Dynamic Inverse Model for Multi-Sensor Measurement System Based on Wavelet Neural Network

2012 ◽  
Vol 499 ◽  
pp. 335-339
Author(s):  
Dong Zhi Zhang ◽  
Bo Kai Xia ◽  
Kai Wang ◽  
Jun Tong ◽  
Nian Zhen Yang
Keyword(s):  
Neural Network ◽  
Measuring Method ◽  
Wavelet Neural Network ◽  
Inverse Model ◽  
Measuring System ◽  
Water Mixture ◽  
Accuracy Measurement ◽  
Sensing Technology ◽  
Dynamic Inverse ◽  
Digital Oilfield

As traditional measuring method based on dielectric coefficients shows cross-sensitivity for multi-factor in the measurement of oil/water mixture, it can not meet the requirements of digital oilfield construction. Therefore, this paper presents an inverse model of wavelet neural network (WNN) combining with multi-sensing technology for achieving high-accuracy measurement of water content in crude oil. The simulation and experimental results demonstrate that the proposed method is available to eliminate the cross-coupling effects of multi-factors. The method has higher measurement accuracy and stronger generalization than the model built by BP-NN, and opens a versatile approach in nonlinear error calibration for multi-factors measuring system.

Wavelet Neural Network Based Adaptive Robust Control for a Class of MIMO Nonlinear Systems

2011 ◽  
Vol 383-390 ◽  
pp. 290-296
Author(s):  
Yong Hong Zhu ◽  
Wen Zhong Gao
Keyword(s):  
Neural Network ◽  
Robust Control ◽  
Nonlinear Systems ◽  
Wavelet Neural Network ◽  
Adaptive Robust Control ◽  
Loop Control ◽  
Control Approach ◽  
Mimo Nonlinear Systems ◽  
Output Tracking Control ◽  
Dynamic Inverse

Wavelet neural network based adaptive robust output tracking control approach is proposed for a class of MIMO nonlinear systems with unknown nonlinearities in this paper. A wavelet network is constructed as an alternative to a neural network to approximate unknown nonlinearities of the classes of systems. The proposed WNN adaptive law is used to compensate the dynamic inverse errors of the classes of systems. The robust control law is designed to attenuate the effects of approximate errors and external disturbances. It is proved that the controller proposed can guarantee that all the signals in the closed-loop control system are uniformly ultimately bounded (UUB) in the sense of Lyapunov. In the end, a simulation example is presented to illustrate the effectiveness and the applicability of the suggested method.

A Local Linear Wavelet Neural Network Based on a Bayesian Design Method

2009 ◽  
Vol 129 (7) ◽  
pp. 1356-1362
Author(s):  
Kunikazu Kobayashi ◽  
Masanao Obayashi ◽  
Takashi Kuremoto
Keyword(s):  
Neural Network ◽  
Design Method ◽  
Wavelet Neural Network ◽  
Bayesian Design ◽  
Local Linear
Sign in / Sign up

Export Citation Format

Share Document