Search Results

The authors present the supervisory control of a parallel hybrid powertrain, focusing on several issues related to the real-time implementation of optimal control based techniques, such as the Equivalent Consumption Minimization Strategies (ECMS). Real-time implementation is introduced as an intermediate step of a complete chain of tools aimed at investigating the supervisory control problem. These tools comprise an offline optimizer based on Pontryagin Minimum Principle (PMP), a two-layer real-time control structure, and a modular engine-in-the-loop test bench. Control results are presented for a regulatory drive cycle with the aim of illustrating the benefits of optimal control in terms of fuel economy, the role of the optimization constraints dictated by drivability requirements, and the effectiveness of the feedback rule proposed for the adaptation of the equivalence factor (Lagrange multiplier).

The paper presents a new control strategy to compensate the mutual influence of multiple injections in diesel and HCCI engines. The approach is based on a control-oriented model of the process, which represents the dependencies between injection timing, rail pressure, and masses injected. The model is conveniently inverted to yield the injection timing required to obtain a desired mass pattern. The model-based compensator developed is calibrated against measurements taken both on a dedicated injection bench and on a HCCI engine test bench. The compensator is then implemented in the control unit of the latter and validated against measurements of fuel consumption.

Fuel consumption is an essential factor that requires to be minimized in the design of a vehicle powertrain. Simple energy models can be of great help - by clarifying the role of powertrain dimensioning parameters and reducing the computation time of complex routines aiming at optimizing these parameters. In this paper, a Fully Analytical fuel Consumption Estimation (FACE) is developed based on a novel GRaphical-Analysis-Based fuel Energy Consumption Optimization (GRAB-ECO), both of which predict the fuel consumption of light- and heavy-duty series hybrid-electric powertrains that is minimized by an optimal control technique. When a drive cycle and dimensioning parameters (e.g. vehicle road load, as well as rated power, torque, volume of engine, motor/generators, and battery) are considered as inputs, FACE predicts the minimal fuel consumption in closed form, whereas GRAB-ECO minimizes fuel consumption via a graphical analysis of vehicle optimal operating modes.

Due to more and more complex powertrain architectures and the necessity to optimize them on the whole driving conditions, simulation tools are becoming indisputable for car manufacturers and suppliers. Indeed, simulation is at the basis of any algorithm aimed at finding the best compromise between fuel consumption, emissions, drivability, and performance during the conception phase. For hybrid vehicles, the energy management strategy is a key driver to ensure the best fuel consumption and thus has to be optimized carefully as well. In this regard, the coupling of an offline hybrid strategy optimizer (called HOT) based on Pontryagin’s minimum principle (PMP) and an online equivalent-consumption-minimization strategy (ECMS) generator is presented. Additionally, methods to estimate the efficiency maps and other overall characteristics of the main powertrain components (thermal engine, electric motor(s), and battery) from a few design parameters are shown.

In this paper, an integrated simulation-based methodology demonstrating feasibility and performance of several electric-hybrid concepts is developed. Several advanced tools are coupled to define the specifications of each component of the hybrid powertrain, to select the most promising hybrid architecture and finally to assess the proposed powertrain with regard to CO2 and pollutants emissions. Concurrent minimization of NOx and CO2 emissions enables to find the best compromise to fulfil Euro 6 standards while lowering fuel consumption. This stage consists in an iterative co-optimization of the power split strategies between the electric drive and the Diesel engine and of the engine settings (injection pressure, EGR rate, etc.). The methodology combines optimal control laws and optimization methodology based on global statistical models using single-cylinder design of experiments. After several iterations, this method allows to find the optimal NOx/CO2 trade-off curve.