DocumentCode :
724191
Title :
The research of loading model of eddy current dynamometer based on DRNN with double hidden layers
Author :
Fan Shuang ; Wang Junzheng ; Shen Wei ; Chen Guangrong
Author_Institution :
Beijing Inst. of Technol., Beijing, China
fYear :
2015
fDate :
23-25 May 2015
Firstpage :
2485
Lastpage :
2490
Abstract :
Since the loading model of eddy current dynamometer is difficult to establish as well as the experimental data may be insufficient in the process of modeling, this paper proposed a method, based on diagonal recurrent neural network (DRNN) with double hidden layers, to predict the data which can reflect system characteristics but can´t be measured by experiments, and then establish the loading model of eddy current dynamometer. Comparing the performance of DRNN with that of recursive least square (RLS) with forgetting factor method, this proposed model is much closer to the practical input and output characteristics of eddy current dynamometer. Appling it to control the loading system as a reference is conducive to the improvement of control precision and the enhancement of response characteristic.
Keywords :
dynamometers; eddy currents; engines; least mean squares methods; mechanical engineering computing; recurrent neural nets; DRNN; RLS; control precision; diagonal recurrent neural network; double hidden layers; eddy current dynamometer; forgetting factor method; loading model; recursive least square; response characteristic; Data models; Eddy currents; Load modeling; Loading; Mathematical model; Predictive models; Torque; DRNN; Dynamometer; Loading model;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Control and Decision Conference (CCDC), 2015 27th Chinese
Conference_Location :
Qingdao
Print_ISBN :
978-1-4799-7016-2
Type :
conf
DOI :
10.1109/CCDC.2015.7162339
Filename :
7162339
Link To Document :
بازگشت