基于ISCA-BP算法的天然气水合物预测模型

Prediction model of natural gas hydrate based on ISCA-BP network

为了建立更加准确、稳定的天然气水合物预测模型,利用改进正余弦算法优化神经网络,在正余弦算法(SCA)位置变化中加入非线性权重,对个体位置进行修正,提高算法收敛精度;融入Levy飞行改进SCA算法,强化局部搜索能力;将改进的正余弦算法用于BP神经网络参数优化,搭建ISCA-BP天然气水合物预测模型。与传统的热力学模型和BP模型模拟结果进行了对比,并且将该模型应用到气田现场测试。结果表明,ISCA-BP模型预测结果绝对相对误差为1.990%,平均绝对相对误差仅为0.339%,与其他热力学模型和BP模型相比,误差最小,预测结果精度高,稳定性好。在酸性体系和含醇盐体系中,都具有更准确的表现。ISCA-BP模型现场应用效果良好,可为抑制剂注入量的确定和现场安全运行策略的制定提供理论依据。

In order to establish a more accurate and stable natural gas hydrate prediction model,the improved sine and cosine algorithm was used to optimize the neural network,and the nonlinear weight was added to the position change of the sine and cosine algorithm(SCA) to correct the individual position and improve the convergence accuracy of the algorithm.Levy flight was integrated to improve SCA algorithm and strengthen local search ability.The improved sine and cosine algorithm was used for parameter optimization of BP neural network,and an ISCA-BP gas hydrate prediction model were built.The simulation results of traditional thermodynamic model and BP model were compared,and the model was applied to the field test of gas field.The result shows that the absolute relative error of ISCA-BP model prediction results is 1.990%,and the average absolute relative error is only 0.339%.Compared with other thermodynamic models and BP models,the error is the smallest,and the prediction result has high precision and good stability,more accurate performance in both acidic and alkoxide-containing systems.The field application of the ISCA-BP model has a good effect,which can provide a theoretical basis for the determination of the inhibitor injection amount and the formulation of the field safety operation strategy.