Proposal of new methodology

As vehicles are subjected to different driving conditions in the real-world, powertrain component sizes need to be optimum over different driving patterns simultaneously to perform optimally in the real-world. With this hypothesis, a new methodology is proposed for the optimisation of powertrain components through computer simulation model, so that powertrain components are optimum over a range of different driving patterns simultaneously. The main concept of the new methodology is that no ideal driving pattern is possible and no single driving pattern can represent the real-world and for that reason different driving patterns need to be considered during the optimisation of powertrain component sizes to represent the real-world in a realistic way. The new methodology improves the conceptual weakness of the traditional methodology by considering a range of driving patterns representing different traffic conditions and driving styles simultaneously to make the new methodology more applicable for real-world application, and for this reason the new methodology can be considered conceptually original. In the new methodology, driving patterns are categorised into different traffic conditions and each traffic condition is further classified into different driving styles. For example, traffic conditions could be categorised into urban and highway and each traffic condition could be further classified into different driving styles e.g., conservative, normal, and aggressive. After categorisation, all the driving patterns are considered during the evaluation of an objective (e.g., FE, emissions etc.). During optimisation, the

driving patterns. The optimisation needs to be decided based on the cumulative value of the objective over all the categorised driving patterns.

For example, it is assumed that driving patterns are classified into four different driving patterns and the objective of optimisation is the minimisation of FE, as shown in Figure 5.1. If the FE values over the four driving patterns are FE1, FE2, FE3, and FE4 respectively, then the optimisation decision needs to be done based on the summation of the FE1 to FE4 for the new methodology, as shown in Figure 5.1. This improves upon the traditional methodology which considers FE over only one driving pattern and therefore, provides different optimum designs over different driving patterns, as shown in Figure 5.2.

Figure 5.2: Concept of optimisation decision in the traditional methodology

As the optimisation in the new methodology is done based on the objective value over a range of different driving patterns representing different traffic conditions and driving styles, the powertrain components are actually optimum over a range of driving patterns representing different traffic conditions and driving styles. As the optimum design produced by the new methodology is optimum over a range different driving patterns, it is expected that FE variability of the optimum design will be reduced. The new methodology will be called as proposed methodology and termed as M2 in the remaining discussion.

The proposed methodology is as follows and shown in Figure 5.3.

1) Assumption of initial sizes of powertrain components.

2) The characteristics of those components are decided based on the base components. The base components are the components whose performance characteristics are used to determine the characteristics of new components during the optimisation process.

3) Driving patterns are categorised into different traffic conditions (e.g., urban, highway etc.) and driving styles (e.g., conservative, normal, aggressive etc.).

5) The components are evaluated according to a vehicle supervisory control (VSC) strategy for an objective or objectives (e.g., FE, emissions etc.) over the combination of categorised driving patterns.

6) The components are checked against design constraints to ensure the minimum performance requirements.

7) The optimisation process is checked against a termination criterion which might be a fixed number of iterations or until a stable objective value is achieved.

8) If the optimisation termination criterion is not met, the current component sizes are fed into an optimisation method.

9) The optimisation method generates new sizes of the powertrain component. 10) If a parameter of the VSC strategy also needs to be optimised, the

optimisation method generates a new value of the parameter.

11) Repeat 2 and 5 to 9, if only the powertrain components need to be optimised or repeat 2 and 5 to 10, if both the powertrain components and parameters of the VSC strategy need to be optimised, until the optimisation termination criterion is achieved.

12) If the optimisation termination criterion is met, the optimisation process reaches the optimum sizes of the powertrain components.

1) Initial size of powertrain components

5) Evaluation of objective over the combination of categorised driving

patterns (DP1, DP2,…, DPn)

6) Design constraints

9) New size of powertrain components

7) Is optimisation termination criterion met?

12) Optimum powertrain components No

Yes

8) Optimisation method Base powertrain components

(ICE, generator, motor, battery)

2) Characteristics of powertrain components

3) Categorisation of driving patterns into different traffic conditions and driving styles (DP1, DP2, …, DPn) Vehicle supervisory control strategy 10) New values of parameters of vehicle supervisory control strategy 4) Combination of all the categorised

driving patterns in series

Figure 5.3: Proposed methodology of powertrain component size optimisation (M2)

The conceptual difference between the proposed and traditional methodologies is that the proposed methodology considers a range of driving patterns representing different traffic conditions and driving styles simultaneously, as shown in Figure 5.3, whereas the traditional methodology generally considers a single standard driving pattern for the optimisation of powertrain component sizes, as discussed in chapter 3

Figure 3.1 (Repeated): Traditional methodology of powertrain component size optimisation (M1)

5.2 Summary

 A new methodology has been proposed for the optimisation of powertrain

component sizes to address FE variability due to variation in driving patterns in HEVs.

 The proposed methodology considers a range of driving patterns representing

different traffic conditions and driving styles simultaneously instead of a single standard driving pattern used in the traditional methodology followed in the reviewed literature.