A no-cost time extension was requested, to permit additional laboratory testing prior to undertaking field data collection. This was received in this reporting period. To minimize program cost, this additional testing is planned to be performed in concert with EPRI-funded testing at the Coal Flow Test Facility. Since the EPRI schedule was undecided, a hiatus occurred in the test effort. Instead, a significant effort was exerted to analyze the available laboratory test data to see whether the source and nature of noise behaviors could be identified, or whether the key flow information could be extracted even in the presence of the noise. One analysis approach involved filtering the data numerically to reject dynamics outside of various frequency bands. By varying the center frequency and width of the band, the effect of signal frequency on flow dynamics could be examined. Essentially equivalent results were obtained for all frequency bands that excluded a neighborhood of the transducer resonance, indicating that there is little advantage to be gained by limiting the experimental frequency window. Another approach examined the variation of the dynamics over a series of 1-second windows of data, producing an improvement in the prediction of coal flow rate. Yet another approach …

Additional calibration data were collected in the Coal Flow Test Facility early in this reporting period. These data comprised a total of 181 tests for stud and magnetic accelerometer mounts, with two mounting locations relative to two different pipe elbows, and including some tests with out-of-plane elbows upstream of the test section to produce coal ''roping''. The results found in analyzing these new data were somewhat disappointing: correlations for coal flow rate for a given mount type and mounting location were less accurate than desired, and degraded badly when data from other locations were included in the same analysis. Reviewing all of the data files (from both the earlier testing and recent calibration testing) disclosed a significant fraction of cases with several forms of noise. Eliminating these cases improved the correlations somewhat, but the number of cases that remained did not permit general conclusions to be drawn. It was finally learned that yet another type of noise is present in some data files, producing a strong effect on the correlation accuracy. The cases not subject to this noise correlated very well. It would be desirable to collect additional data in the Coal Flow Test Facility prior to moving on to …