Recently, the western part of Bohai has obtained significant commercial oil discoveries in Guantao group for the first time. The old seismic data is wide, but the bandwidth is narrow. The new data bandwidth is large, but the coverage is small. Besides, the impedance difference of sand and mudstone is small. The traditional impedance inversion can't distinguish them effectively. It's very important for the next evaluation to get seismic data that can reflect reservoir effectively.
We propose to use geophysical technology to fuse high-frequency and low frequency information of new data into old data. First, the reflection coefficient of the drilling bypass is extracted. Second combined it's high frequency information with the low frequency information of the reflection coefficient which is calculated by the drilled wells. After that, the fusion coefficient is used as the sample information to invert the frequency attributes of the old data. Finally, the high frequency, the low frequency information of frequency inversion and the intermediate frequency information of the old data are fused to obtain the broadband inversion data.
In the new and old seismic data superposition area, the frequency band characteristics of the comparison inversion data and the new data are found to be quite similar. Frequency band features are also very similar. It shows that the inversion data are true and reliable. Comparison between drilled wells and the inversion data, compared to the old seismic data, the inversion data of the Guantao formation reservoir thickness is more than 8m can be effectively reflected, and the old seismic data have seismic response when the reservoir thickness is more than 15m, and the response of some reservoirs is not obvious. The seismic resolution of the inversion data is improved effectively. To sum up, the frequency division attribute inversion based on seismic well fusion can effectively improve the bandwidth of the difference data in the rear area of seismic data, and improve the resolution of seismic data.