SOME EXPERIMENTS ON MULTIPLE REFLECTION CANCELLATION

Abstract

One of the most promising methods of identification or cancellation of multiple reflections on seismic records involves the calculation of synthetic records with all primaries and multiples, and the matching of the synthetic record with the field record. Such matching suffers today from the lack of precise information about the velocities and densities of the formations, dips of beds nonvertical transmission, etc. One possibility of improving this match involves the use of the earth itself as the “synthetic record computer.” In this process, the upcoming (or downgoing) primary signals are fed back into the earth with a vibrator in proper amplitude and phase to create a synthetic record of multiples only, which should match the multiples on the field record. Of course, only those multiple reflections which include a downward reflection from beds above the primary signal detectors will be included in the synthetic record of multiples only. The paper reports two experimental programs. One was carried out on an analog network to simulate the near‐surface and deeper formations, with means to feed back the upcoming signals in proper timing and polarity to cancel the multiples. These experiments indicated the theoretical workability of the process. The second program of experiments involved the use of a vertical spread to detect the upcoming and downgoing signals, and the use of a hydraulic vibrator to impress those signals back into the earth. These experiments were not conclusive because of insufficient power in the vibrator and high noise level. However, they indicated possible ways in which these limitations might be reduced, and the method applied to routine field operations.