Drilling loss circulation prediction method based on Hippopotamus algorithm to optimize hyperparameters

During the development of oil fields, drilling loss circulation may lead to serious con-sequences such as stuck drill, blowout, and extended non-productive time. To address this key technical problem, this paper uses machine learning to construct a combination of 16 different classification models and optimization algorithms to accurately predict the occurrence of drilling loss circulation. First, this paper verifies the effectiveness of the algorithm optimization problem on the benchmark function and evaluates the optimization performance of the four intelligent optimization algorithms. The results show that the Hippopotamus Optimization has the highest convergence speed and optimization accuracy on the four benchmark functions. Subsequently, in order to improve the model performance and generalization ability, the above optimization algorithms are used to optimize the hyperparameters of the four classic classification algorithms to determine the best configuration of each combination and construct a drilling loss circulation prediction model. The results show that the support vector machine model optimized by the Hippopotamus Optimization algorithm achieves an accuracy of 98.72% and an F1 score of 0.983, which is significantly better than the other 15 comparison models, and the prediction results are highly accurate and consistent. The research results can provide a theoretical basis for the prediction of drilling loss circulation.