Institute of Metal research, Chinese Academy of Science, Wencui Rd 62, Shenyang, 110016, China.
Sci Rep. 2017 Jan 17;7:40827. doi: 10.1038/srep40827.
A concept for prediction of organic coatings, based on the alternating hydrostatic pressure (AHP) accelerated tests, has been presented. An AHP accelerated test with different pressure values has been employed to evaluate coating degradation. And a back-propagation artificial neural network (BP-ANN) has been established to predict the service property and the service lifetime of coatings. The pressure value (P), immersion time (t) and service property (impedance modulus |Z|) are utilized as the parameters of the network. The average accuracies of the predicted service property and immersion time by the established network are 98.6% and 84.8%, respectively. The combination of accelerated test and prediction method by BP-ANN is promising to evaluate and predict coating property used in deep sea.
提出了一种基于交变静压（AHP）加速试验预测有机涂层的概念。采用不同压力值的 AHP 加速试验来评估涂层降解。建立了反向传播人工神经网络（BP-ANN）来预测涂层的使用性能和使用寿命。将压力值（P）、浸泡时间（t）和使用性能（阻抗模|Z|）用作网络参数。所建立网络预测使用性能和浸泡时间的平均准确率分别为 98.6%和 84.8%。加速试验和 BP-ANN 预测方法的结合有望用于评估和预测深海用涂层的性能。
