Search Results

This work describes the application of Non-Linear Autoregressive Models with Exogenous Inputs (NLARX) in order to predict the NOx emissions of heavy-duty diesel engines. Two experiments are presented: 1.) a Non-Road-Transient-Cycle (NRTC) 2.) a composition of different engine operation modes and different engine calibrations. Data sets are pre-processed by normalization and re-arranged into training and validation sets. The chosen model is taken from the MATLAB Neural Network Toolbox using the algorithms provided. It is teacher forced trained and then validated. Training results show recognizable performance. However, the validation shows the potential of the chosen method.

The control of Homogeneous Charge Compression Ignition (HCCI) (also known as Controlled Auto Ignition (CAI)) has been a major research topic recently, since this type of combustion has the potential to be highly efficient and to produce low NOx and particulate matter emissions. Ion current has proven itself as a closed loop control feedback for SI engines. Based on previous work by the authors, ion current was acquired through HCCI operation too, with promising results. However, for best utilization of this feedback signal, advanced interpretation techniques such as artificial neural networks can be used. In this paper the use of these advanced techniques on experimental data is explored and discussed. The experiments are performed on a single cylinder cam-less (equipped with a Fully Variable Valve Timing (FVVT) system) research engine fueled with commercially available gasoline (95 ON).

Closed-loop electronic control is a proven and efficient way to optimize spark ignition engine performance and to control pollutant emissions. In-cylinder pressure sensors provide accurate information on the quality of combustion. The conductivity of combustion flames can alternatively be used as a measure of combustion quality through ion-current measurements. In this paper, combustion diagnostics through ion-current sensing are studied. A single cylinder research engine was used to investigate the effects of misfire, ignition timing, air to fuel ratio, compression ratio, speed and load on the ion-current signal. The ion-current signal was obtained via one, or both, of two additional, remote in-cylinder ion sensors (rather than by via the firing spark plug, as is usually the case). The ion-current signals obtained from a single remote sensor, and then the two remote sensors are compared.

Proton exchange membrane fuel cell (PEMFC) provides a promising future low carbon automotive powertrain solution. The catalyst layer (CL) is its core component which directly influences the output performance. PEMFC performance can be greatly improved by the effective optimization of CL composition. This work demonstrates a deep optimization of CL composition for improving the PEMFC performance, including the platinum (Pt) loading, Pt percentage of carbon-supported Pt and ionomer to carbon ratio of the anode and the cathode,. The simulation results by a PEMFC three-dimensional (3D) computation fluid dynamics (CFD) model coupled with the CL agglomerate model is used to train the artificial neural network (ANN) which can efficiently predict the current density under different CL composition. Squared correlation coefficient (R-square) and mean percentage error in the training set and validation set are 0.9867, 0.2635% and 0.9543, 1.1275%, respectively.

Energy management of battery electric vehicle (BEV) is a very important and complex multi-system optimisation problem. The thermal energy management of a BEV plays a crucial role in consistent efficiency and performance of vehicle in all weather conditions. But in order to manage the thermal management, it requires a significant number of temperature sensors throughout the car including high voltage batteries, thus increasing the cost, complexity and weight of the car. Virtual sensors can replace physical sensors with a data-driven, physical relation-driven or machine learning-based prediction approach. This paper presents a framework for the development of a neural network virtual sensor using a thermal system hardware-in-the-loop test rig as the target system. The various neural network topologies, including RNN, LSTM, GRU, and CNN, are evaluated to determine the most effective approach.

The importance of designing and sizing a thermal management system for electric vehicle powertrains cannot be overstated. Traditional approaches often rely on model-based system design using supplier reference component data, which can inadvertently lead to undisclosed errors arising from the interactions between the components and the environment. This paper introduces a novel test facility for battery electric vehicle thermal management technology, which has been designed for neural network virtual sensor and non-linear multi-in multi-out control development. The paper demonstrates how a digital twin of the test bench can used to support the development of such technology. Additionally, this paper presents preliminary results from the test bench revealing insights into the performance and interactions of key components. For instance, there is an observed 30% reduction in the maximum flow rate of the pump integrated into the test bench compared to the specified value.