Development of neural network calibration algorithms for multi-port pressure probes /
algorithm and program
| Main Author: | |
|---|---|
| Format: | Thesis eBook |
| Language: | English |
| Published: |
[Place of publication not identified] :
[publisher not identified] ;
1996.
|
| Subjects: | |
| Online Access: | Link to OAKTrust copy |
| Summary: | algorithm and program application of neural network calibration to this new design architecture optimization. Techniques for local minima architecture, and network optimization capabilities. avoidance and convergence rate improvement, incorporated into batch mode processing. The study compares the performance of calibrate multi-port pressure probes for use in wind tunnel calibrating multi-port probes for flow measurement by commercial packages' in terms of maximum allowable network contributes to the current techniques and methods for demonstrate PROBENET's real world application. demonstrates the flexibility and utility of the method. To dimensional velocity data about a delta wing model to flow analysis. The code offers distinct advantages over for the calibration of velocity measurement instrumentation function per layer and multiple activation per layer, and implementing a robust neural network and training algorithm. In addition, a unique mini-5-hole probe having a diameter of in terms of convergence rate and accuracy. The, research is intended to be a robust self learning code, developed to is presented. The backpropagation-based program, PROBENET, layer, as well as heuristics-based procedures for network of traditional 5- and 7-hole probes is detailed and the only 0.065" is calibrated and used to measure three probe is used to demonstrate the accuracy of threecomponent probe to overcome the flow angularity measurement limitations PROBENET incorporates multiple activation functions per prove the capabilities of PROBENET, a 5-hole hemispherical shows that the latter consistently produces a better solution size, training convergence rates, flexibility in network the algorithm, include: momentum, variable learning rate, and The development of a novel nearly-onmi-directional 18-hole The development of an enhanced neural network training two types of network architectures: single activation velocity prediction over a large range of flow angularity. |
|---|---|
| Item Description: | "Major subject: Aerospace Engineering". Vita. |
| Physical Description: | x, 108 leaves : illustrations ; 28 cm. Also available online. Issued also on microfiche from Lange Micrographics. |
| Bibliography: | Includes bibliographical references: pages 61-62. |