23 October 2022 in Expert Systems with Applications
7 May 2020 in IEEE Sensors Journal
MHMohammad Reza Hasheminasab
Mohammad Reza Hasheminasab
Proton exchange membrane fuel cell (PEMFC) is a rich source of renewable energy. A non-destructive prediction method is needed to determine the content of water in the PEMFC. In the gas channel of a transparent PEMFC, water is detected with image processing. This method has a high computational cost and is sensitive to the initial position of the camera and ambient lighting. In this paper, the deep neural network (DNN) has been trained to learn the transparent PEMFC’s labeled images as a way to determine the content of water, limit human interference and employed in a real-time process. This DNN model is a virtual sensor for measuring the water coverage ratio. To produce the label of images, all data are divided into 6 classes based on the percentage of water coverage ratio. Through analyzing the number of each class, the imbalance data is unfolded. To overcome this problem, random oversampling and undersampling techniques are used. The images and the classes are considered as the input and output of the DNN, respectively. Also, the region of water accumulation in the gas channel can be recognized with robustness to the environmental conditions. Final results of 4-41-64-120-99-6 nodes for DNN layers were derived due to GA optimization. Accuracy of 96.77% and 94.23% in train and test data have been achieved. This DNN, processes the images up to 10 times faster than image processing. Also, the region of water accumulation in the gas channel can be recognized.
10 May 2025 in International Journal of Hydrogen Energy
5 August 2025 in Control Engineering Practice
2 March 2026 in Control Engineering Practice
1 January 2026 in Transactions of the Canadian Society for Mechanical Engineering
Hydrogen/diesel dual-fuel (HDDF) engines are being investigated as a promising solution for reducing carbon dioxide and providing a transition to zero-carbon fuels. However, modeling these engines presents challenges due to the complex interactions between the two fuels. Computational fluid dynamics modeling is often inaccurate across various operating conditions, while conventional data-driven approaches frequently face issues related to a lack of interpretability. Furthermore, based on the complexity of the system, there is a shortage of accurate physics-based modeling. This study explores temporal Kolmogorov–Arnold networks (T-KANs) as an alternative to conventional machine learning (ML) models. T-KANs offer a structured framework for function approximation that can efficiently learn the underlying dependencies with historical data while ensuring interpretability. A T-KAN model is developed and trained using experimental HDDF engine data, demonstrating its ability to predict performance metrics, including indicated mean effective pressure and nitrogen oxides. A comparative analysis of seven different models of well-known ML methods highlights the advantages of T-KANs in terms of accuracy, generalization, and computational efficiency. Given its computational efficiency and a coefficient of determination of 0.956, the T-KAN with a 10-lookback is suitable for a model-based controller. This configuration effectively utilizes historical data while operating independently of sensor feedback.