Search

High-performance and low-cost emission characteristic prediction is very crucial for diesel engine design optimization and emission aftertreatment control and diagnosis. In this paper, a novel hybrid model that combines Convolutional Neural Network (CNN) and Gated Recurrent Unit Network (GRU) was proposed to predict the emission characteristic from diesel engines, encompassing CO, THC, CO2, NOx, exhaust temperature and exhaust pressure. Nine operating parameters from WHTC and WHSC cycles, including speed, torque, intake pressure, intake flow, intake temperature, oil pressure, fuel rate, oil temperature, water temperature, were considered as inputs. Firstly, the importance of each variable is evaluated by Random Forests algorithm to determine the optimal inputs for each emission characteristic parameter and reduce redundancy. Then the effect of different hyperparameters on the model performance was discussed in detail and PSO algorithm was used to obtain the optimal hyperparameters. Finally, the CNN-GRU hybrid model was assessed for its generalization and compared with ANN, LSTM and GRU models. The result demonstrates that the CNN-GRU hybrid model with PSO optimization has excellent prediction performance in either the training dataset or the validation dataset. The average value of R² is 0.993 in the training dataset and 0.985 on the validation dataset. In the test dataset, the average R² is 0.961, showing a minor decrease of 3.19% and 2.47% compared to the training and validation dataset, respectively. This indicates that the CNN-GRU hybrid model has strong generalization ability. Compared with other algorithms in the test dataset, the CNN-GRU hybrid model exhibits better comprehensive performance, with the average R² value exceeding that of ANN, LSTM and only GRU by 5.96%, 2.69% and 3.23%, respectively. [ABSTRACT FROM AUTHOR]

Predictive analytics utilizing machine learning algorithms play a pivotal role in various domains, including the profiling of carbon dioxide (CO2) emissions. This research paper delves into an extensive exploration of different algorithms, encompassing neural networks with diverse architectures, optimization, training, ensemble, and specialized algorithms. The primary objective of this research is to evaluate the efficacy of supervised and unsupervised algorithms, including Deep Belief Networks, Feed Forward Neural Networks, Gradient Boosting, and Regression, as well as Convolutional Neural Networks, Gaussian, Grey, and Markov models, and clustering and optimization algorithms. The study places particular emphasis on data-driven methodologies and cross-validation techniques with an evaluation of the learning models entailing comprehensive training, validation, and testing, employing evaluation metrics such as R2, MAE, and RMSE. The study employs correlation analysis to examine the relationship between input parameters and emission characteristics. The research highlights the advantageous attributes of these algorithms in accurately forecasting CO2 emissions, evaluating energy sources, improving prediction accuracy, and estimating emissions. Notably, deep learning, Artificial Neural Networks (ANN), and Support Vector Machines (SVM) demonstrate effectiveness across diverse industries, while the Modified Regularized Fast Orthogonal-Extreme Learning Machine (MRFO-ELM) algorithm optimizes predictions specifically related to coal chemical emissions. Hybrid techniques demonstrate accuracy in predicting carbon emissions and energy consumption, whereas gray prediction models provide reliable estimates even with limited data. However, it is important to acknowledge certain limitations, including data requirements, potential inaccuracies arising from complex factors, constraints faced by developing countries, and the impact of electric vehicle expansion on the power grid. To optimize models, a survey is conducted, involving customization of parameters and learning rates, while exploring various performance metrics to evaluate model accuracy. The research outcomes contribute to the effective monitoring of CO2 emissions in operational environments, thereby aiding executive decision-making processes. [ABSTRACT FROM AUTHOR]

Solar energy is a feasible alternative to traditional sources of energy. However, the intermittent and random nature of photovoltaic power generation poses a challenge to the reliable and cost-effective functioning of active distribution systems. Photovoltaic power forecasting is a critical instrument in solar photovoltaic power plants for improving energy delivery quality to the grid and lowering weather-related ancillary expenses. This paper proposes a new hybrid deep learning model called convolutional neural network (CNN) and gated recurrent unit (GRU). Convolutional layers have the potential to learn complicated characteristics from raw data automatically. GRU layers, on the other hand, can learn numerous parallel sequences of input data immediately. The generated prediction model was compared to newly established deep-learning-based algorithms of CNN, GRU, recurrent neural network (RNN), and long short-term memory (LSTM) to test CNN-GRU performance. The photovoltaic power data collected in Alice Springs, Australia, is utilized. In terms of accuracy and efficacy, the values of mean absolute error (MAE), root mean squared error (RMSE), mean bias error (MBE), mean squared error (MSE) and the coefficient of determination (R2) were 0.068 kW, 0.103 kW, 0.032 kW, 0.010 kW, and 99.91%, respectively. The experimental findings show that the proposed model outperforms alternative baseline models. [ABSTRACT FROM AUTHOR]