基于MEA-BP神经网络的钻井机械钻速预测

Prediction of Drilling ROP Based on MEA-BP Neural Network

机械钻速是衡量钻井效率的重要指标,如何高效地预测出机械钻速对提高钻井效率、降低钻井成本具有重要意义。目前机械钻速预测模型多以物理实验和经验公式为主,缺乏对现场实际工程数据的应用,基于此,该文提出了一种基于思维进化算法(MEA)优化BP神经网络的机械钻速预测新模型。该模型以现场实际工程数据为基础,通过小波降噪、标准化处理和灰色关联度分析对数据进行预处理,利用思维进化算法实现BP神经网络的初始权值和阈值的优化,从而实现机械钻速预测。将该模型预测结果与单一BP神经网络、遗传算法(GA)优化后的预测结果对比,结果表明,经过思维进化算法优化后的机械钻速精度更高,拟合优度达到0.936,为机械钻速预测带来一种新思路。

ROP is an important indicator to measure drilling efficiency.How to efficiently predict ROP is of great significance to improve drilling efficiency and reduce drilling costs.However,the current ROP prediction models are mostly based on physical experiments and empirical formulas,and lack the application of actual engineering data in the field.Based on this,a new model of ROP prediction based on the evolutionary thinking algorithm(MEA)optimization of BP neural network is proposed.The model is based on the actual engineering data in the field,preprocesses the data through wavelet noise reduction,standardization processing and grey correlation analysis,and uses the evolutionary algorithm to realize the optimization of the initial weights and thresholds of the BP neural network,so as to realize the ROP predict.Comparing the prediction results of this model with the prediction results optimized by a single BP neural network and genetic algorithm(GA),the results show that the ROP optimized by the thinking evolution algorithm has higher precision,and the goodness of fit reaches 0.936,it brings a new idea for ROP prediction.