基于贝叶斯优化BP神经网络的压力容器最小壁厚预测研究

Research on Minimum Wall Thickness Prediction of Pressure Vessel Based on Bayesian Optimization BP Neural Network

压力容器最小壁厚对压力容器的结构强度和稳定性具有重要作用。鉴于此,利用贝叶斯优化方法结合BP神经网络,实现对压力容器最小壁厚的精确预测。首先,采用贝叶斯方法优化隐含层神经元数量,得到最优隐含层神经元数量。然后,使用该最优神经元数进行模型训练。最后,将验证集输入到模型中进行验证,以得到压力容器最小壁厚的预测结果。结果表明,基于贝叶斯方法优化BP神经网络的模型能够精确地预测压力容器筒体和封头的最小壁厚,且均方误差仅为0.2678和0.7448,该模型能够提供更可靠的设计和决策依据,为提高压力容器结构设计的效率和安全性做出重要贡献。

The minimum wall thickness of pressure vessels plays an important role in the structural strength and stability of pressure vessels.In view of this,the Bayesian optimisation method combined with BP neural network is used to achieve accurate prediction of the minimum wall thickness of pressure vessels.First,the Bayesian method is used to optimise the number of neurons in the hidden layer to obtain the optimal number of neurons in the hidden layer.Then,the model is trained using this optimal number of neurons.Finally,the validation set is input into the model for validation to obtain the prediction of the minimum wall thickness of the pressure vessel.The results show that the model based on the Bayesian approach to optimise the BP neural network is able to accurately predict the minimum wall thickness of the pressure vessel cylinder and head with a mean square error of only 0.2678 and 0.7448,and the model is able to provide a more reliable design and decision-making basis,which can make an important contribution to the improvement of the efficiency and safety of the structural design of pressure vessels.