2016 IEEE Electrical Power and Energy Conference (EPEC)

Dynamic Modeling of Diesel Generator Based on

Electrical and Mechanical Aspects

S. Benhamed1,2*, H. Ibrahim1, K. Belmokhtar1, H. A. Ilinca2, D. Rousse3, A. Chandra3, D. Ramdenee4

Hosni1,2, 2
Université du Québec à Rimouski, [email protected]
1
TechnoCentre éolien, 70 rue bolduc, Gaspé, QC, Canada 3
École de Technologie Supérieure, [email protected],
[email protected], [email protected], [email protected]
[email protected], [email protected] 4
ITMI, Sept-Îles, Canada, [email protected]

Abstract– Nowadays the studies of diesel generators are parameters and they didn’t represent a complete model as well
limited to present mechanical dynamic of the process or as not coupling both mechanical and electrical events. Some of
the electrical one, this is due to the complexity and the high these models are based on an experimental or a numerical
no linearity of the DG. This paper gives a revue of study. However experimental studies is not always feasible as
different model which can describe the total dynamic it is too expensive to instrument a GD and for safety reasons it
process of the diesel generator. Then a developed model is is advisable to always carry a numerical study to meet the
proposed to study the interaction between the mechanical challenges and analyze the dynamic performance of a model.
and electrical aspects in a DG. The validity of the proposed The aim of this paper is to present a proposed complete
model is verified by an application in a study case. The dynamic model that takes into account the mechanical and
developed model has been implemented in electrical aspects of DG developed in real time with Matlab /
Matlab/Simulink, and the simulation result confirms the Simulink. The work presented is part of a research project on
dynamic performances of the system compared with the the development of a dynamic simulator for a wind-diesel
operational data of the DG of TechnoCentre éolien (TCE). hybrid system with compressed air storage for a project
designed to Esker Camp, Canada [8]. The rest of the paper is
Keywords—diesel generator; fuel; diesel engine; dynamic organized as follows:
modeling; speed generator; governor; mechanical & electrical Section 2 presents a brief description of the operation of DG
components; AVR then a review of different types of models of DG implemented
on Matlab / Simulink with a brief analysis of the advantages
I. INTRODUCTION
and limitations of each model. In section 3 the developed
Diesel engines currently find applications in most heavy-load model is presented with the principal blocks of the
mobile and in many stationary power-generation units, mathematical model. In Section 4, a case study was essential
because of their high torque output, size flexibility, durability to validate the performances of the proposed model where the
and fuel efficiency [1]. In fact, in remote communities in results are presented and analyzed. The last section is
Canada, the most demanding application for the diesel engines considered to the remarks and the conclusion.
is when it’s coupled to a synchronous machine operating in
parallel with renewable energy sources used for power supply II. OPERATING DIESEL GENERATOR AND MODEL
and providing light and energy services to these places. A. General Operating of the DG
Customer expectations for higher quality power generation
with lower operating costs have continued to increase in recent The basic components of a diesel generation system are:
years which attracted more researchers to investigate in a new engine, governor, excitation and synchronous generator.
technology for modeling and control of diesel generators
(DG). An efficient model of the diesel generator set can
significantly improve the analysis stage prior to the design of
the hybrid wind-diesel plan as well as its control,
identification, performance prediction and diagnosis processes
[2].
Under that context, a literature review has been done and
several different model of diesel generator set has been found
[3, 4] [2, 5-7] but their development was either very specific
and cannot be used for other applications or not developed for
dynamic studies. In addition, most of these researches have Figure 1. Block diagram of the diesel generator module: principal functional
been conducted for a reduced number of diesel generator blocks and their interconnections[10]

Authorized licensed use limited to: Universite du Quebec a Trois-Rivieres. Downloaded on May 23,2025 at 15:41:54 UTC from IEEE Xplore. Restrictions apply.
978-1-5090-1919-9/16/$31.00 ©2016 IEEE
 2016 IEEE Electrical Power and Energy Conference (EPEC)

