基于径向基函数神经网络的平台许用重心高度预测

为保障平台的安全生产,得到更精确的许用重心高度曲线,提出采用人工智能预测平台许用重心高度的方法以代替粗略估算的线性插值法。以某自升式风电安装平台的不同吃水作为神经网络输入层、相对应的完整稳性许用重心高度作为输出层进行训练,使用经过训练的神经网络预测许用重心高度曲线。对预测结果与实际值、线性插值法计算值进行比较,神经网络预测结果的最大相对误差为1.43%,神经网络法比线性插值法精度更高,证明其能够精确预测完整稳性许用重心高度曲线,可为平台的安全运营提供支持。

Prediction of Allowable Vertical Center of Gravity of Platform Based on Radial Basis Function Neural Network

In order to ensure the safe production of the platform and obtain more accurate allowable vertical center curve of gravity, an artificial intelligence method for predicting the allowable vertical center of gravity of the platform is proposed to replace the conventional linear interpolation method. Taken different drafts of a jack-up wind power installation platform as the input layer of the neural network and the corresponding intact stability allowable vertical center of gravity as the output layer for training, the trained neural network is used to predict the allowable vertical center curve of gravity. By comparing the predicted results with the actual values and the calculated values of linear interpolation method, the maximum relative error of the predicted results of the neural network is 1.43%, which shows that the neural network method is more accurate than the linear interpolation method. It proves that the neural network can accurately predict the allowable vertical center curve of gravity of the complete stability and provide support for the safe operation of the platform.