基于多层前馈神经网络的涡轮导叶过渡态温度场预估方法

• Format: Patent
• Language: Chinese
• Published: 05.11.2024
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS

本发明公开了一种基于多层前馈神经网络的涡轮导叶过渡态温度场预估方法，首先，将在不同运行工况切换时涡轮导叶上关键点温度的过渡态实验数据离散化为多个状态点，并建立过渡态中各时刻的涡轮导叶温度场数据库，以此作为训练样本；然后，引入多层前馈神经网络对训练样本进行训练，通过多层全连接层、批量归一化层及激活层的组合，快速提取输入特征与对应输出之间的非线性映射关系，得到了不同运行工况变化时涡轮导叶过渡态温度场快速预测模型。本发明能够在工况变化时快速、准确的预测导叶上关键点温度的动态变化，提供了一种高效、可靠的涡轮导叶过渡态温度场预测工具。 The invention discloses a turbine guide vane transition state temperature field estimation method based on a multilayer feedforward neural network, and the method comprises the steps: discretizing the transition state experiment data of the temperature of a key point on a turbine guide vane into a plurality of state points when different operation conditions are switched, and building a turbine guide vane temperature field database at each moment in the transition state; taking the sample as a training sample; then, a multi-layer feedforward neural network is introduced to train a training sample, a nonlinear mapping relation between input features and corresponding output is rapidly extracted through combination of a multi-layer full connection la