Performance prediction of hydrogen enriched compressed natural gas engine based on IMPSO-BPNN
-
摘要: 为进一步提高传统粒子群算法优化反向传播神经网络(PSO-BPNN)模型的性能，基于惯性权重和加速因子对粒子群优化的影响机制，提出一种采用非线性递减惯性权重和非线性加速因子调整策略的改进粒子群算法优化反向传播神经网络(IMPSO-BPNN)方法；将IMPSO-BPNN方法应用于天然气掺氢(HCNG)发动机扭矩、等效燃料比消耗和NOx比排放等性能参数的回归分析与预测，并从预测精度、泛化能力和收敛速度3个方面与其他神经网络方法进行了比较，包括PSO-BPNN、遗传算法优化反向传播神经网络(GA-BPNN)和反向传播神经网络(BPNN)方法。研究结果表明：燃空比和点火提前角均可显著影响HCNG发动机的扭矩、等效燃料比消耗和NOx比排放；以扭矩为预测变量，最优IMPSO-BPNN模型的平均绝对百分比误差分别比PSO-BPNN、GA-BPNN和BPNN方法所建立的最优模型小5.85%、12.62%和17.96%，且最优IMPSO-BPNN模型的相关系数也最大，达到了0.999 86，说明IMPSO-BPNN方法所建立模型的预测性能和泛化能力总体上优于其他方法；以NOx比排放为预测变量，最优PSO-BPNN和最优IMPSO-BPNN模型的CPU运行时间比最优GA-BPNN模型均减少约95%，说明与GA-BPNN方法相比，PSO-BPNN和IMPSO-BPNN方法在时间维度上优越性明显。可见，本文提出的IMPSO-BPNN方法相比PSO-BPNN和GA-BPNN方法在预测性能和泛化能力方面均有显著的优势，同时能够保证较高的计算效率。Abstract: To further improve the performance of traditional particle swarm optimization back-propagation neural network (PSO-BPNN) model, based on the influence mechanisms of inertia weight and acceleration factor on particle swarm optimization, an improved particle swarm optimization back-propagation neural network (IMPSO-BPNN) method adopting non-linear decreasing inertia weight and non-linear acceleration factor was proposed. The IMPSO-BPNN method was applied to the regression analysis and prediction of performance parameters such as torque, equivalent brake specific fuel consumption, and brake specific NOx emission of a hydrogen enriched compressed natural gas (HCNG) engine. It was also compared with other neural network methods in terms of prediction accuracy, generalization ability, and convergence speed, including PSO-BPNN, genetic algorithm optimized back-propagation neural network (GA-BPNN), and back-propagation neural network (BPNN) methods. Research results show that the fuel-air ratio and spark advance angle can significantly affect the torque, equivalent brake specific fuel consumption, and brake specific NOx emissions of the HCNG engine. With torque as the predictive variable, the average absolute percentage error of the optimal IMPSO-BPNN model is 5.85%, 12.62%, and 17.96% smaller than those of PSO-BPNN, GA-BPNN, and BPNN models, respectively, and the correlation coefficient of the optimal IMPSO-BPNN model is 0.999 86, also the highest among these models, which indicates that the prediction performance and generalization ability of the model established by the IMPSO-BPNN method are generally superior to those established by other methods. With brake specific NOx emission as the predictive variable, the CPU running times reduce by 95% in both the optimal PSO-BPNN model and the optimal IMPSO-BPNN model compared with the optimal GA-BPNN model, which demonstrates the superiority of PSO-BPNN and IMPSO-BPNN methods to the GA-BPNN method in terms of time dimension. Therefore, compared with PSO-BPNN and GA-BPNN methods, the proposed IMPSO-BPNN method has significant advantages in prediction performance and generalization ability, and ensures high computing efficiency.
