摘要
针对电涡流缓速器制动力矩数学模型在高速时的计算力矩与实际输出力矩存在较大偏差的缺点,提出了一种基于数学模型和神经网络模型相结合的混合制动力矩模型,在电涡流缓速器低速时采用数学模型计算力矩,在高速时采用神经网络模型来逼近非线性输出力矩.分析了电涡流缓速器的输出制动力矩在低、高速时的特性,提出了数学模型与神经网络模型切换点的选择方法.通过实验对比了基于数学模型和基于混合模型的制动力矩曲线的逼近效果,结果表明混合模型更为有效.
The mathematical model of braking torque of eddy current retarder was analyzed. Aimed at the mathematical model＇s shortage of braking torque inherent bias decreasing to practical value at high velocity, a novel hybrid braking torque model composed of mathematical model and neural network was proposed. The hybrid model took advantage of the good characteristics of mathematical model at low velocity and adopted neural network to approach the nonlinear output torque of eddy current retarder at high velocity. The characteristics of eddy current retarder were analyzed at low and high velocity respectively. A selecting method for switch point between mathematical model and neural network was proposed. The capabilities of the mathematical model and the hybrid model were compared, with the experimental results showing that the hybrid braking torque model was more accurate.
出处
《浙江大学学报（工学版）》
EI
CAS
CSCD
北大核心
2006年第11期1980-1984,共5页
Journal of Zhejiang University：Engineering Science
基金
国家自然科学基金资助项目(60374057)
浙江省科技攻关计划重点资助项目(021110514)
关键词
电涡流缓速器
混合模型
制动力矩
神经网络
eddy current retarder
hybrid model
braking torque
neural network
