Feature Memory and Deep Cross Network for Parameter Prediction of Offshore drilling

The Rate of Penetration (ROP) is a critical indicator of drilling efficiency in offshore drilling engineering. Accurately predicting ROP can significantly enhance drilling efficiency and reduce enterprise costs, making it a crucial topic in the oil industry. In this paper, we propose a deep learning method called the Feature Memory and Deep Cross Network (FMDCN). This model has the unique ability to remember important features and analyze time-series data, resulting in a more accurate prediction of ROP. Our experiments, conducted on real datasets, demonstrate that FMDCN consistently outperforms general time-series models, showing a smaller prediction error.