In this paper, a Self-learning Collective Intelligence Differential Evolution (SLCIDE) algorithm was proposed to optimize both the architecture and parameters of a Feedforward Neural Network (FNN). In order to improve the exploration-exploitation capability, a new Collective Intelligence (CI) based mutation operator was proposed by mixing some promising donor vectors in the current population. Besides, a self-learning mechanism which was designed to adaptively select m top ranked donor vectors was developed by using a widely used unsupervised learning method, k-means. As a result, the proposed approach can be more adaptive and statistically powerful on versatile problems. Then, we evaluated the performances of the proposed SLCIDE approach by studying some numerical optimization problems of CEC 2013 with D = 30 and D = 50. Further, the proposed SLCIDE method was applied to train a FNN on four most popular datasets, resulting in very competitive performances. The comprehensive experimental results have demonstrated that the presented SLCIDE method obtain better results compared with other state-of-the-art algorithms.

An accurate model plays an important role in control strategy development of smart ships. For the control-oriented engine models, calibration by experienced personnel is key to outputting high accuracy. However, the dual-fuel engine runs in liquid fuel mode, gas fuel mode, and fuel sharing mode. It is impossible to tune a single model or a set of parameters for the dual-fuel engine under all operating modes and loads. On the basis of our experience and practice, a Wiebe-based heat release rate model is used. To make the Wiebe model available for the dual-fuel engine, the Wiebe parameters are assumed to be linear functions. The combustion beginning angle is modeled as a function of fuel quantity in liquid fuel mode and as a look-up table in gas fuel mode for all loads. The combustion duration and the combustion distribution factor are modeled as a function of fuel quantity and engine revolution both in liquid fuel mode and in gas fuel mode. In fuel sharing mode, the heat release rate is modeled as a combination of the heat release rate models in liquid fuel mode and gas fuel mode. This model is called the SL model. For a further discussion, four types of combinations in fuel sharing mode are investigated. In addition, in liquid fuel mode and gas fuel mode, the combustion duration model and the combustion distribution factor model are replaced by the Woschni/Anisits model, which was specifically used in the diesel engine. This variation of model is called the WA model. To validate our hypothesis and models, the Wiebe parameters in liquid fuel mode and gas fuel mode are given, four types of combinations and two cases of comparisons in fuel sharing model are discussed, and the engine performance is checked and analysed. Results show that for the SL model, the average RMSE is 1.45% in the liquid fuel mode, 2.22% in the gas fuel mode, and 2.53% in the fuel sharing mode. For the WA model, the RMSE of the NOx is 9.79% in liquid fuel mode and 45.20% in gas fuel mode. Its maximum error reaches −65.54%. The proposed SL model is accurate and can generate Wiebe parameters that are better than the carefully tuned parameters. The WA model is not suitable for engine models that require NOx-emission-related parameters.

To enhance technical support for monitoring the condition and managing the overall health of marine diesel engines, this study introduces a baseline model for marine diesel engine exhaust gas temperature (EGT). This model integrates the Improved Slime Mould Algorithm (ISMA) with the Generalized Regression Neural Network (GRNN) for real-time EGT monitoring. For the experimental dataset, this study introduces methods such as Mahalanobis distance, sliding window filtering algorithm, and Pearson’s algorithm to establish a comprehensive data processing workflow. Subsequently, the ISMA algorithm is employed to optimize the hyperparameter (σ) value in the GRNN, and the optimized σ value is then input into the GRNN model to establish the optimal baseline model for EGT prediction. Validation is conducted using historical ’8L20’ diesel engine data. Experimental results demonstrate that the ISMA-GRNN baseline model leverages the respective advantages of both methods and exhibits superior performance in prediction accuracy, convergence speed, and robustness compared to RNN, BiLSTM, GRNN, GA-GRNN, and PSO-GRNN models. The model effectively utilizes historical monitoring data to deeply extract internal information, enabling accurate baseline model establishment of diesel engine exhaust gas temperature. This not only contributes to the realization of condition monitoring and early warning functions for ship engines, improving ship energy efficiency and environmental performance, but also provides a new approach for intelligent operation and maintenance of ships.

The jet impingement technique is currently one of the most efficient cooling solutions for highly reinforced pistons of large two-stroke engines. To study the heat transfer characteristics of piston, experimental and numerical investigations with a piston cooling bore impinged by SAE 30 oil were carried out. To investigate the heat transfer coefficient distributions over the target bore, the wall temperatures of the cooling bore were measured by thermocouples, which will also be used in the numerical calculation. The jet Reynolds number (Re) ranges from 220 to 330, and the jet-to-plate spacing ratios (H/D) range from 10 to 30. Results show that jet-to-plate spacing ratios have a slight effect on the heat transfer coefficient for this low Reynolds numbers impingement which is quite different from high Reynolds numbers flow. There are both three peaks of the local heat transfer coefficient for Re = 330 and 280 along the x-axis direction. However, only two peaks occur when Re = 280. The heat transfer coefficient increases with the increase of Reynolds number when x/D < 0.22 or x/D > 1.77 while the variation is contrary when 0.22 < x/D < 1.77. The average heat transfer coefficient of the top surface region is far larger than other regions and decreases significantly in the upper chamfered region. While it is almost identical in the cylindrical region for different Reynolds numbers. This study provides the heat transfer characteristics of piston cooling with SAE30 oil and can be used for the piston optimization of large two-stroke engines with high cooling performance requirements.