基于IAGA-BP神经网络的海洋平台健康状况预测

Health Status Prediction of Offshore Platforms Based on IAGA-BP Neural Network

对海洋平台的健康状况进行客观合理的评估是保障海洋平台正常安全运作的重要前提。针对海洋平台日益突出的健康状况评估问题,提出了基于改进的遗传算法和BP神经网络相结合的海洋平台健康状况预测模型。该模型首先对传统的遗传算法进行改进,保留精英个体,大概率选择优秀个体,直接淘汰较差个体,通过自适应算法改进交叉概率和变异概率,再通过改进的遗传算法对BP神经网络的初始权值和阈值进行优化,以优化了初始权值和阈值的BP神经网络对海洋平台健康状况进行评估。通过对海洋平台的历史监测数据进行了仿真验证,结果表明,传统BP神经网络预测误差大多在0.4~0.8之间浮动,GA-BP神经网络预测误差大多在0.2~0.4之间浮动,而IAGA-BP神经网络预测误差大多在0.1~0.2之间浮动,IAGA-BP神经网络模型的收敛速度和精度明显优于传统的BP神经网络模型和GA-BP神经网络模型,可以更好的适用于对海洋平台健康状况的预测。

An objective and reasonable assessment of the health status of offshore platforms is an important prerequisite for ensuring the normal and safe operation of offshore platforms.To solve the increasingly prominent health assessment problem of offshore platforms,an offshore platform health prediction model based on a combination of improved genetic algorithm andBPneural network is proposed.First,the model improves the traditional genetic algorithm,which retains the elite individuals,selects the outstanding individuals with a high probability,directly eliminates the poor individuals,regulates the crossover probability and the mutation probability through an adaptive algorithm,and then the improved genetic algorithm is used to optimize the initial weights and thresholds of the BP neural network,finally,BP neural network with optimized initial weights and thresholds is used to evaluate the health status of offshore platforms.Through the simulation verification of the historical monitoring data of the offshore platform,the results show that the traditional BP neural network prediction error mostly varies between 0.4 and 0.8,the GA-BP neural network prediction error mostly varies between 0.2 and 0.4,while the IAGA-BP neural network prediction error mostly varies between 0.1 and 0.2,and the convergence speed and accuracy of IAGA-BP neural network model are obviously better than the traditional BP neural network model and GA-BP neural network model,which is betterfor predicting the health of offshore platforms.