Long-Term Prediction of Atmospheric Corrosion Loss in Various Field Environments

CORROSION ◽  
10.5006/2706 ◽  
2018 ◽  
Vol 74 (6) ◽  
pp. 669-682 ◽  
Author(s):  
Yi-kun Cai ◽  
Yu Zhao ◽  
Xiao-bing Ma ◽  
Kun Zhou ◽  
Hao Wang
Keyword(s):  
Neural Network ◽  
Linear Function ◽  
Environmental Variables ◽  
Atmospheric Corrosion ◽  
Accurate Estimation ◽  
Corrosion Loss ◽  
Surface Method ◽  
Long Term Prediction ◽  
Corrosion Data

This paper deals with the prediction of long-term atmospheric corrosion in different field environments using the power-linear function. A method for the calculation of exponent n and stationary corrosion rate α in the power-linear function is proposed based on the 1- and 8-y corrosion loss results (C1 and C8) of the ISO CORRAG program. The response surface method and the artificial neural network methodology are used to obtain the accurate estimation of C1 and C8 in different locations using environmental variables. Considering the uncertainty of the model and the experimental data, the confidence intervals of n and α are also calculated. It is shown that the long-term predictions obtained by the proposed method coincide with the actual corrosion loss within ±30% relative error. The estimations for the range of the long-term corrosion loss are also reliable. The proposed method is helpful in extrapolating the knowledge of corrosion management to different field environments where corrosion data are not available.

Long-term prediction on atmospheric corrosion data series of carbon steel in China based on NGBM(1,1) model and genetic algorithm

2019 ◽  
Vol 66 (4) ◽  
pp. 403-411 ◽  
Author(s):  
Yuanjie Zhi ◽  
Dongmei Fu ◽  
Tao Yang ◽  
Dawei Zhang ◽  
Xiaogang Li ◽  
...  
Keyword(s):  
Genetic Algorithm ◽  
Carbon Steel ◽  
Atmospheric Corrosion ◽  
Data Series ◽  
Data Sets ◽  
Content Type ◽  
Term Prediction ◽  
Long Term Prediction ◽  
Corrosion Data

PurposeThis study aims to achieve long-term prediction on a specific monotonic data series of atmospheric corrosion rate vs time.Design/methodology/approachThis paper presents a new method, used to the collected corrosion data of carbon steel provided by the China Gateway to Corrosion and Protection, that combines non-linear gray Bernoulli model (NGBM(1,1) with genetic algorithm to attain the purpose of this study.FindingsResults of the experiments showed that the present study’s method is more accurate than other algorithms. In particular, the mean absolute percentage error (MAPE) and the root mean square error (RMSE) of the proposed method in data sets are 9.15 per cent and 1.23 µm/a, respectively. Furthermore, this study illustrates that model parameter can be used to evaluate the similarity of curve tendency between two carbon steel data sets.Originality/valueCorrosion data are part of a typical small-sample data set, and these also belong to a gray system because corrosion has a clear outcome and an uncertainly occurrence mechanism. In this work, a new gray forecast model was proposed to achieve the goal of long-term prediction of carbon steel in China.

A spiking neural network-based long-term prediction system for biogas production

2020 ◽  
Vol 129 ◽  
pp. 271-279 ◽  
Author(s):  
Giacomo Capizzi ◽  
Grazia Lo Sciuto ◽  
Christian Napoli ◽  
Marcin Woźniak ◽  
Gianluca Susi
Keyword(s):  
Neural Network ◽  
Biogas Production ◽  
Spiking Neural Network ◽  
Prediction System ◽  
Term Prediction ◽  
Long Term Prediction

Long-Term Prediction of a Sine Function Using a LSTM Neural Network

2016 ◽  
pp. 159-173 ◽  
Author(s):  
Magdiel Jiménez-Guarneros ◽  
Pilar Gómez-Gil ◽  
Rigoberto Fonseca-Delgado ◽  
Manuel Ramírez-Cortés ◽  
Vicente Alarcón-Aquino
Keyword(s):  
Neural Network ◽  
Sine Function ◽  
Term Prediction ◽  
Long Term Prediction

Rock Mass Deformation Long-Term Forecast via DRNN Neural Network Model

2012 ◽  
Vol 433-440 ◽  
pp. 770-774
Author(s):  
Sha Ma ◽  
Zhi Quan Huang
Keyword(s):  
Neural Network ◽  
Rock Mass ◽  
Network Stability ◽  
Term Forecast ◽  
Link Weight ◽  
Network Output ◽  
Rock Mass Deformation ◽  
Long Term Prediction ◽  
Mass Deformation

The question of rock mass deformation Long-term forecast is researched base on DRNN. The construction of neural network is optimized via reconstructed chaotic phase space, and the all nodes of DRNN are interconnected, and the feedback between nodes and that of node itself is included, and mult-linkage branch is build between two nerves, and the linkage branch stands for the link weight and the time delay of regular step. So the current moment network output of node depends on not only current moment network iutput, but also the node output of Some moment before current, so the chaotic prediction sensibility to initial condition is reduced effectively. The calculating velocity and network stability is improved effectively. Examples show that results are reasonable and the long-term prediction is reasonable and feasible.

Sign in / Sign up

Export Citation Format

Share Document