A DG set (a unit of diesel engine and governor) is a device 2) Model 2

which converts fuel (diesel oil) energy into mechanical energy This model developed in SPS(simpower systems)/Matlab
in diesel engine and subsequently converts mechanical energy environement inspired from a dynamic model of a gasoline
into electrical energy in a governor. A governor can be defined engine in SPS library [13] deals with purely mechanical
as a mechanical or electromechanical device for automatically phenomena such as modeling of combustion process by
controlling the speed of an engine by relating the intake of controlling engine speed with a fast throttle actuator. This
fuel. The controller for the engine is a simple speed governor model present a detailed mechanical part of the engine, so we
that keeps the turbine operating at its designed speed. The tried to adapt it to a DG with adding the electrical section (GS)
output of the speed governor is throttle signal that controls the and with minor modifications to the inputs in the admission
fuel going into the engine[9].
bloc because the inputs of principal model is based on a
B. Model DG and criterion selection gasoline concept.
After a literature review, four models has selected and analysis
in order to show the most efficient model which presents a
dynamic study of the DG and the more flexible to be used with
several technologies. The models are implemented in
Matlab/Simulink and the simulation results will not be
presented here. Only models structure are presented and
discuss by considering the interaction between mechanical and
electrical part of DG, the ability to integrate supercharging,
and others many criteria which are presented in the end of this
section.
1) Model 1
A developed model on Simulink described by physical
equations is shown in figure 2.It is defined by mathematical
model and physics characteristic. The electrical part is
ontained from the laws of physics, such as the Park Figure 3. Block diagram Generator diesel model
transformation in order to model the synchronus machine and
the grid connection. Mechanical part of diesel engine is Physical and empirical relationships were used to describe the
carrried out by the combustion process [11] that takes into dynamic behavior of the system. The model was able to
account the air mass flow and the speed regulation acts in the respond only for a little power rate and for a short simulation
rack position to control the fuel injection in the cynlinders. time (in order of milliseconds). The model is cannot be easily
The combustion efficiency and the fuel consumption are used for another application and is too specific, evidently it
improved by optimizing air/fuel ratio thanks to the CAES cannot get along with variation load and the interaction
assistance [12]. between mechanical and electrical parameter of DG does not
exist. Then the model is not flexible, a few parameter that can
be controlled and they confirm the limited validity of this
model.
3) Model 3
It is a dynamic model, ready to use from the Simulink library
that addresses the electrical aspects (Figure 4). The internal
combustion chain is based on actuators with simple traditional
models to represent the mechanical part of DG [13] .

Fig.2 Simulink model of a DG overeating with CAES

The construction of this model is simplistic and one of its

benefits consists to introduce overeating. Except that it takes a
long time for simulation; this is due to the large amount of
equations to treat. This amount of calculation is a real barrier
for dynamic simulation and creates convergence problems
because of the algebraic loops that present a real challenge. As
a consequence it slows down the simulation and can even
sometimes stop it completely. Although the phenomena
discuss in this model are important but its simulation is very
difficult, due to the highly nonlinear model. Figure 4. Simulation DG system's structure based of ACA1 exciter

Authorized licensed use limited to: Universite du Quebec a Trois-Rivieres. Downloaded on May 23,2025 at 15:41:54 UTC from IEEE Xplore. Restrictions apply.
 2016 IEEE Electrical Power and Energy Conference (EPEC)

