吴玥,梁兴雨,屠丹红.基于神经网络的柴油机活塞环组窜气量预测方法研究[J].内燃机工程,2024,45(6):60-70.
基于神经网络的柴油机活塞环组窜气量预测方法研究
Study on the Prediction Method of Diesel Engine Piston Ring Pack Blow-by Based on Neural Network
DOI：10.13949/j.cnki.nrjgc.2024.06.007
关键词:柴油机  活塞环组  窜气量  预测模型  粒子群优化
Key Words:diesel engine  piston ring pack  blow-by  prediction model  particle swarm optimization（PSO）
基金项目:国家自然科学基金项目（52342606）；工信部船机重大专项之船用发动机高可靠性设计和验证关键技术项目
作者单位E-mail
吴玥 天津大学 先进内燃动力全国重点实验室，天津 300072 wuyue220@tju.edu.cn 
梁兴雨* 天津大学 先进内燃动力全国重点实验室，天津 300072 lxy@tju.edu.cn 
屠丹红 中船动力研究院有限公司，上海 200129 18d00084@cspi.net.cn 
摘要点击次数: 349
全文下载次数: 87
摘要:针对发动机中出现密封不严而造成的发动机动力性和经济性下降及重要零部件损坏的现象，以某柴油机的单缸试验机为研究对象，对活塞环组的密封性能进行仿真计算，针对开口间隙、倒角长度、径向弹力、工作温度等5个输入和窜气量1个输出，建立窜气量反向传播神经网络（back propagation neural network， BPNN）预测模型，并通过灰狼优化（grey wolf optimization， GWO）算法、鲸鱼优化算法（whale optimization algorithm， WOA）、遗传算法（genetic algorithm， GA）、粒子群优化（particle swarm optimization， PSO）算法进行优化，提高模型的预测性能。结果表明，粒子群优化–反向传播（particle swarm optimiation-back propagation， PSO-BP）预测模型对窜气量具有较强的泛化能力和预测性能。PSO-BP预测模型的高准确性和稳定性为发动机设计和维护提供了强有力的决策支持工具，有助于实现更精确的故障诊断和预测性维护，降低运营成本，提高发动机的整体性能和经济效益。
Abstract:In response to the phenomenon of poor sealing in the engine， which leads to a decrease in engine power and economy， as well as damage to important components， a single cylinder test engine of a certain diesel engine was taken as the research object. The sealing performance of the piston ring pack was simulated and calculated. A back propagation neural network （BPNN） prediction model for gas leakage was established for five inputs including opening clearance， chamfer length， radial elasticity， working temperature， and one output of blow-by. Four algorithms were used to improve the prediction performance of the model， namely grey wolf optimization （GWO）， whale optimization algorithm （WOA）， genetic algorithm （GA）， and particle swarm optimization （PSO）.The results indicate that the PSO-BP prediction model has strong generalization ability and predictive performance for blow-by. The high accuracy and stability of the particle swarm optimization-back propagation （PSO-BP） prediction model provide a powerful decision support tool for engine design and maintenance， helping to achieve more accurate fault diagnosis and predictive maintenance， reduce operating costs， and improve the overall performance and economic benefits of the engine.
查看全文  HTML   查看/发表评论