next up previous contents

Subsections


16 Journal Bearing

Given the eccentricity $\epsilon$ of the journal bearing, find the pressure distribution in the lubricant separating the shaft from the bearing.

Formulation

The journal bearing problem [10] requires determining the pressure between two circular cylinders of length $L$ and radii $R$ and $ R + c $. The separation between the cylinders is $ \varepsilon c $, where $ \varepsilon $ is the eccentricity. The pressure $v$ minimizes the quadratic $q:K \mapsto \mbox{${\mathbb{R}}$}$,

\begin{displaymath}
q(v)= \int_{\cal D}
\left\{ \frac{1}{2} w_q(x) \Vert\nabla v(x)\Vert^2 - w_l(x) v(x) \right\} dx ,
\end{displaymath}

over the convex set $K$, where ${\cal D}=(0,2 \pi) \times (0,2 b)$,

\begin{displaymath}
K=\{v \in H_0^1 ({\cal D}): v \ge 0 \} ,
\end{displaymath}

$ H_0^1 ({{\cal D}}) $ is the space of functions with gradients in $ L^2 ( {\cal D} ) $ that vanish on the boundary of $ \cal {D} $, and the functions $w_q:{\cal D} \mapsto \mbox{${\mathbb{R}}$}$ and $w_l:{\cal D} \mapsto \mbox{${\mathbb{R}}$}$ are defined by

\begin{displaymath}
w_q (\xi_1,\xi_2)=(1+\varepsilon \cos \xi_1)^3, \qquad
\ w_l(\xi_1,\xi_2) = \varepsilon \sin \xi_1 ,
\end{displaymath}

with $ \varepsilon \in (0,1) $ the eccentricity of the bearing.

A finite element approximation to the journal bearing problem is obtained by triangulating ${{\cal D}}$ and minimizing $q$ over the space of piecewise linear functions with values $ v_{i,j} $ at the vertices of the triangulation. We follow [3] by using a triangulation with, respectively, $ n_x $ and $ n_y $ internal grid points in the coordinate directions. Data for this problem appears in Table 16.1.



Table 16.1: Journal bearing problem data
Variables $ n_x n_y$
Constraints 0
Bounds $ n_x n_y $
Linear equality constraints 0
Linear inequality constraints 0
Nonlinear equality constraints 0
Nonlinear inequality constraints 0
Nonzeros in $ \nabla ^2 f(x) $ $5n_x n_y -2(n_x+n_y)$
Nonzeros in $ c'(x) $ 0

Performance

We provide results with the AMPL formulation in Table 16.2 with $b=10$ and $\epsilon = 0.1$. For these results we fix $ n_x = 50 $ and vary $ n_y $. The starting guess is the function $\max\{\sin (x),0\}$ evaluated at the grid nodes. Figure 16.1 shows the pressure distribution for the journal bearing problem.



Table 16.2: Performance on pressure in journal bearing problem
Solver $n_y = 25$ $n_y = 50$ $n_y = 75$ $n_y = 100$
LANCELOT 3.02 s 7.19 s 11.77 s 17.89 s
$f$ -1.54015e-01 -1.54824e-01 -1.54984e-01 -1.55042e-01
$c$ violation 0.00000e+00 0.00000e+00 0.00000e+00 0.00000e+00
iterations 12 11 10 10
LOQO 3.36 s 5.71 s 9.56 s 13.33 s
$f$ -1.54015e-01 -1.54824e-01 -1.54984e-01 -1.55042e-01
$c$ violation 2.2e-16 3.1e-16 3.7e-16 4.2e-16
iterations 26 19 20 21
MINOS 173.65 s 964.59 s 2850.41 s $\ddagger$
$f$ -1.54015e-01 -1.54824e-01 -1.54984e-01 $\ddagger$
$c$ violation 0.0e+00 0.0e+00 0.0e+00 $\ddagger$
iterations 1 1 1 $\ddagger$
SNOPT 722.68 s $\ddagger$ $\ddagger$ $\ddagger$
$f$ -1.54015e-01 $\ddagger$ $\ddagger$ $\ddagger$
$c$ violation 0.0e+00 $\ddagger$ $\ddagger$ $\ddagger$
iterations 197 $\ddagger$ $\ddagger$ $\ddagger$
$\dagger$ Errors or warnings. $\ddagger$ Timed out.

Figure 16.1: Journal bearing problem with $ \epsilon = 0.1 $
\includegraphics[width=5in]{ps/bearing.eps}


next up previous contents
Liz Dolan
2001-01-02