The intelligent reservoir classification and evaluation platform adopts deep learning to simulate the professional logic of geological experts, at the same time, it analyzes and models the logging curve and logging image data, and has reservoirs got intelligent division and classification through multi-model fusion technology.
Taking reservoir evaluation as the starting point and the intelligent processing and interpreting of the logging information as the premise , it solves the two major problems of massive data in multiple wells of the old area and too much individual analysis proportion in the previous conclusions. It improves the work efficiency of technicians effectively and reduces the multiplicity of solutions.
Penetrating the intelligent technology into oil and gas exploration and development, and it has become an indispensable means to solve the basic links of mature exploration areas, deepen reservoir knowledge, serve rolling exploration and efficient development.
Well related data files could be read, including logging curves, imaging logging data and manual evaluation results (for comparison).
Including missing value processing, gray processing, etc
To realize the logging data and reservoir classification results visualized, and it mainly includes the original logging curve, reservoir classification and evaluation results, manual analysis results, and the consistency between the model intelligent recognition and manual analysis.
New well Planning Evaluation
Rechecking the old well to extract
the potential oil reservoir
One click to intelligent divide and identify the reservoir
The use of random forest interpretation model can refine the reservoir discrimination logic and quantitative indicators, which is convenient for users to understand.
The reservoir is divided by comprehensively using imaging logging data and logging curve data, and the final division result is obtained after weighting by using random forest algorithm and convolution neural network algorithm.
The platform can complete the processing, division and comprehensive research and judgment of multi-source data of single target well in one minute, and quickly locate all kinds of high-quality reservoirs