Although this model, it can see the phenomena of electric - Type of intended application and facilitate use for several
stability and performance of a DG (transient voltage and application-specific (standalone application, mobile)
current, power system stability, load flow study). - The possibility of integrating the overeating in the model
It models the engine as a single entry in the form of a transfer - Use with different village load profiles
function. Also the model is effective just with the high power - Model used (behavioral or functional block, Simulink
load (MW) and does not adapt to the low-power load. Model or SPS)
4) Model 4 - Model performance (accuracy, stability, error ...)
This is a new model developed based on the strengths of the - Clear and easy-to-understand modeling
models mentioned above in order to create a complete 6) Model selected
dynamic model that take into account the mechanical and Model 4 that was chosen, selected for its flexibility and
electrical aspects simultaneously. The diesel engine is based integration of various electrical and mechanical approaches. It
on a mechanical model of internal combustion engine and the was concluded that the combined detailed of mechanical
electrical generator is modeled as detailed in model 3. model to the electrical one meet the best objectives study. In
fact, this level of modeling is used to evaluate the influence of
mechanical disturbances in electricity production.
III. DEVELOPED MODEL
The developed model takes into account of the mechanical and
electrical process. The mechanical aspect is defined by
combustion process and engine model. Regarding the
electrical aspect are shown by the synchronous machine which
used from SPS/Matlab environment and the voltage
regulation. Hereafter, the modeling methodology will be
presented along with the main assumptions used to model
principal component of the DG.
Figure 5. Simulation diagram of the Generator diesel with AVR A. Air mass flow model
This model is useful for different load condition and flexible to The air mass flow is modeled in a way to integrate overeating
act on several parameters. It is easy to make modifications on in case of the other use of the model. The amount of air
the system and examine immediately their effects on the entering in the cylinder is written as follows:
system.
It shows the intake events and challenges that must be built. = . ( ) (1)
.
This system analyze both static and dynamic performance and
shown the interaction between electrical and mechanical Where the volumetric efficiency motor [11], WDE is the
process (air mass flow input, engine mechanical power, diesel speed, is displacement volume, engine inlet
generator speed, fuel injected). It enables the study of temperature and engine inlet pressure.
mechanical interactions, regardless of a particular B. Fuel injection system and speed control
configuration of the electrical conversion system. Indeed, the
use of electric model allows focusing on the dynamics of The flow rate of injected fuel is determined from the control of
mechanical components. In addition, a simulation based on a the error between rotational speed of generator and speed
detailed mechanical model represents a preliminary analysis reference (Figure 6) . In general terms, this control works as
for a possible global model. It can target the influence of follows: a disturbance (e.g. a change in the electrical load)
mechanical components on the frequency stability of the produces a change in the power balance of the Gen Set,
system in a model where the electric part is detailed. rotating speeds within the Gen Set change accordingly, a
change in the speed of the diesel engine is sensed by the
5) Criterion for selection governor, the control error is processed and then the
To choose the most representative model of a dynamic system corresponding control signal is sent to the fuel injection
that takes into account the mechanical and the electric process, system, which produces the corresponding change in the fuel
several criteria are required: flow to the engine[3].
- Modeling of the mechanical and electrical part of the
generator (combustion process, diesel engine, electrical
generator ...).
- Observation of the interaction between the mechanical and
the electrical components.
- The level of flexibility in the model which meaning the Figure 6. Black diagram of fuel injection system
facility to create modification (regulator’s parameter,
controller of the fuel injected …) This figure present the model of fuel injected as implemented
in Matlab, the output of the PID is limited to the range [0.1] in

Authorized licensed use limited to: Universite du Quebec a Trois-Rivieres. Downloaded on May 23,2025 at 15:41:54 UTC from IEEE Xplore. Restrictions apply.
 2016 IEEE Electrical Power and Energy Conference (EPEC)

