Rebuttal - IMECE2016 66312 1
Rebuttal - IMECE2016 66312 1
Rebuttal - IMECE2016 66312 1
We thank the referees and the editor for their constructive comments. We have
addressed all the reviewers comments below. The corresponding changes have also been
incorporated in the revised draft appropriately and highlighted in red color font.
Reviewer: 1
The paper presents an initial sizing and optimization of the powertrain components for a
series hybrid, heavy duty truck. The powertrain is optimized for an Indian drive cycle with
vehicle performance constraints. Improvement from a baseline design is presented. The
paper could be improved by considering the following:
We thank reviewer (1) for his/her valuable comments and suggestions. The responses to the
same are given below.
Comment 1: Authors should clarify the type of hybrid technology, i.e., plug-in hybrid or
battery hybrid, that is considered in this paper.
Response: The hybrid vehicle technology considered in the paper is a battery hybrid and the
same has been mentioned now in the revised draft.
Comment 2: Justification for series hybrid architecture should be relative to the vehicle
class that is considered in the paper. An architecture suitable for a passenger vehicle may not
be suitable for a heavy duty truck. Considering the application to be a heavy duty truck in
this paper, it is hard to justify the use of series architecture since high torque requirements in
heavy vehicles make the series design undesired in general.
Response: As correctly pointed out by the reviewer, a hybrid architecture suitable for a
passenger vehicle may not be suitable for a heavy duty truck. The drive cycle considered in
this paper is a city drive cycle with frequent stop and go conditions. This paper aims at
optimizing the class of heavy road vehicles available and used particularly in urban driving
conditions such as city buses, dump trucks, towing trucks, pickup trucks, etc. Because of the
frequent stop-and-go condition in city driving, the efficiency of a conventional heavy road
vehicle is poor since the engine is not able to run in its optimized region. But, in a series
hybrid configuration, the engine is not directly linked to the transmission or the drive wheels.
The electric motor system provides tractive power to turn the wheels of the vehicle and
recharge the batteries.
Since the engine does not follow the load requirements of the vehicle,
it can be operated at its most optimum points in its speed-torque map, regardless of vehicle
speed and load.
An electric motor provides maximum torque at low speed regions (up to its
base speed), beyond which it enters the constant power region. Hence, the use of an electric
motor for initial high torque requirement in a city drive cycle is justified by a series hybrid
configuration, where the vehicle can run almost in a pure electric mode.
For simulation purpose, a conventional heavy duty truck9 was considered for the SHEV
design.
Response: Thank you. The literature review has been updated in the second page of the
revised draft as follows:
The optimization of the vehicle powertrain is the next important step to fulfil the vehicle
performance and achieve best fuel economy. Various optimization algorithms for the design
of powertrain systems of electric and hybrid electric vehicles can be found in the literature
[5]. Gradient-based algorithms such as sequential quadratic programming (SQP) [5] use the
derivative information to find the local minima. Derivative-free algorithms such as divided
rectangles (DIRECT) and genetic algorithm (GA) [6], work well for objective functions that
do not rely on derivatives. Also, the DIRECT algorithm is a deterministic global optimization
algorithm ensuring that the objective function converges to a global optimum in finite time.
Gao et al. [5] optimized a PHEV using PSAT software for maximum fuel economy on a
composite drive cycle using four global optimization algorithms (DIRECT, GA, Simulated
Annealing and Particle Swarm Optimization) and found out that the derivativefree
algorithms (DIRECT, Simulated Annealing and GA) were efficient for the complex HEV
design problem.
The DIRECT algorithm is used for optimization of the vehicle [6].
Comment 4: Following up with the previous comment, a recent review article discusses
various powertrain design studies for different vehicle applications.
DOI:10.1109/TVT.2016.2547897
Response: We thank you for pointing out this paper. The paper has been referred to in the
revised draft.
Comment 5: In the third paragraph of the introduction, authors list three design constraints.
A vehicles ability to follow a drive cycle and battery SOC limits are also commonly used as
design constraints during a powertrain design process.
Response: The three main operations that provide the basic design constraints for a hybrid
powertrain are discussed in the third paragraph of the introduction. For the powertrain design
process, the battery SOC limits and the vehicles ability to follow a drive cycle were also
taken into consideration along with the acceleration performance and gradeability, as given in
table 5 of the draft. The same has been reproduced here:
delta_trace Difference between drive cycle requested speed and vehicle 3.20 kmph
achieved speed at every second during the drive cycle
delta_soc Difference between final and initial battery state of charge 0.50%
Response: The variable R(t) is the total resistance force of the vehicle that includes the tire
rolling resistive force (Rf(t)), aerodynamic force (Ra(t)) and the grade resistance (Rg(t)):
R(t ) = R f (t ) + Ra (t ) + Rg (t ).
The same has been mentioned in the revised draft.
Response: Thank you. This suggestion has been incorporated in the revised manuscript as
follows:
Figure 3 shows an approximated induction motor torque speed characteristic curve where
the natural mode region was not considered.
Comment 8: Authors cite Ross et al. in page 4 but it is missing in the references.
Response: Thank you. This reference has been updated in the revised draft.
Response: Thank you. This sentence has been corrected in the revised draft as follows:
The power curve was obtained from Eqn. (13) as a result of numerical simulation by
changing the speed ratios from 2 to 10 for five different motors with a maximum speed
ranging from 9000 rpm to 13000 rpm.
Response: The parameter S represents the driving range of the vehicle in pure electric mode.
The same has been defined in the revised draft.
Comment 11: It is not clear why iterative optimization is used. Since iterative optimization
methods might have convergence problems, the need for an iteration loop should be clarified.
If motor and battery weights are defined as functions of the design variables, then a single
optimization formulation can be made.
