OCR Text |
Show near 15 MHz and the other was operated near 45 MHz. The output signal from each mixer was low-pass filtered to pass only the difference frequency between the raw signal and the local oscillator. The signals for the horizontal (15 MHz) and vertical (45 MHz) channels were thus separated and the signal frequencies were significantly lowered, allowing more accurate signal processing. The processors measure the average period of the signals by counting the number of cycles of a 100 MHz clock during some fixed number (usually 8) of signal cycles. For each measurement location the LV provided 1300 to 1400 simultaneous measurements of instantaneous axial and vertical velocities, u and v. An error analysis determined that the uncertainty in the instantaneous velocities generally ranged from 2% to 4% of the measured values(4). Statistical methods were used to obtain mean velocities, TT and "v", turbulence velocities, u' and v', and the cross correlation u'v'. In general statistical tolerances for the mean and turbulence velocities establish, with 95% probability, that true values are within about 5% of calculated values for each of the ~400 locations per test case. Typical tolerances are presented in Table II. TYPICAL LASER VELOCIMETER MEASUREMENT CONFIDENCE TOLERANCES AXIAL u (tn/s) -.741 -.511 -.068 .070 .226 .318 .492 .637 A«(%) tolerance ±4.0 2.2 25.0 17.1 7.5 •1.7 4.1 .17 u' (m/s) .546 .216 .314 .224 .317 .281 .368 .210 tolerance ±3.8 12.5 3.2 4.0 3.8 3.6 2.4 3.8 VERTICAL V (m/s) -.575 -.050 -.039 .203 .305 .746 1.367 1.614 Av(%) tolerance ±3.7 38.0 43.6 16.7 4.6 3.2 1.8 1.4 v' (m/s) .398 .364 .318 .618 .260 .451 .439 .544 Av'{%) tolerance ±3.8 0.3 3.1 2.8 5.4 6.2 4.3 2.6 fbx = 0.44 Mhz 6X = 25.73 /im -1.76 m/s < u < 1.53 m/s fhy = 1.47 Mhz f)y = 8.46 /urn -4.47 m/s < v < 4.47 m/s 8 |