基于GRNN算法的船用柴油机性能曲线模拟与油耗率预测  

Performance Curve Simulation and Fuel Consumption Rate Prediction Based on GRNN Algorithm for Marine Diesel Engine

作　　者：黄加亮[1,2] 刘磊[1] 陈景锋[1,2] 

机构地区：[1]集美大学轮机工程学院,福建厦门361021 [2]福建省船舶与海洋工程重点实验室,福建厦门361021

出　　处：《船舶工程》2013年第3期37-40,共4页Ship Engineering

基　　金：国家自然科学基金资助项目(51279066);福建省自然科学基金资助项目(2012J01230)

摘　　要：根据4190ZLC-2船用四冲程增压柴油机实际实验测得数据,建立基于广义回归神经网络(GRNN)的柴油机性能曲线和燃油消耗率预测模型。在所得实验数据中,选取柴油机油门、转速、扭矩等参数数值作为网络输入,柴油机的燃油消耗率作为网络输出。仿真结果表明:基于GRNN模型的神经网络学习速度快,预测精度高,可以很好地适用于柴油机燃油消耗率的性能预测中,并且能很好的实现预测仿真的效果。模型建立之后,可以根据测得数据实时了解柴油机的运行工况及性能状态。According to the actual experimental data of 4190ZLC-2 four-stroke turbochatged marine diesel engine, the performance curve and the prediction model of fuel consumption rate are established by using General Regression Neural Network （GRNN） algorithm. In the obtained experimental data, the diesel engine throttle, speed, torque and other parameters are selected as the network inputs, while the diesel engine fuel consumption rate as the outputs. The simulation results show that the neural network based on GRNN model has the advantages of learning quickly and high prediction accurac？A, which is well applied in the prediction of the diesel engine fuel oil consumption rate. It also well meets the needs of the prediction simulation. After the establishment of the model, the diesel engine operating conditions and performance status can be known by real-time based on the measured data.

关 键 词：船用柴油机 GRNN神经网络 性能曲线模拟 油耗率预测 

分 类 号：U664.21[交通运输工程—船舶及航道工程]