Comment 12: Derivative-free algorithms do not rely on derivatives and can therefore
work well for any kind of objective function, thus, often being the best global algorithms.
Derivative-free methods have their own drawbacks. They might fail for highly constrained
design problems for instance. Authors should be careful when justifying the optimization
methods they use.
Response: Gradient based algorithms use the derivative information to find the local minima.
For the optimization problem described in the paper, smooth gradients for the gradient based
optimization tools (FMINCON) could not be provided in the entire design space, leading to
premature termination of the optimization process. Also, a gradient based local optimizer
does not generally search the entire design space to find the global minimum. Hence, a
derivative free algorithm was used because of the characteristic of the objective function,
since the optimization algorithm could not rely on its derivatives in the entire design space.
The DIRECT algorithm used in this paper is a deterministic global optimization algorithm
ensuring that the objective function converged to a global optimum in finite time. It was
found that the optimization process in this study fortunately did not suffer from any
convergence issue.
Comment 13: What are the lower and upper limits on the battery SOC variation?
Response: We apologize for not highlighting the limits on the battery SOC variation. Based
on an extensive literature review, the battery SOC variation was found out to be between 0.4
and 0.8. The same SOC limits has been considered in the paper and updated in the revised
draft on page 6.
Comment 14: Only one drive cycle is used for optimization. Generally, designs optimized for
one drive cycle might not work well for different driving conditions. Maybe as a future work,
authors should consider integrating other driving conditions such as a highway cycle.
Response: An on road measured Indian drive cycle has been considered in the paper for the
design and optimization of the SHEV. As a future work, we would definitely consider a drive
cycle consisting of both a city and a highway cycle.
Comment 15: Different architecture choices (parallel or series-parallel) might be a better fit
for the vehicle application selected in the paper. Considerations such as cost or space might
make the bulky motor design in a series architecture undesired. It should be acknowledged
and stated at least as a future study.
Response: Thank you for the suggestion. We could try to formulate a minimization
optimization problem as a future study by implementing the objective function as a cost
function or a weighted function to take care of the cost of the powertrain components.
The same has been updated in the conclusion of the revised draft:
The optimization of the SHEV done in the paper is a single objective function of achieving
the maximum fuel economy in a city drive cycle. As a future study, a multi-objective
function can be formulated to take into account the cost factor and space constraint of the
powertrain components for a city-highway drive cycle. Optimization of a parallel and a
series-parallel configuration can also be made to select the best possible configuration for the
vehicle under study for different drive cycles.
Reviewer: 2
We thank reviewer (2) for his/her valuable comments and suggestions. The responses to the
same are given below.
Comment 1: A better layout of SHEV needs to be provided, the current figure has missing
component names.
Response: The layout of the SHEV has been updated in the revised draft as follows:
Comment 2: More little literature review on prior work done on the topic needs to be
included.
Response: Thank you for the suggestion. The literature review has been updated in the
revised draft by including the work done in the optimization of a hybrid powertrain design as
follows:
The optimization of the vehicle powertrain is the next important step to fulfil the vehicle
performance and achieve best fuel economy. Various optimization algorithms for the design
of powertrain systems of electric and hybrid electric vehicles can be found in the literature
[5]. Gradient-based algorithms such as sequential quadratic programming (SQP) [5] use the
derivative information to find the local minima. Derivative-free algorithms such as divided
rectangles (DIRECT) and genetic algorithm (GA) [6], work well for objective functions that
do not rely on derivatives. Also, the DIRECT algorithm is a deterministic global optimization
algorithm ensuring that the objective function converges to a global optimum in finite time.
Gao et al. [5] optimized a PHEV using PSAT software for maximum fuel economy on a
composite drive cycle using four global optimization algorithms (DIRECT, GA, Simulated
Annealing and Particle Swarm Optimization) and found out that the derivativefree
algorithms (DIRECT, Simulated Annealing and GA) were efficient for the complex HEV
design problem.
Response: The drive cycle considered in the paper is a real time drive cycle in the city of
Coimbatore, India, that was provided by Mr C S Nandakumar, Robert Bosch Engineering and
Business solutions limited (RBEI). The authors have acknowledged Mr. C S Nandakumar for
the same. Citation for the drive cycle has been done in the revised draft as reference number
[16].
Comment 4: There seems to be some discrepancy in the vehicle parameter values used for
simulation versus the values in reference 7. Correction or explanation required.
Response: The vehicle in reference 7 was used as the base vehicle to make a comparison
between a conventional, a series hybrid and an optimized series hybrid electric vehicle.
In the software used for simulation (ADVISOR), individual powertrain component masses
need to be quantified and updated. The software then calculates the gross vehicle weight. The
glider mass (mass of the vehicle without the powertrain), the cargo mass and the vehicle
parameters (Cd, Af, fr, r, L) were kept the same as given in the reference 7. Depending on the
vehicle performance constraints, individual powertrain component specifications were
calculated. Since the individual components specification data were not available in the
reference, approximated values were updated in the software. Thus, the values of a few
vehicle parameters such as the GVW and the engine rated torque used in simulation were
different from that of the reference vehicle7. The difference in the parameter values are given
in the following table:
Response: Thank you. The following symbols have been included in the nomenclature of the
revised draft:
Response: We apologize for not highlighting the limits on the battery SOC variation. Based
on an extensive literature review, the battery SOC variation was found out to be between 0.4
and 0.8. The same SOC limits has been considered in the paper and updated in the revised
draft on page 6.
Draft Recommendations/Comments:
Comment:
Based on the detailed comments of the reviewers, I am requesting you submit a
revised version of your draft paper in order to consider it for publication in the conference
proceedings. Upon review of the revision, the paper will be accepted.
Response: Thank you for your comment. The comments of the reviewers have been
addressed and corresponding changes have been made in the revised draft.