This paper describes an intelligent control system that uses electro-optics and neural networks to control the flow of air over a flexible wall. In this investigation a pressure sensor which was part of the wall of the wind tunnel and an optical apparatus were used to produce moire´ fringes. A back propagation neural network was used to analyze the fringe patterns and to classify the pressures into four levels. A second neural network was used to recognize the pressure patterns and to provide the input to a control system that was capable of modifying the shape of the flexible wall in order to preserve the stability of the flow. The flexible wall was part of the wall of the wind tunnel and was installed in the upstream of the flow. It was made of silicone rubber and had an area of 76 mm by 76 mm. There were 15 rows of actuators installed under the flexible wall which were used to change the shape of the wall. In the downstream of the flow was an optical pressure sensor which had the same dimensions as the flexible wall and consisted of a 15 × 15 array of small diaphragms. These diaphragms responded to the pressure fluctuations in the boundary layer flow and were the source of the signals for the optical system. A CCD camera viewed the pressure sensor through an optical apparatus which produced moire´ fringes. A back propagation neural network analyzed the fringe patterns and classified the pressures into four levels. The classified pressures which was a 15 × 15 array of numbers ranging from 1 to 4 was the input to a second back propagation neural network which was used to recognize the pressure patterns. The output from the back propagation neural network used for pattern recognition provided the input to a control system that changed the shape of the flexible wall. This paper presents the experimental results as well as the computer simulations which were created for this project. This includes the complete process of creating the slope fringes, classifying the pressures into four levels, recognizing the wall pressure patterns and generating the output signals to the actuator for changing the shape of the flexible wall.
Skip Nav Destination
ASME 2005 International Mechanical Engineering Congress and Exposition
November 5–11, 2005
Orlando, Florida, USA
Conference Sponsors:
- Dynamic Systems and Control Division
ISBN:
0-7918-4216-9
PROCEEDINGS PAPER
Pressure Measurement and Pattern Recognition by Using Neural Networks
Parham Piroozan
Parham Piroozan
Indiana Institute of Technology
Search for other works by this author on:
Parham Piroozan
Indiana Institute of Technology
Paper No:
IMECE2005-79745, pp. 897-905; 9 pages
Published Online:
February 5, 2008
Citation
Piroozan, P. "Pressure Measurement and Pattern Recognition by Using Neural Networks." Proceedings of the ASME 2005 International Mechanical Engineering Congress and Exposition. Dynamic Systems and Control, Parts A and B. Orlando, Florida, USA. November 5–11, 2005. pp. 897-905. ASME. https://doi.org/10.1115/IMECE2005-79745
Download citation file:
11
Views
Related Proceedings Papers
Related Articles
Small Silicon Pressure Transducers for Space-Time Correlation Measurements in a Flat Plate Boundary Layer
J. Fluids Eng (September,1996)
A Three-Dimensional Human Trunk Model for the Analysis of Respiratory Mechanics
J Biomech Eng (January,2010)
A Method for the Construction of Flow Nozzles and Wind Tunnel
Contraction Sections
J. Fluids Eng (December,1977)
Related Chapters
Experimental Investigation of Ventilated Supercavitation Under Unsteady Conditions
Proceedings of the 10th International Symposium on Cavitation (CAV2018)
An Intelligent Control System Based on FPGA for Telephone Communication Network
International Conference on Electronics, Information and Communication Engineering (EICE 2012)
Engineering Design about Electro-Hydraulic Intelligent Control System of Multi Axle Vehicle Suspension
International Conference on Instrumentation, Measurement, Circuits and Systems (ICIMCS 2011)