Proposal Title

Aircraft High Speed PM Synchronous Machine Basic Sizing Regression Neural Function

Primary Faculty Mentor’s Name

Dr Adel el Shahat

Proposal Track

Student

Session Format

Poster

Abstract

The requirement for electrical power on board aircraft is forecasted to rise dramatically in the future as a result of increased loads for improved in-flight information services, and the advent of more new electrical loads such as electrical actuation for flight surfaces and landing gear. This paper proposes basic sizing of high speed permanent magnet synchronous machine (HSPMSM), which utilizes the windmill effect of the low-pressure turbine of the aircraft engine for emergency power generation. Because this backup power would be immediately available following a main generator failure and this would potentially have a number of benefits. This paper presents relations among basic sizing parameters like, the required output power, tip speed, stack length, rotor diameter, stress, and r.p.m speed. This is done within the most probable range of power (5:500 Kw) in order to suit its applications. All results are presented with the aid of MATLAB environment. Permanent magnet synchronous machine operates at high speed to be convenient with its function in aircraft, moreover, high speed PM generators provide a substantial reduction in size and weight, also higher in power density, with simple structure. Also, this type of machine could be used in spacecraft especially in flywheel. This paper introduces neural network unit, using the back propagation (BP) learning algorithm due to its benefits. This unit generates the basic sizing function for the initial sizing parameters, with suitable number of network layers and neurons at minimum error and precise manner. This neural model takes the desired output power, with the required tip speed as inputs, and outs rpm speed, stack length, diameter. This modeling is done with a prescribed number of pole pairs, and stress value. The neural model has the ability to predict values in – between learning values, also make interpolation between learning curves data at various tip speed values as shown in the paper. Finally, deduction of algebraic nonlinear function which, connects between inputs and outputs for neural network is presented, to can aid any designer without the need of training the neural network each time. The validity of this algebraic function is achieved from comparisons between their values with such coming from Matlab relations, and Neural Network results.

Keywords

Design, High speed, PM synchronous machine, Neural network, Aircraft, MATLAB

Award Consideration

1

Location

Concourse/Atrium

Presentation Year

2014

Start Date

11-15-2014 2:55 PM

End Date

11-15-2014 4:10 PM

Publication Type and Release Option

Presentation (Open Access)

This document is currently not available here.

Share

COinS
 
Nov 15th, 2:55 PM Nov 15th, 4:10 PM

Aircraft High Speed PM Synchronous Machine Basic Sizing Regression Neural Function

Concourse/Atrium

The requirement for electrical power on board aircraft is forecasted to rise dramatically in the future as a result of increased loads for improved in-flight information services, and the advent of more new electrical loads such as electrical actuation for flight surfaces and landing gear. This paper proposes basic sizing of high speed permanent magnet synchronous machine (HSPMSM), which utilizes the windmill effect of the low-pressure turbine of the aircraft engine for emergency power generation. Because this backup power would be immediately available following a main generator failure and this would potentially have a number of benefits. This paper presents relations among basic sizing parameters like, the required output power, tip speed, stack length, rotor diameter, stress, and r.p.m speed. This is done within the most probable range of power (5:500 Kw) in order to suit its applications. All results are presented with the aid of MATLAB environment. Permanent magnet synchronous machine operates at high speed to be convenient with its function in aircraft, moreover, high speed PM generators provide a substantial reduction in size and weight, also higher in power density, with simple structure. Also, this type of machine could be used in spacecraft especially in flywheel. This paper introduces neural network unit, using the back propagation (BP) learning algorithm due to its benefits. This unit generates the basic sizing function for the initial sizing parameters, with suitable number of network layers and neurons at minimum error and precise manner. This neural model takes the desired output power, with the required tip speed as inputs, and outs rpm speed, stack length, diameter. This modeling is done with a prescribed number of pole pairs, and stress value. The neural model has the ability to predict values in – between learning values, also make interpolation between learning curves data at various tip speed values as shown in the paper. Finally, deduction of algebraic nonlinear function which, connects between inputs and outputs for neural network is presented, to can aid any designer without the need of training the neural network each time. The validity of this algebraic function is achieved from comparisons between their values with such coming from Matlab relations, and Neural Network results.