In order to validate this data procedure and estimation method, we carried out experiment on the ROMAC stability test rig as depicted in Fig. ef{fig2:subfig:a}. The fluid and magnetic bearing configurations were in cite{cloud2007stability}, and during the excitation process, the temperature of inlet oil was 37 ${}^{circ}mathrm{C}$. The shaft rotates at 3500 rpm, and the actuator #2 that is located at the node 47 in the FE model excited this test rig at six frequencies (55 Hz to 80 Hz by step of 5 Hz). The excitation force is expressed in Eq.( ef{eq32}). egin{equation}label{eq32}
m{F}{
m{ = }}left( {egin{array}{*{20}{c}}
{{F_{ex}}}\
{{F_{ey}}} end{array}} ight) = {F_e}left( {egin{array}{*{20}{c}}
{Xcosleft( …show more content…
egin{table}[htbp] …show more content…
ef{fig15}. The first forward log decrement remain nearly invariant in the excitation frequency range, while the first backward frequency drops a little.
egin{figure}[htbp] centering includegraphics[width=8cm]{Fig15.pdf} caption{Mode paramenters} label{fig15} end{figure} Using the sine-swept excitation method, we did both backward and forward excitaions for the test rig, and Fig. ef{fig16:subfig} shows the full spectrum waterfall for the left bearing (Node 10). Using the idenitification method [Li], the first mode parameters are estimated (1B: f=85.3Hz, LogDec=0.726; 1F: 86.3Hz, LogDec=0.442). By comparing the prdicted and measured first forward mode parameters, we can drow conclusion that the force coeffients of bearings are reasonably