order to represent the operating stroke of the actuator and the Here mechanical and electrical dynamic process of the DG is
delay time representing the internal combustion delay. shown by tracing the rotational generator speed and the
C. Engine model quantity of fuel injected in the cylinders and the terminal
voltage of the synchronous generator. The proposed model is
The model of the engine mechanical power applied to the
simulated for a real case study in Camp Esker, Canada in order
synchronous generator is determined in [1]. The model takes
for a load of 30KW rated power.
into account the physical phenomena of combustion process
The dynamic performance of diesel generator is simulated
by considering the compression ratio of an internal-
when the diesel generator load at time t=40s used for two
combustion engine.
scenarios as presented below.
This model involves the following two parameters:
In scenario 1, the results of the developed model of the DG are
• the quantity of fuel injected into the cylinders
numerically compared with the operation data of the TCE’s
• the speed governor (WDE) DG, main component of the TCE’s test site, based at Rivière
Therefore we chose to model the engine mechanical power [1] au Renard, Canada.
as a function expressed by: The 2nd scenario is illustrated in case by adding a model of
the continously variable transmission (CVT) to the developed
= ( , _ ) system, in order to see its influence on the generator speed and
Where _ is the diesel engine performance expressed by : on the mass flow rate of fuel injected compared with to the
first scenario.
_ = + + Si >
A. Scenario1
_ =( + + ). Si < The model is simulated for 60s with a load of 30 KW power
rate, at time t= 40s the load decreased by 10 KW in order to
a, b, and c are constants, and is the compression ratio (air / verify the robustness of control system and the dynamic
fuel) which is 14.7 for diesel engine [1]. performance of the model. The results are illustrated in figures
D. Electrical Generator 8 to 10: they give (with respect to time) the variation of:
• The rotational generator speed (fig 8);
The electrical generator is represented by present model in SPS
machines library in Matlab [14]. The machine modeling the • The quantity of fuel injected in the cylinders (fig 9)
dynamics of three phase salient pole. • The terminal voltage (fig 10).
Notice that at t=0 the diesel generator starts up and the load is
E. Automatic voltage regulator
connected.
The excitation system of the synchronous generator is based Figure 8 shows that the generator speed is overlaid on the
on automatic voltage regulator (AVR). Its main objective is to reference speed, except during the transitional period. It is
control the rotor angle and the generator terminal voltage, to observed the first action of the diesel speed control strategy
improve the stability properties and to get a good dynamic appears where generator speed increased from its initial value
response. to 2300 rpm during the start-up of the DG, then after 6s the
The proposed model for the AVR is represented on the Figure response stabilize and it becomes coincident with the reference
7 [15]. The model is controlled by a proportional- integral (PI) speed. Following the load decreased at 40s, the generator
controller. speed increased from 1800rpm to 1920 rpm, a maximum
deviation of approximately 5% is obtained and it takes about
2s before the speed returns to the reference value

Figure 7. Simulation diagram of the automatic voltage regulator block

IV. VALIDATION AND DISCUSSION

For validation purpose, a diesel generator of 50 KW rated
power has been implemented in the Matlab/Simulink software
according to the proposed model. It should be noted that it is a Figure 8. Engine speed
digital validation of the model in this study.

Authorized licensed use limited to: Universite du Quebec a Trois-Rivieres. Downloaded on May 23,2025 at 15:41:54 UTC from IEEE Xplore. Restrictions apply.
 2016 IEEE Electrical Power and Energy Conference (EPEC)

Figure 10 shows when the load is decreased at 40s, a very little

variation on the voltage is observed for a very short during
(0.1s) and it stabilizes quickly. This small variation is mainly
due to the AVR parameter. Therefore this variation is
acceptable and the error is ignored, as long as the voltage
remains at its normal operating value.
Whereas for the load decrease, the corresponding figures are
respectively 5% and 1s over compared to the DG data of the
TCE. Note that is very much dependant on the parameters of
the regulators in the developed model.
These characteristics are in accordance with the expected
behavior of a DG and present a model with a relatively small
error. They thus confirm the relevance of the developed
Figure 9. Mass flow fuel injected. model.
B. Scenario2
It is cleared that the variation of the speed is symmetrical to
that observed for the variation of the mass flow fuel injected In scenario 2 the model is simulated for the same parameters
(Figure 9). These characteristics are in accordance with the with the integration of CVT technology. The addition of
expected behavior of a diesel engine, as shown by the continuously variable transmission (CVT) is in order to
proposed model (Figure 6) of the fuel injected is physically control the generator speed to be at its best value. This
related to the generator speed. operating system is evaluated according to a correspondence
A more complete explanation of the phenomenon is given table between the diesel power rate and generator speed. The
below. When a load is connected, and the DG starts up, the correspondence table implemented in the model provided from
generator speed increased and decreased and the speed a document of CVTCorp Company of the DG of the TCE.
regulation acts on the rack to inject more fuel until the Simulation results obtained with CVT are illustrated in figures
combustion is total, however it takes account of air mass flow below compared to the simulation results of the first scenario,
if it is sufficient, then the combustion is complete. As a in objective to see the impact of the CVT in the developed
consequence, the generator speeds stabilize and the fuel model.
injected is superimposed to the fuel curve of TCE model. This section compares the simulation results of the developed
After a load is decreased, the generator speed first goes up. model obtained in first scenario with those obtained with
The speed regulation acts to reduce the fuel injection in the adding of the CVT. It can be seen from the above figures that
cylinders. The air mass flow is sufficient and the combustion the amplitudes of the oscillations in scenario 1 and scenario 2
is total. As a result, the generator speed stabilizes faster and are quite close, and show approximately the same performance
the speed deviation is smaller at 40s. but the results from scenario 2 are suite accuracy and faster.
Figure 10 shows the three-phase voltage Vabc. The voltage of
the PMSG is formed is a sinusoidal waveform, the amplitude
is V=600v.

Figure 11. Engine speed with CVT

Figure 11 and 12 show that the response time of the simulation

model in scenario 2 (red curve) is quite faster than the
Figure 10.Terminal Voltage response in scenario1.

Authorized licensed use limited to: Universite du Quebec a Trois-Rivieres. Downloaded on May 23,2025 at 15:41:54 UTC from IEEE Xplore. Restrictions apply.
 2016 IEEE Electrical Power and Energy Conference (EPEC)

