基于PSO-BP的钻井机械钻速预测模型

机械钻速预测是优化钻进过程、提高钻井效率的关键技术,现有的计算模型主要建立在物理实验和理论分析的基础上,缺少对钻井工程实测数据的应用,导致计算精度难以满足复杂的现场需求.基于此,提出一种人工智能算法与BP(back propagation)神经网络相结合的钻井机械钻速预测模型.首先,利用小波滤波方法对实测数据进行降噪处理,并依据互信息关联分析优选输入参数降低模型冗余.其次,利用粒子群优化(particle swarm optimization,PSO)算法实现对BP神经网络初始权值、阈值的优化,建立机械钻速预测新模型,并将PSO-BP新模型与标准BP、BAS(Beetle Antennae Search,天牛须算法)-BP及GA(genetic algorithm,遗传算法)-BP等三种模型进行对比分析.最后,根据实际工况对PSO-BP钻井机械钻速预测模型进行模型评价.结果表明,PSO-BP机械钻速预测模型不仅具有良好的预测精度,而且为钻进过程中提高机械钻速提供科学的参考.

Rate of Penetration for Drilling Prediction Model Based on PSO-BP

The prediction of ROP is a key technology to optimize the drilling process and improve drilling efficiency. The existing calculation models are mainly based on physical experiments and theoretical analysis. The lack of application of measured data in drilling engineering makes the calculation accuracy difficult to meet the complexity On-site demand. Based on this, a new ROP prediction model combining artificial intelligence algorithm and BP neural network is proposed. First, the wavelet filtering method is used to reduce the noise of the measured data, and the input parameters are optimized according to the mutual information correlation analysis to reduce the model redundancy. Secondly, the PSO algorithm is used to optimize the initial weights and thresholds of the BP neural network, and establish a new model of ROP prediction. Finally, according to the actual data, the PSO-BP drilling speed prediction model is evaluated experimentally, and the new PSO-BP model is compared with the standard BP, BAS-BP and GA-BP models. The results show that the PSO-BP rate of penetration prediction model not only has a good prediction accuracy, but also provides a scientific reference for increasing the ROP during drilling.