A COMPREHENSIVE METHODOLOGY for the OPTIMIZATION of the OPERATING STRATEGY of HYBRID ELECTRIC VEHICLES

The sustainable exploitation of energy and reduction of pollutant emissions are main concerns in our society. Driven by more stringent international standards, automobile manufacturers are developing new technologies such as the Hybrid Electric Vehicles (HEVs). These innovative systems combine the main benefits of traditional Internal Combustion Engines (ICEs) with those of Battery Electric Vehicles (BEVs), while overcoming their main drawbacks. HEVs can offer significant improvements in the efficiency of the propulsion system, but they also lead to higher complexities in the design and in the control. In order to exploit all the expected advantages, a dedicated optimization of the Hybrid Operating Strategy (HOS) is required. In this framework, simulation plays a key role in identifying the optimal HOS, where the primary design targets are the fuel economy, emission reduction and improvement in the vehicle performance (including acceleration, driving range, operational flexibility and noise). With such a perspective, a simulation study was performed involving the implementation, in Matlab environment, of zero-dimensional models of a Series Hybrid Electric Vehicle (SHEV) and a Parallel Hybrid Electric Vehicle (PHEV). As far as the hybrid operating strategy is concerned, three different approaches were investigated: _ A novel Benchmark Optimizer (BO), that determines the best possible operating strategy for the selected target, mission profile and powertrain design. The single solution is characterized by a vector, in which every scalar independently defines the mechanical power of the electric machine, for the PHEV, or the engine speed, for the SHEV, at each time step of the selected driving cycle _ A real-time optimizer based on the Minimization of the Total system Losses (TLM). It involves a vector-approach, in order to select, at each time step, the power split that guarantees the minimum system losses. It requires a reduced number of calibration parameters and, therefore, is computationally fast and adequate to work in real-world applications. Based on this technique, two different methodologies concerning the engine component are considered: the Total engine losses (TLM TOT) and the Recoverable (with respect to the optimal operating point) engine losses (TLM REC) _ A real-time optimizer based on the Total Load Switch Thresholds. It switches the operating mode depending on the load and speed signals. It involves a scalar-approach and requires a reduced number of calibration parameters. It is by far the method that requires the least computational effort In all the three cases, the numerical optimizer is based on Genetic Algorithm (GA) techniques. GAs are inspired by the mechanism of natural selection, in which better individuals are likely to be the winners in a competing environment. It is a statistical approach able to solve optimization problems whose objective function is non-continuous, non-differentiable, stochastic and highly non-linear. The study analyses the optimization of the well-to-wheel CO2 emissions of a Parallel and a Series Hybrid Electric Vehicles along the New European Driving Cycle (NEDC) and the Artemis Driving Cycles. In the case of the only compression ignition engine, also NOx emissions were considered as optimization criteria along the NEDC.