The internal combustion engine faces increasing societal and governmental pressure to improve both efficiency and engine out emissions. Currently, research has moved from traditional combustion methods to new highly efficient combustion strategies such as Homogeneous Charge Compression Ignition (HCCI). However, predicting the exact value of engine out emissions using conventional physics-based or data-driven models is still a challenge for engine researchers due to the complexity the of combustion and emission formation. Research has focused on using Artificial Neural Networks (ANN) for this problem but ANN’s require large training datasets for acceptable accuracy. This work addresses this problem by presenting the development of a simple model for predicting the steady-state emissions of a single cylinder HCCI engine which is created using an metaheuristic optimization based Support Vector Machine (SVM). The selection of input variables to the SVM model is explored using five different feature sets, considering up to seven engine inputs. The best results are achieved with a model combining linear and squared inputs as well as cross correlations and their squares totaling 26 features. In this case the model fit represented by R² values were between 0.72 and 0.95. The best model fits were achieved for CO and CO₂, while HC and NO_x models have reduced model performance. Linear and non-linear SVM models were then compared to an ANN model. This comparison showed that SVM based models were more robust to changes in feature selection and better able to avoid local minimums compared to the ANN models leading to a more consistent model prediction when limited training data is available. The proposed machine learning based HCCI emission models and the feature selection approach provide insight into optimizing the model accuracy while minimizing the computational costs.

Homogeneous Charge Compression Ignition (HCCI) is a low temperature combustion technique with a high potential for reducing emissions while simultaneously improving fuel consumption. However, the high sensitivity to changing boundary conditions and low combustion stability at the edges of the operating range has lead to implementation challenges. Additionally, cyclic coupling through internal exhaust gas recirculation causes cyclic variations of the process, resulting in incomplete combustion, or even misfiring. Thus, consecutive cycles must be decoupled to increase the process stability. To achieve an accurate description of the coupling effects on a cycle-to-cycle and an inner-cyclic timescale, a novel measurement methodology is presented to generate data with a high variance. For this purpose, an active process excitation is performed to capture all relevant interactions between operating and feedback variables to enable modeling of the coupling effects on both timescales. To demonstrate the potential of the methodology, the generated data is used to design multiple input, multiple output (MIMO) models for both cyclic and inner-cyclic timescales. Artificial neural networks are then utilized to address the highly nonlinear process by taking advantage of the large amount of training data. Inverse process models are then used to implement a pure cycle-to-cycle and a multiscale MIMO closed-loop controller. Compared to state-of-the-art rule-based control approaches, the process stability and its thermodynamic efficiency are significantly improved. For the multiscale MIMO controller, a reduction of the standard deviation of the indicated mean effective pressure and the combustion phasing of more than 65% is achieved. In particular, the additional inner-cyclic feedback loop achieves a remarkable reduction of the standard deviation of approximately 35% and a 1.2% higher indicated efficiency compared to the cycle-to-cycle MIMO controller. The dynamic measurement with active in-cycle process excitation has proven to be an enabler for data-driven MIMO control of HCCI on multiple timescales.

Homogeneous Charge Compression Ignition (HCCI) combustion has been widely reported to offer high efficiency and ultra-low nitrogen oxide emissions compared to conventional spark-ignited (SI) combustion. However, reported efficiency benefits strongly depend on boundary conditions, engine hardware, and the chosen reference concept. This study presents a systematic experimental comparison between HCCI and SI combustion using gasoline and ethanol on the same single-cylinder research engine under unthrottled and otherwise identical operating conditions. Combustion stability, indicated efficiency, combustion phasing, and gaseous emissions are evaluated. The results show that HCCI combustion provides substantially reduced CO (ethanol: −61.1%; gasoline: −80.6%) and NO_x (ethanol: −96.1%; gasoline: −86.3%) emissions and superior combustion stability for both fuels. Ethanol further improves efficiency and emissions compared to gasoline. Contrary to common expectations reported in the literature, no universal efficiency advantage of HCCI combustion over SI operation is observed under the specific boundary conditions and with the investigated engine configuration of this study. A detailed loss analysis shows that, for the present setup, increased gas exchange and heat transfer losses offset the higher working cycle efficiency (without gas exchange) of HCCI combustion.