unsteadiness on the rotor was 0.22%. This is exceptionally small relative to other cases run with the VBI code. For highly loaded transonic cases involving shocks, a difference between the maximum pressure and minimum pressure on the rotor of 30% is not uncommon. The very low Mach numbers in the Purdue rig presented difficulties for the VBI code because the flow solver in the code is optimized for compressible flow. The general rule is that the code is best suited for Mach numbers above 0.2. The inlet Mach number for the Purdue University low speed turbine rig was about 0.08, well below the normal limit Because of this, several of the standard input parameters used for transonic cases could not be used. For example, if the standard number of 4 sub-iterations per real time step was used, instabilities near the inlet developed and were visible as active spots of alternating high and low flow variable values. Increasing the number of sub-iterations per time step to 10 and reducing the CFL number per time step stabilized the calculation. These changes also increased the required run time by three times compared to transonic cases, making each case require at least three weeks run time on a Silicon Graphics Power Challenge machine.

0 0

Post a comment