Prediction of Cement Slurry Formulation Based on RMAPCN

Cement slurry formulation design is a very important process in cementing operations, and the performance of the formula will directly affect the quality of cementing. In traditional cementing operations, professionals perform experiments based on information from adjacent Wells to obtain fuzzy cement formulas, and then constantly adjust the final cement formula in the actual operation. This operation can cost a lot of labor and time, which is not good for real-time operation. To solve the above problems, a new forecasting model RMAPCN (IndRNN with Multi-Head Attention Mechanism and PSO-CNN) is proposed in this paper for the prediction of cement slurry formulation. The model can quickly obtain the cement slurry formulation with higher accuracy and better timeliness by mining the characteristics of the existing data information and using the well information. The comparison between the RMAPCN model and the baseline model shows that the RMAPCN model has obvious advantages in each performance index, and can be used to guide the actual cement slurry formulation design.