Case Study: Sinopec achieves high-accuracy seismic inversion with MathWorks MATLAB and deep learning

Sinopec Develops High Accuracy Intelligent Seismic Inversion with Deep Learning

Sinopec, one of China’s largest oil and gas companies, needed a higher-precision approach to seismic inversion for petroleum exploration. Traditional model-based methods were sensitive to noise and often produced low-resolution results, so Sinopec turned to MathWorks and MATLAB, using tools such as Signal Processing Toolbox, Optimization Toolbox, and Statistics and Machine Learning Toolbox to support geophysical modeling.

MathWorks helped Sinopec develop a new frequency-phase intelligent inversion method that combines seismic frequency-phase features with deep learning. Using MATLAB, the team extracted seismic wavelets, performed high-resolution time-frequency analysis, built large labeled datasets, and trained deep networks, resulting in improved high-resolution impedance inversion for both simulated and real seismic data and a marked boost in practical application.

Sinopec