基于基因表达式编程算法的磨料射流切割深度预测模型

A model for predicting cutting depth of casing using AWJ technique based on gene expression programming algorithm

针对磨料射流切割套管深度与水力参数、工作参数及磨料参数之间存在复杂的耦合关系及传统预测方法的不足,建立基于基因表达式编程算法的磨料射流切割深度的预测模型。通过将函数表达式基因化,利用选择算子、变异算子、插串算子、变换算子等对群体实施遗传操作,得出最优函数表达式,并将其与人工神经网络预测模型、回归预测模型进行对比分析。结果表明,基于基因表达式编程算法的预测值与试验值的平均误差为3.93%,标准误差为0.251,预测精度明显高于其他预测模型,可直观、准确地反映磨料射流切割深度与水力参数、工作参数及磨料参数之间的关系,为磨料射流切割技术的定量控制提供可靠的理论支撑。

The cutting depth of casing using an abrasive water jet（ AWJ） technique is determined by the hydraulic power,the properties of the abrasives used and operation conditions. A prediction model for the cutting depth using the AWJ technique was established based on a gene expression programming algorithm. In the construction of the gene expression of functions,the operators of selection,mutation operator,plugging string,insertion and transform were used to perform the genetic operations in order to obtain an optimal function expression. A comparative analysis of the optimal function expression,an artificial neural network model and an regression correlation model was carried out. The results show that the average error between experimental results and the prediction using the gene expression algorithm is 3. 93%,and the standard error is0. 251,with accuracy significantly higher than that of other prediction models. The gene expression algorithm model proposed can be effectively used to support the quantitative control of casing cutting using the AWJ technique.