Integrated machine learning-based intelligent optimization for clean and efficient operation of diesel engines and diesel particulate filters in dual modes

Abstract

To ensure the clean and efficient operation of diesel engines while achieving the safe and effective regeneration of the diesel particulate filter (DPF), an intelligent optimization method is proposed for dual-mode engine operation (normal and regeneration modes). Specifically, an integrated prediction model combining the Extreme Learning Machine (ELM) and Backpropagation Neural Network (BPNN) is developed, while a multi-strategy Salp Swarm Algorithm (MSSA) is introduced to improve model accuracy. By utilizing the MSSA-ELM-BP model combined with the Pareto Envelope-based Selection Algorithm II (PESA-II), a multi-objective optimization method based on multi-model fusion is proposed. This method optimizes both engine emissions and DPF regeneration performance under dual modes. Experimental results from steady-state, transient, and regeneration conditions demonstrate significant performance improvements. In normal mode, NO_x emissions were reduced from 886.72 ppm to 829.88 ppm, CO from 113.84 ppm to 90.83 ppm, and smoke opacity from 0.17 FSN to 0.16 FSN under steady conditions. During WHTC testing, emission reductions were evident, with NO_x dropping to 392.6 ppm and carbon smoke mass concentration decreasing to 1.03 mg/m³ . In regeneration mode, T4, T5, and O₂ concentrations improved by 6.10 %, 2.90 %, and 18.86 %, respectively, complemented by a 36.40 % increase in T4 temperature, while NO_x, CO, HC, smoke, and BSFC decreased by an average of 10.72 %, 4.12 %, 3.24 %, 11.48 %, and 0.24 %, respectively. Additionally, after-optimization regeneration tests confirmed a safe increase in DPF inlet temperature to 560°C and substantial reductions in peak pressure drop and residual smoke levels, demonstrating the effectiveness and safety of the proposed method. This intelligent optimization strategy provides a novel and practical approach to enhancing the clean and efficient operation of diesel engines and aftertreatment systems, contributing to lower emissions, improved fuel efficiency, and sustainable energy utilization.