-
图 4 不同转速下T、BSFC和BSNOx随Φ和θsp的变化曲线
Figure 4. Variation curves of T, BSFC and BSNOx with Φ and θsp at various rotate speeds
图 5 预测变量为T时不同方法所建立预测模型的EMAPE
Figure 5. EMAPE of prediction models established by various methods with T as predictive variable
图 6 预测变量为T时不同方法所建立预测模型的R
Figure 6. R of prediction models established by various methods with T as predictive variable
图 7 预测变量为T时各方法所建立最优预测模型的性能对比
Figure 7. Performance comparison of optimal prediction models established by various methods with T as predictive variable
图 8 预测变量为T时最优IMPSO-BPNN模型的预测结果
Figure 8. Predictin results of optimal IMPSO-BPNN model with T as predictive variable
图 9 预测变量为BSFC时不同方法所建立预测模型的EMAPE
Figure 9. EMAPE of prediction models established by various methods with BSFC as predictive variable
图 10 预测变量为BSFC时不同方法所建立预测模型的R
Figure 10. R of prediction models established by various methods with BSFC as predictive variable
图 11 预测变量为BSFC时各方法所建立最优预测模型的性能对比
Figure 11. Performance comparison of optimal prediction models established by various methods with BSFC as predictive variable
图 12 预测变量为BSFC时最优IMPSO-BPNN模型的预测结果
Figure 12. Prediction results of optimal IMPSO-BPNN model with BSFC as predictive variable
图 13 预测变量为BSNOx时不同方法所建立预测模型的EMAPE
Figure 13. EMAPE of prediction models established by various methods with BSNOx as predictive variable
图 14 预测变量为BSNOx时不同方法所建立预测模型的R
Figure 14. R of prediction models established by various methods with BSNOx as predictive variable
图 15 预测变量为BSNOx时各方法所建立最优预测模型的性能对比
Figure 15. Performance comparison of optimal prediction models established by various methods with BSNOx as predictive variable
图 16 预测变量为BSNOx时最优IMPSO-BPNN模型的预测结果
Figure 16. Prediction results of optimal IMPSO-BPNN model with BSNOx as predictive variable
表 1 WP6NG240E50发动机性能参数
Table 1. Performance parameters of WP6NG240E50 engine
参数 数值 总排量/L 6.75 气缸数 6 压缩比 11.5 缸径/mm 105 行程/mm 130 额定功率/kW 177 额定转速/(r·min-1) 2 300 
下载: 导出CSV
表 2 发动机试验工况范围
Table 2. Ranges of engine test conditions
参数 范围 转速/(r·min-1) 1 000~2 000 进气歧管压力/kPa 60~180 燃空比 0.60~0.84 点火提前角/(°) CA BTDC 9~36 注：BTDC，Before Top Dead Center，上止点前。
下载: 导出CSV
-
[1] GANDHIDASAN P, ERTAS A, ANDERSON E E. Review of methanol and compressed natural gas (CNG) asalternative for transportation fuels[J]. Journal of Energy Resources Technology, 1991, 113(2): 101-107. doi: 10.1115/1.2905782 [2] KHAN M I, YASMEEN T, KHAN M I, et al. Research progress in the development of natural gas as fuel for road vehicles: a bibliographic review (1991—2016)[J]. Renewable and Sustainable Energy Reviews, 2016, 66: 702-741. doi: 10.1016/j.rser.2016.08.041 [3] TANG Cheng-long, ZHANG Ying-jia, HUANG Zuo-hua. Progress in combustion investigations of hydrogen enriched hydrocarbons[J]. Renewable and Sustainable Energy Reviews, 2014, 30: 195-216. doi: 10.1016/j.rser.2013.10.005 [4] SANCHEZ A L, WILLIAMS F A. Recent advances in understanding of flammability characteristics of hydrogen[J]. Progress in Energy and Combustion Science, 2014, 41: 1-55. doi: 10.1016/j.pecs.2013.10.002 [5] 范英杰. 车用氢气发动机研究进展综述[J]. 