Search Results

In recent years, the urgent need to fully exploit the fuel economy potential of the Electrified Vehicles (xEVs) through the optimal design of their Energy Management System (EMS) have led to an increasing interest in Machine Learning (ML) techniques. Among them, Reinforcement Learning (RL) seems to be one of the most promising approaches thanks to its peculiar structure, in which an agent is able to learn the optimal control strategy through the feedback received by a direct interaction with the environment. Therefore, in this study, a new Soft Actor-Critic agent (SAC), which exploits a stochastic policy, was implemented on a digital twin of a state-of-the-art diesel Plug-in Hybrid Electric Vehicle (PHEV) available on the European market. The SAC agent was trained to enhance the fuel economy of the PHEV while guaranteeing its battery charge sustainability.

Traditional CACC systems utilize inter-vehicle wireless communication to maintain minimal yet safe inter-vehicle distances, thereby improving traffic efficiency. However, introducing communication delays generates system uncertainties that jeopardize string stability, a crucial requirement for robust CACC performance. To address these issues, we introduce a decentralized Model Predictive Control (MPC) approach that incorporates Kalman Filters and state predictors to counteract the uncertainties posed by noise and communication delays. We validate our approach through MATLAB Simulink simulations, using stochastic and mathematical models to capture vehicular dynamics, Wi-Fi communication errors, and sensor noises. In addition, we explore the application of a Reinforcement Learning (RL)-based algorithm to compare its merits and limitations against our decentralized MPC controller, considering factors like feasibility and reliability.

The REAL NOx regulation requires tracking and reporting of NOx emissions starting in 2022MY for both medium-duty and heavy-duty diesel vehicles with potential to be considered during the next light-duty rulemaking. The regulation includes minimum NOx mass measurement accuracy requirements of either +/−20 percent or +/− 0.1 g/bhp-hr. Existing NOx sensor technology may not be able to meet the regulated accuracy requirements especially when exposed to other sources of variation within the emissions control system. This paper provides an assessment of real-world NOx sensor accuracy and the impact of other sources of variation and noise factors on NOx measurement accuracy. Noise factors investigated include NOx sensor tolerance, exhaust flow rate estimation, NOx sensor ammonia (NH3) cross sensitivity, mass air flow (MAF) sensor accuracy, NOx sensor placement, and laboratory emissions measurement capability.

A model-based torque control strategy which is simple and easily adaptable to various types of engines is developed in this paper. A torque model is derived from constant-volume combustion model, and applications of the model to engine torque control problem are also discussed. As examples, the torque model is calibrated with experimental data collected from two different engines, and simulation and experimental results from the torque control strategy are presented as well.

The paper presents a robust adaptive control technique for precise regulation of a port fuel injection + direct injection (PFI+DI) system, a dual fuel injection configuration adopted in modern gasoline engines to boost performance, fuel efficiency, and emission reduction. Addressing parametric uncertainties on the actuators, inherent in complex fuel injection systems, the proposed approach utilizes an indirect model reference adaptive control scheme. To accommodate the increased control complexity in PFI+DI and the presence of additional uncertainties, a nonlinear plant model is employed, incorporating dynamics of the exhaust burned gas fraction. The primary objective is to optimize engine performance while minimizing fuel consumption and emissions in the presence of uncertainties. Stability and tracking performance of the adaptive controller are evaluated to ensure safe and reliable system operation under various conditions.

When an SI engine is equipped with individual cylinder pressure transducers, combustion timing of each cylinder can be precisely controlled by adjusting spark timing in real-time. In this paper, a novel method based on principal component analysis (PCA) is introduced to control the combustion timing with a significantly less computational burden than a conventional method.

