Scholarlink

<p>This study investigates the impact of plastic-based fuel and diesel additives on biodiesel blends’ combustion performance and emissions in a single-cylinder diesel engine. The primary objective is to evaluate the effects of fuel blends on brake-specific fuel consumption (BSFC), brake power (BP), and exhaust emissions, including carbon monoxide (CO), hydrocarbons (HC), and nitrogen oxides (NO<sub>x</sub>). FTIR analysis identified key functional groups in all fuel samples, showing characteristic C–H, C=O, and C–O stretching peaks. Three fuel samples were tested under varying engine loads (10% to 100%) and speeds (2000&#xa0;rpm and 3000&#xa0;rpm). Additionally, a Medium Neural Network (MNN) model was developed to predict NO<sub>x</sub> emissions, with performance assessed using the correlation coefficient (R) and mean square error (MSE). The results revealed that the PP10 + Add blend achieved the lowest BSFC, with a reduction of up to 8.5% compared to B10, indicating enhanced combustion efficiency. BP increased proportionally with engine load for all fuel blends at both 2000 and 3000&#xa0;rpm. While B10 delivered the highest BP at 2000&#xa0;rpm, PP10 showed significant BP improvement at 3000&#xa0;rpm. CO emissions were significantly reduced, attributed to improved oxidation and more complete combustion. HC emissions decrease as the engine load increases from 10 to 100%, which indicates more complete combustion at higher loads. At both 2000 and 3000&#xa0;rpm, the B10 and PP10 + Add blends exhibit the lowest HC emissions at 75% load compared to PP10. The MNN model demonstrated strong predictive performance for NO<sub>x</sub> emissions, with R values of 0.9531 for training, 0.9488 for validation, and 0.9297 for testing, confirming its reliability in analysing combustion and emission behaviours. Overall, while the incorporation of plastic-based fuel and additives enhances combustion efficiency and reduces CO emissions, it also leads to increased HC and NO<sub>x</sub> emissions, likely due to altered ignition characteristics and fuel–air mixing dynamics. These findings offer valuable insights into fuel modification strategies aimed at optimizing diesel engine performance and emission profiles.</p> Graphical abstract <p></p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

NOx Prediction in a Diesel Engine Using a Medium Neural Network (MNN) with Additive-Enhanced Biodiesel and Polypropylene

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年4月27日

查看全文
  • Mohammad Nor Khasbi Jarkoni,
  • Muhamad Akhil Zakhwan Shahriman,
  • Norazlina Abdul Nasir,
  • Wan Nurdiyana Wan Mansor,
  • Sheikh Alif Ali,
  • Che Wan Mohd Noor,
  • Anuar Abu Bakar,
  • Nurul Huda Abd Kadir,
  • Nurul Ashraf Razali

摘要

This study investigates the impact of plastic-based fuel and diesel additives on biodiesel blends’ combustion performance and emissions in a single-cylinder diesel engine. The primary objective is to evaluate the effects of fuel blends on brake-specific fuel consumption (BSFC), brake power (BP), and exhaust emissions, including carbon monoxide (CO), hydrocarbons (HC), and nitrogen oxides (NOx). FTIR analysis identified key functional groups in all fuel samples, showing characteristic C–H, C=O, and C–O stretching peaks. Three fuel samples were tested under varying engine loads (10% to 100%) and speeds (2000 rpm and 3000 rpm). Additionally, a Medium Neural Network (MNN) model was developed to predict NOx emissions, with performance assessed using the correlation coefficient (R) and mean square error (MSE). The results revealed that the PP10 + Add blend achieved the lowest BSFC, with a reduction of up to 8.5% compared to B10, indicating enhanced combustion efficiency. BP increased proportionally with engine load for all fuel blends at both 2000 and 3000 rpm. While B10 delivered the highest BP at 2000 rpm, PP10 showed significant BP improvement at 3000 rpm. CO emissions were significantly reduced, attributed to improved oxidation and more complete combustion. HC emissions decrease as the engine load increases from 10 to 100%, which indicates more complete combustion at higher loads. At both 2000 and 3000 rpm, the B10 and PP10 + Add blends exhibit the lowest HC emissions at 75% load compared to PP10. The MNN model demonstrated strong predictive performance for NOx emissions, with R values of 0.9531 for training, 0.9488 for validation, and 0.9297 for testing, confirming its reliability in analysing combustion and emission behaviours. Overall, while the incorporation of plastic-based fuel and additives enhances combustion efficiency and reduces CO emissions, it also leads to increased HC and NOx emissions, likely due to altered ignition characteristics and fuel–air mixing dynamics. These findings offer valuable insights into fuel modification strategies aimed at optimizing diesel engine performance and emission profiles.

Graphical abstract