基于GRNN算法的船用柴油机性能曲线模拟与油耗率预测

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。

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

A mathematical model of 4190ZLC-2 four-stroke turbocharged marine diesel engine performance prediction is established by using General Regression Neural Network (GRNN) algorithm classification methods, according to the actual measured data of the 4190 series marine diesel engine. Taking into account factors that affect the running state, the diesel engine throttle, speed, torque and other parameters from the obtained data are selected as the input, and the diesel engine fuel oil consumption rate as the output. The simulation results showed that based on General Regression Neural Network classification model has the advantages of learning quickly and high prediction accuracy. And it can be applied in the diesel engine fuel oil consumption rate prediction very well, better meet the needs of the diesel engine performance prediction simulation. After the establishment of the model, the engineer in real-time grasp the diesel engine operating conditions and performance status based on measured data.