In the face of growing concerns about environmental sustainability and urban congestion, the integration of eco-driving strategies has emerged as a pivotal solution in the field of the urban transportation sector. This study explores the potential benefits of a CAV functioning as a virtual eco-driving controller in an urban traffic scenario with a group of following human-driven vehicles. A computationally inexpensive and realistic powertrain model and energy management system of the Chrysler Pacifica PHEV are developed with the field experiment data and integrated into a forward-looking vehicle simulator to implement and validate an eco-driving speed planning and energy management strategy assuming longitudinal automation. The eco-driving algorithm determines the optimal vehicle speed profile and energy management strategy.

In this paper, we present a novel algorithm designed to accurately trigger the engine coolant flow at the optimal moment, thereby safeguarding gas-engines from catastrophic failures such as engine boil. To achieve this objective, we derive models for crucial temperatures within a gas-engine, including the engine combustion wall temperature, engine coolant-out temperature, engine block temperature, and engine oil temperature. To overcome the challenge of measuring hard-to-measure signals such as engine combustion gas temperature, we propose the use of new intermediate parameters. Our approach utilizes a lumped parameter concept with a mean-value approach, enabling precise temperature prediction and rapid simulation. The proposed engine thermal model is capable of estimating temperatures under various conditions, including steady-state or transient engine performance, without the need for extra sensors.

To meet the stringent NOx and particulate emissions requirements of Euro 6 and China 6 standard, Selective Catalyst Reduction (SCR) catalyst integrated with wall flow particulate filter (SCR-DPF) has been found to be an effective solution for the exhaust aftertreatment systems of diesel engines. NOx is reduced by ammonia generated from urea injection while the filter effectively traps and burns the particulate matter periodically in a process called regeneration. The engine control unit (ECU) effectively manages urea injection quantity, timing and soot burning frequency for the stable functioning of the SCR-DPF without impacting drivability. To control the NOx reduction and particulate regeneration process, the control unit uses lookup tables generated from extensive hardware testing to get the current soot load and NOx slip information of SCR-DPF as a function of main exhaust state variables.

Cast austenitic stainless steels, such as 1.4837Nb, are widely used for turbo housing and exhaust manifolds which are subjected to elevated temperatures. Due to assembly constraints, geometry limitation, and particularly high temperatures, thermomechanical fatigue (TMF) issue is commonly seen in the service of those components. Therefore, it is critical to understand the TMF behavior of the cast steels. In the present study, a series of fatigue tests including isothermal low cycle fatigue tests at elevated temperatures up to 1100°C, in-phase and out-of-phase TMF tests in the temperature ranges 100-800°C and 100-1000°C have been conducted. Both creep and oxidation are active in these conditions, and their contributions to the damage of the steel are discussed.

The engine acoustic character has always represented the product DNA, owing to its strong correlation with in-cylinder pressure gradient, components design and perceived quality. Best practice for engine acoustic characterization requires the employment of a hemi-anechoic chamber, a significant number of sensors and special acoustic insulation for engine ancillaries and transmission. This process is highly demanding in terms of cost and time due to multiple engine working points to be tested and consequent data post-processing. Since Neural Networks potentially predicting capabilities are apparently un-exploited in this research field, the following paper provides a tool able to acoustically estimate engine performance, processing system inputs (e.g. Injected Fuel, Rail Pressure) thanks to the employment of Multi Layer Perceptron (MLP, a feed forward Network working in stationary points).

The past five years have seen significant research into autonomous vehicles that employ a by-wire steering rack actuator and no steering wheel. There is a clear synergy between these advancements and the parallel development of complete Steer-by-Wire systems for human-operated passenger vehicle applications. Steer-by-Wire architectures presented thus far in the literature require multiple layers of electrical and/or mechanical redundancy to achieve the safety goals. Unfortunately, this level of redundancy makes it difficult to simultaneously achieve three key manufacturer imperatives: safety, reliability, and cost. Hindered by these challenges, as of 2020 only one production car platform employs a Steer-by-Wire system. This paper presents a Steer-by-Wire architectural solution featuring fail-operational steering control architected with the objective of achieving all system safety, reliability, and cost goals.