内燃机与配件, 2021(3): 40-42.FAN Ying-jie. Summary of research progress on hydrogen engines for vehicles[J]. Internal Combustion Engine and Parts, 2021(3): 40-42. (in Chinese) [6] MA Fan-hua, WANG Yu. Study on extension of operation limit through hydrogen enrichment in a natural gas spark-ignition engine[J]. International Journal of Hydrogen Energy, 2008, 33: 1416-1424. doi: 10.1016/j.ijhydene.2007.12.040 [7] MA Fan-hua, WANG Yu, LIU Hai-quan, et al. Effects of hydrogen addition on cycle-by-cycle variations in a lean burn natural gas spark-ignition engine[J]. International Journal of Hydrogen Energy, 2008, 33(2): 823-831. doi: 10.1016/j.ijhydene.2007.10.043 [8] LUO Si-jie, MA Fan-hua, MEHRA R K, et al. Deep insights of HCNG engine research in China[J]. Fuel, 2019, 263: 116612. [9] MA Fan-hua, LI Shun, ZHAO Jian-biao, et al. Effect of compression ratio and spark timing on the power performance and combustion characteristics of an HCNG engine[J]. International Journal of Hydrogen Energy, 2012, 37: 18486-18491. doi: 10.1016/j.ijhydene.2012.08.134 [10] MA Fan-hua, WANG Ming-yue, JIANG Long, et al. Performance and emission characteristics of a turbocharged spark-ignition hydrogen-enriched compressed natural gas engine under wide open throttle operating conditions[J]. International Journal of Hydrogen Energy, 2010, 35: 12502-12509. doi: 10.1016/j.ijhydene.2010.08.053 [11] ZHENG Jian-jun, HU Er-jiang, HUANG Zuo-hua, et al. Combustion and emission characteristics of a spray guided direct-injection spark-ignition engine fueled with natural gas-hydrogen blends[J]. International Journal of Hydrogen Energy, 2011, 36: 11155-11163. doi: 10.1016/j.ijhydene.2011.05.119 [12] MIAO Hai-yan, LU Lin, HUANG Zuo-hua. Flammability limits of hydrogen-enriched natural gas[J]. International Journal of Hydrogen Energy, 2011, 36: 6937-6947. doi: 10.1016/j.ijhydene.2011.02.126 [13] ORTENZI F, CHIESA M, SCARCELLI R, et al. Experimental tests of blends of hydrogen and natural gas in light-duty vehicles[J]. International Journal of Hydrogen Energy, 2008, 33: 3225-3229. doi: 10.1016/j.ijhydene.2008.01.050 [14] MUNSHI S R, NEDELCU C, HARRIS J, et al. Hydrogen blended natural gas operation of a heavy duty turbocharged lean burn spark ignition engine[C]//SAE. 2004 Powertrain and Fluid Systems Conference and Exhibition. Warrendale: SAE, 2004: 1-15. [15] BHASKER J P, PORPATHAM E. Effects of compression ratio and hydrogen addition on lean combustion characteristics and emission formation in a compressed natural gas fuelled spark ignition engine[J]. Fuel, 2017, 208: 260-270. doi: 10.1016/j.fuel.2017.07.024 [16] 周锐, 郑建, 王明达, 等. 外部燃料重整掺氢对天然气发动机性能的影响[J]. 车用发动机, 2021(3): 20-25.ZHOU Rui, ZHENG Jian, WANG Ming-da, et al. Influence of hydrogen mixing based on external fuyel reforming on natural gas engine performance[J]. Vehicle Engine, 2021(3): 20-25. (in Chinese) [17] ZHANG Pan, GAO Wen-zhi, LI Yong, et al. Misfire detection of diesel engine based on convolutional neural networks[J]. Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering, 2021, 235(8): 2148-2165. doi: 10.1177/0954407020987077 [18] CHEN Wei, PAN Jian-feng, ZUO Qing-song, et al. Combustion performance improvement of a diesel fueled Wankel stratified-charge combustion engine by optimizing assisted ignition strategy[J]. Energy Conversion and Management, 2020, 205: 112324. doi: 10.1016/j.enconman.2019.112324 [19] LIN S W, YING K C, CHEN S C, et al. Particle swarm optimization for parameter determination and feature selection of support vector machines[J]. Expert Systems with Applications, 2008, 35: 1817-1824. doi: 10.1016/j.eswa.2007.08.088 [20] MEHRA K R, DUAN Hao, LUO Si-jie, et al. Experimental and artificial neural network (ANN) study of hydrogen enriched compressed natural gas (HCNG) engine under various ignition timings and excess air ratios[J]. Applied Energy, 2018, 228: 736-754. doi: 10.1016/j.apenergy.2018.06.085 [21] MARIANI V C, OCH S H, COELHO L D, et al. Pressure prediction of a spark ignition single cylinder engine using optimized extreme learning machine models[J]. Applied Energy, 2019, 249: 204-221. doi: 10.1016/j.apenergy.2019.04.126 [22] VONG C M, WONG P K. Engine ignition signal diagnosis with wavelet packet transform and multi-class least squares support vector machines[J]. Expert Systems with Applications, 2011, 38(7): 8563-8570. doi: 10.1016/j.eswa.2011.01.058 [23] HOANG A T, NIZETIC S, ONG H C, et al. A review on application of artificial neural network (ANN) for performance and emission characteristics of diesel engine fueled with biodiesel-based fuels[J]. Sustainable Energy Technologies and Assessments, 2021, 47: 101416. doi: 10.1016/j.seta.2021.101416 [24] SIMSEK S, USLU S, SIMSEK H. Proportional impact prediction model of animal waste fat-derived biodiesel by ANN and RSM technique for diesel engine[J]. Energy, 2022, 239: 122389. doi: 10.1016/j.energy.2021.122389 [25] SABOUR M H, BEHESHTI A, ESFAHANIAN V. Reduction of experimental effort in conventional engine calibration process by using reduced order model[J]. Proceedings of the Institution of Mechanical Engineers, Part D: Journal of Automobile Engineering, 2021, 235(2/3): 639-648. [26] DUAN Hao, YIN Xiao-jun, KOU Hai-liang, et al. Prediction of combustion promotion effect of high/low-frequency AC electric fields based on machine learning method[J]. Fuel, 2023, 346: 128348. doi: 10.1016/j.fuel.2023.128348 [27] MA Fan-hua, WANG Yu, DING Shang-fen, et al. Twenty percent hydrogen-enriched natural gas transient performance research[J]. International Journal of Hydrogen Energy, 2009, 34: 6523-6531. doi: 10.1016/j.ijhydene.2009.05.135 [28] 李沁璘. 人工神经网络综述[J]. 科学与信息化, 2021(7): 181-182.LI Qin-lin. Review of artificial neural networks[J]. Science and Informatization, 2021(7): 181-182. (in Chinese) [29] 崔长彩, 李兵, 张认成. 粒子群优化算法[J]. 华侨大学学报(自然科学版), 2006, 27(4): 343-347.CUI Chang-cai, LI Bing, ZHANG Ren-cheng. Particle swarm optimization[J]. Journal of Huaqiao University (Natural Science), 2006, 27(4): 343-347. (in Chinese) [30] 崔峰, 王汉封, 舒卓乐. 基于PSO-BP神经网络的隧道内气动压力幅值预测[J]. 中南大学学报(自然科学版), 2023, 54(9): 3752-3761.CUI Feng, WANG Han-feng, SHU Zhuo-le. Prediction of aerodynamic pressure amplitude in tunnel based on PSO-BP neural network[J]. Journal of Central South University (Science and Technology), 2023, 54(9): 3752-3761. (in Chinese) [31] BAI Bin, ZHANG Jun-yi, WU Xuan, et al. Reliability prediction-based improved dynamic weight particle swarm optimization and back propagation neural network in engineering systems[J]. Expert Systems with Applications, 2021, 177: 114952. -
点击查看大图
计量
- 文章访问数: 160
- HTML全文浏览量: 55
- PDF下载量: 34
- 被引次数: 0
