Hybrid Electric Vehicle Design

Although the sizes and types of components used in a series hybrid drive train may vary, their

functional roles remain pretty much the same. An onboard generator maintains the state of charge of

an energy accumulator and the energy from this accumulator is transformed by an energy converter

into torque to drive the wheels. The main design parameters in a series hybrid are the selection

and sizing of the generation and storage devices. Generators can be an ICE/generator pair, a gas

turbine generator, fuel cells, and many others. Candidates for the accumulator include batteries,

ultra capacitors, °ywheels, and hydraulic accumulators. Typically, an electric motor is used to

provide torque at the wheels. Conventional design methodology for a series hybrid consists mainly

of sizing the propulsion motor to provide the desired performance and then choosing a generator

and storage device to provide the required range on a speci¯c drive cycle. Although the devices

chosen to ¯ll those roles might be distinct, their functions keep unchanged. Thus, a series power

plant can be designed based almost entirely on the functional roles and then components can be

selected to best meet power requirements. Figure 2 shows the design variables of di®erent system

components involved in the system-level series HEV design. We can see that the HEV has a number

of design variables as well as multiple design objectives, and thus HEV design is a complicated and

cumbersome task. Di®erent design objectives of HEV performance should be optimized while the

con°icting design constraints should also be ful¯lled simultaneously.

The design problem can be formulated based on the HEV operation principles. The major

components are identi¯ed and their characteristics are de¯ned. The important equations relating

the components to the performance characteristics, as well as the inter-relating components are

4

derived. The derived equations, which involve these design variables, are used to design the power

train. The formulated problem is the basis for the future system level design of the series hybrid

drive train. In the conventional HEV design, the design °ow is unidirectional and predetermined.

It goes from the performance speci¯cations to the other components as shown in Figure 3. The

conventional design process looks at the design problem from a physical perspective; each variable

has its own distinct identity; it has its characteristics, importance and in a conventional design, is

always calculated in a speci¯c way. For example, acceleration is a performance speci¯cation, while

gear ratio is a physical variable. For design, acceleration has to be de¯ned before hand; otherwise

conventional design is not possible. The gear ratio will be calculated from the maximum velocity of

the vehicle and that of the motor. The designer always starts with the performance speci¯cations

and work towards the parameters. A major drawback of this approach is its low e±ciency and

the fact that it is not able to consider multiple design objectives simultaneously. A reasonable

tradeo® is hard to achieve in this way. The hybrid electric vehicle design optimization used to

be a multivariable constrained nonlinear problem. Unfortunately, conventional local optimization

methods may often lead to an undesired local solution dependent on the chosen starting points.

HEV design should be treated as a globally constrained problem. This allows the designers to use

more design variables and fewer simpli¯cations. The optimized HEV design can only be reached

with the correct balance of multiple criteria, because the improvement of one criterion will be at

the expense of at least one other objective. This research intends to develop a multi-objective

optimization approach to the HEV design, which controls the searching process toward the global

solution via an evolutionary algorithm.