基于深度学习的晶间腐蚀试验温度自适应控制系统设计

Study on Adaptive Temperature Control During Intergranular Corrosion Test of Corrosion Resistant Alloys Based on Deep Learning

针对耐蚀合金晶间腐蚀试验中试验装置和温度控制存在的问题,应用深度学习技术实现了试验温度的自适应控制。首先,通过图像处理技术获得了腐蚀溶液表面沸腾运动特征图,形成了训练数据集;随后搭建了卷积神经网络(CNN)模型并利用搭建的数据集进行了训练,基于CNN模型和modbus TCP协议开发了协同温度控制算法;最后形成了原型试验系统并对其进行了测试。研究结果表明,基于数据集训练后的CNN模型能够以95%以上的准确率对溶液沸腾状态进行识别。在晶间腐蚀验证试验中,开发的智能控制算法能够根据溶液沸腾状态输出正确的温控指令,实现随沸腾状态变化的自适应温度控制。

The deep learning computer vision technology is applied on adaptive temperature control during intergranular corrosion test of corrosion-resistant alloys.Firstly,the surface boiling motion feature images of the test solution is obtained through image processing to form a training dataset.Then,a convolutional neural network(CNN)is built and trained with the training dataset,and a collaborative temperature control algorithm is developed based on the CNN model and modbus TCP protocol.Finally,an experimental system based on the developed temperature control algorithm was built and tested.Results show that the trained CNN model can accurately recognize the boiling state of the solution with the accuracy of more than 95%.During the intergranular corrosion test,the developed adaptive temperature control algorithm can give correct control instructions according to the boiling state of the solution.The adaptive temperature control with the change of boiling state during the test is achieved.