC 2015,
Hamburg.
Wichert, B. PV-Diesel Hybrid Energy Systems for
Remote Area Power Generation A Review of
Current Practice and Future Developments.
Renewable and Sustainable Energy Reviews 1, no.
3 (September 1997): 20928. doi:10.1016/S13640321(97)00006-3.
Lundsager, Per, and Henrik Bindner. A Simple,
Robust &amp; Reliable Wind Diesel Concept for
Remote Power Supply. Renewable Energy 5, no.
14
1994):
62630.
doi:10.1016/09601481(94)90443-X.
Yamegueu, D., Y. Azoumah, X. Py, and N. Zongo.
Experimental Study of Electricity Generation by
Solar PV/diesel Hybrid Systems without Battery
Storage for off-Grid Areas. Renewable Energy 36,
no. 6 (June 2011): 178087.
doi:10.1016/j.renene.2010.11.011.
Muoz, J., Marroyo, L., Collares-Pereira, M.,
Tyutyuyndzhiev, N., Conlon, M., Elmoussaoui, A.,
and Wilkin, B. An Open-Source Simulation Tool of

Grid-Connected PV Systems. 28th European

Photovoltaic Solar Energy Conference and
Exhibition, 2013, 388287.
doi:10.4229/28thEUPVSEC2013-5BV.4.18.
[6] J. Muoz et al. SISIFO: An online simulator of PV
systems. Technical Reference Manual v1.0.
Available at: www.sisifo.info
[7] Skarstein O. Uhlen K. Design considerations with
respect to long-term diesel saving in wind/diesel
plants. Wind Engineering 13, N 2, 1989, 76-87.
[8] Lim, P. Y., and C. V. Nayar. Modeling and
Simulation of Photovoltaic-Variable Speed Diesel
Generator Hybrid Power System for Off-Grid Rural
Electrification. International Journal of Energy
Science,
February
16,
2012.
http://dpijournals.com/index.php/IJES/article/view/260.
[9] ISO
8528-1:2005.
Reciprocating
internal
combustion engine driven alternating current
generating sets. Part 1: Application, ratings and
performance.
[10] ISO
3046-1:2002.
Reciprocating
internal
combustion engines. Performance. Part 1:
Declarations of power, fuel and lubricating oil
consumptions, and test methods. Additional
requirements for engines for general use.