Formation drillability prediction based on PSO-SVM

By analyzing the relation between mud logging data, well logging data and formation drillability, a novel method for predicting formation drillability based on particle swarm optimization and support vector machine (PSO-SVM) is proposed. The prediction model for formation drillability is established using the data of drilling pressure, rotary speed, hydraulic horsepower, bottom hole differential pressure, rate of penetration, bit diameter, acoustic velocity and formation depth by training the SVM, which is optimized by PSO algorithm. After that, the proposed algorithm is applied to Zhang 105 well in Junggar Basin to predict the formation drillability with the model established using the data of Zhuang 2 well, Zhuang 102 well, Zhuang 103 well and Zhuang 104 well. The experimental results show that the PSO-SVM algorithm has higher prediction precision, faster convergence speed and better generalization effect than BP neural network based approach.