Upload
digvijay-maheshwari
View
231
Download
0
Embed Size (px)
Citation preview
7/28/2019 Nonlinear Feedback linearisation
1/17
A feedback linearizationcontrol the nonlinear 5-Do
flywheel suspended by PMbiased HMB
Authors: Wen Tong, Fang Jiancheng
Presen
Arpit Ch
Digvija
7/28/2019 Nonlinear Feedback linearisation
2/17
Contents Introduction
7/28/2019 Nonlinear Feedback linearisation
3/17
Intro AMBs(Active magnetic bearings) instead of normal ball-bea
Advantage: High Rotation speed
No contact character
Nearly zero friction
Long life span
No auxiliary lubricant system
Types:
EMB(Electro-magnetic bearing)
PM(permanent magnet) biased HMB(Hybrid magnetic bearing)
7/28/2019 Nonlinear Feedback linearisation
4/17
EMB vs PM biased HMB
In EMB, suspending forces for rotor are generated completecurrents.
Whereas in HMBs PM provide the biased flux and current gthe regulating force to realize the stable suspension of the r
Power consumption is reduced greatly in the application of
biased HMB, because the biased flux is proved by the permamagnet without any external power consumption as opposebiased flux is generated by biased current.
7/28/2019 Nonlinear Feedback linearisation
5/17
Need of controller
Effective controller is required for Magnetic suspensionflywheel(MSFW) because of the the instability of the open-lAMB caused by the negative displacement stiffness of themagnetic force .
The dynamic model of the PM biased HMB is nonlinear andcomplicated, and its nonlinearity is caused by the superposi
PM flux and EM flux and the magnetic flux coupling in differdirections.
It is simplified by assuming that permanent magnet in HMB equivalent to biased current in the EMB, so an equivalent EMmodel describing the HMB system can be obtained
7/28/2019 Nonlinear Feedback linearisation
6/17
Need of exact linearization
Based on the equivalent model, using the Taylor series expa
the neighborhood of the operating point(or equilibrium poiapproximate linear model of the nonlinear HMB system canobtained for the controller design.
This method has an advantage that the control model is verand the controller can be realized simply.
For the application of the HMB with high restriction on powand accuracy(such as MSFW), the controller design must bethe nonlinear model of system.
7/28/2019 Nonlinear Feedback linearisation
7/17
Sketch of MSFW
Diagram 1:
7/28/2019 Nonlinear Feedback linearisation
8/17
Model development For convenience, there are three coordinate frames introdu
system, as shown in Figure. OXYZ is the stator-fixed frame, o
rotor-fixed frame and oabc is the intermediate frame.
7/28/2019 Nonlinear Feedback linearisation
9/17
Five DOFs According to the design of flywheel(MSFW), The rotation alon
direction is controlled by the driven motor and independent of tThe translation z is controlled by the axial HMB and can be regardecoupled with other four coupled axial DOFs[x, y, , ].
The four axial DOFs of the rotor could be calculated from the foupicked up by the displacement sensors[xma, yma, xmb, ymb] as showEq.(1):
x = xma+ xmb
= (xma- xmb)/L
y = yma+ ymb
= (yma- ymb)/L
L is the distance between the distance between the two displacesensors located at the two ends the stator of the MSFW, as showfollowing figure
7/28/2019 Nonlinear Feedback linearisation
10/17
7/28/2019 Nonlinear Feedback linearisation
11/17
MSFW description
The MSFW includes four sets of radial HMBs.
As shown in following figures, each set consist of two seriesconnected coils placed diagonally opposite to generate restforces.
7/28/2019 Nonlinear Feedback linearisation
12/17
rotor
7/28/2019 Nonlinear Feedback linearisation
13/17
Take the HMB in X direction for example. It is assumed that the control cur
sent to the opposite connected coils in X+ and X-. Control currents flowing
two coils in X+ and X- are iax and iax, respectively. fax is the resultant restor
acting on the rotor, which is the difference between the two magnetic forc
in X+ and X- directions.
fax = fax+ - fax-= [(pax++ iax++ iax-ax+)2 - (pax- - iax- - iax+ax-)
2 ]/ (20A)
Where,
pax+, pax- are PM flux in air-gap of X+ and X-,
iax+
, iax+ax-
are EM flux in X+ and X- generated by the control current iax
,
iax-, iax-ax+ are EM flux in X+ and X- generated by the control current -iax.
permeability of the vacuum and A is the area of one pole. The other three
forces fay, fbx and fby would be obtained using the same methods described
7/28/2019 Nonlinear Feedback linearisation
14/17
Rotor Dynamics
ma and mb are distances from HMBs to the origin of the OXYZ.
fx, fy, px, and py are the disturbances caused by the residual static and dynamic
imbalances.
The state variables are chosen as [x y ]. The generalized magnetic forces actin
on the system are [fx py fy px].
7/28/2019 Nonlinear Feedback linearisation
15/17
Dynamic model of the nonlinear sys the MSFW is a 4-input, 4-output, 8-order nonlinear system.
The dynamics model is consisted of three parts. The first is sequation f(x) which is regarded as the uncontrollable permamagnetic force and the couple caused by the gyroscopic efferotor. The second is the controllable magnetic force gi (x)uithe input part of the system. The inputs of the system are seu=[iax ibx iay iby]. The last one is the disturbances caused by thdynamic imbalances.
7/28/2019 Nonlinear Feedback linearisation
16/17
where m is themassoftheroto
the moments ofinertialofthe
directions.
7/28/2019 Nonlinear Feedback linearisation
17/17
The input functions of the nonlinear system are described in detail as follows:
The output functions for the selected outputs: