| The fine interpretation of time-frequency domain is based on high-resolution time-frequency analysis
algorithms. Reviewed the evolution of time-frequency analysis algorithms including S-Transform (ST), Generalized STransform
(GST), Synchrosqueezing S-Transform (SST), High-order Synchrosqueezing S-Transform (HSST),
Synchrosqueezing Generalized S-Transform (SGST), Multi-channel Synchrosqueezing Generalized S-Transform
(MSGST), it is claimed that MSGST has high vertical and horizontal resolution. To meet the requirements of highprecision
exploration and development, a High-order Multi-channel Synchrosqueezing Generalized S-Transform
(HMSGST) is proposed to focus on time-frequency resolution and improves the accuracy of reservoir prediction. By
comparing the calculation results of six time-frequency algorithms with theoretical signals and theoretical noise signals, it
is verified that HMSGST has high time-frequency resolution and better characterization ability for complex signals, but
when the signal-to-noise ratio of the signal is below 70 dB, the time-frequency noise resistance of HMSGST begins to
deteriorate, and the time-frequency spectrum of the effective signal is disturbed, making it difficult to accurately
characterize the time-frequency characteristics of the signal. Taking the marine sandstone in a certain 3D survey as an
example, based on HMSGST, time slices are extracted along the main oil and gas reservoirs (T2 layer) in the timefrequency
domain. Due to the corresponding relationship between frequency and geological information, high seismic
frequency is often more advantageous to the identification of thin reservoir. Attribute analysis such as frequency
attenuation gradient and frequency coherence are carried out to enhance the ability of hydrocarbon detection and
recognition of short axis river channels and secondary faults. Furthermore, RGB multi-attribute fusion including the
former and curvature attribute is carried out with fine carving for geological details, especially in the time slice of T2 layer
based on medium-high frequency coherence. For seismic data from different survey and geological condition, parameter
analysis of HMSGST is required, and suitable parameters are a prerequisite for achieving more accurate calculation results. |