Furthermore this rapidity is accompanied with more precision of fuel injected are almost damped and have the same
in scenario 2 compared to the first one. For example the behavior, but the results from the model with a CVT are better
amplitude of fluctuation is lower in scenario 2, in such a way damped than the results from scenario1.
the higher generator speed observed is 2200 rpm and takes 2s The simulation results have shown that the developed models
to stabilize while the higher generator speed in scenario 1 is are efficient and provide performances similar to those
2500 rpm and takes 5 s to stabilize. obtained with DG data of the TCE.
Although the DG model is not sufficiently accurate, it is close
to becoming a useful model. The generator is a complex,
nonlinear, where both output signals are inextricably linked
and cause reactions that affect the performance of the other.
The continuing work of the authors will be made to increase
the accuracy of individual models and the control.
Finally, this first version of DG presents a dynamic model
produced independently flexible and can easily adapt to
experimental test or on a test bench.
REFERENCES
[1] T. Zannis, E. Pariotis, D. Hountalas, D. Rakopoulos, and Y. Levendis,
"Theoretical study of DI diesel engine performance and pollutant
emissions using comparable air-side and fuel-side oxygen addition,"
Energy Conversion and Management, vol. 48, pp. 2962-2970, 2007.
[2] T. Theubou, R. Wamkeue, and I. Kamwa, "Dynamic model of diesel
generator set for hybrid wind-diesel small grids applications," in
Electrical & Computer Engineering (CCECE), 2012 25th IEEE
Canadian Conference on, 2012, pp. 1-4.
[3] M. Torres and L. A. Lopes, "Inverter-based diesel generator emulator for
the study of frequency variations in a laboratory-scale autonomous
power system," Energy and Power Engineering, vol. 5, p. 274, 2013.
[4] B. Asato, K. Uchida, T. Goya, T. Senjyu, A. Yona. "Optimal operation
for diesel generators in small isolated island power system cosidering
controllable load," in Electrical Machines and Systems (ICEMS), 2010
Figure12. Mass flow rate of fuel injected with CVT International Conference on, 2010, pp. 1869-1872.
[5] D. Wang, C. Nayar, and C. Wang, "Modeling of stand-alone variable
From the results of Figure 12, we can see a better control over speed diesel generator using doubly-fed induction generator," in Power
the amount of the fuel injected so that the oscillations Electronics for Distributed Generation Systems (PEDG), 2010 2nd IEEE
presented at the startup of the DG with smaller amplitude International Symposium on, 2010, pp. 1-6.
compared to Scenario 1 and the system stabilizes faster. As a [6] M. Torres and L. A. Lopes, "Inverter-based virtual diesel generator for
laboratory-scale applications," in IECON 2010-36th Annual Conference
consequence we save a quite quantity of fuel in this scenario on IEEE Industrial Electronics Society, 2010, pp. 532-537.
and the system is more stable. Therefore the system was [7] G. Hemmatipoor, M. Monadi, A. Darabi, and S. Shadmani, "Modeling
observed to stabilize the variation after a single passing; this is Small Diesel Generator Connected to Non Linear Loads Base on Space
due to the PID controller parameters implemented in the and Time Harmonics," in Future Computer Sciences and Application
(ICFCSA), 2011 International Conference on, 2011, pp. 114-118.
model. The performance of the CVT technology is designed to [8] M. R. Morgan TR, Brinkworth BJ.ARES:arefined simulation program
operate the diesel generation in an economical way to reduce for the sizing an optimization of autonomous hybrid energy
the fuel consumption and total cost of the generation, which systems.Solar Energy Research Laboratory, Department Of Mechanical
can be observed from the simulation. Engineering, 1997;59(4–6):205–15.
[9] A. Ahmad and A. Siddiqui, "Modeling of A Wind Diesel Integrated
V. CONCLUSION System With No Storage."
[10] J. T. Bialasiewiez, Modular simulation of a hybrid power system with
The synthesis of the models has allows us to choose the most diesel and wind turbine generation: Citeseer, 1998.
representative model of dynamic system which presented [11] H. Ibrahim, Étude et conception d'un générateur hybride d'électricité de
mechanical and electrical aspects. Then a new dynamic model type éolien-diesel avec élément de stockage d'air comprimé: Université
for a DG has been proposed. This model is intended for use in du Québec à Chicoutimi, 2010.
[12] H. Ibrahim, A. Bourji, M. Ghandour, and A. Merabet, "Optimization of
impacts studies of dispersed generation on distribution compressed air storage's volume for a stand-alone wind-diesel hybrid
networks. system," in Electrical Power & Energy Conference (EPEC), 2013 IEEE,
The developed model has a modular structure corresponding 2013, pp. 1-7.
to the different mechanical and electrical components of the [13] Thermal Energy System Specialists. Drake landing solar community:
Okotoks 2006, "http://tess-inc.com/projects/project1."
DG (diesel engine, regulator voltage, speed controller, air [14] Matlab/Simulink.
mass flow model). http://www.mathworks.com/help/physmod/sps/powersys/ref/synchronou
Two scenarios were performed to validate numerically the smachine.html [Online].
proposed Simulink models and to verify their accuracy. [15] G. Sybille. "Simulation of FACTS controllers using the MATLAB
power system blockset and Hypersim real-time simulator," in Power
Simulation results obtained with the two scenarios have shown Engineering Society Winter Meeting, 2002. IEEE, 2002, pp. 488-491.
that the oscillations in generator speed and the mass flow rate

Authorized licensed use limited to: Universite du Quebec a Trois-Rivieres. Downloaded on May 23,2025 at 15:41:54 UTC from IEEE Xplore. Restrictions apply.