The atmospheric corrosion of metallic materials causes great economic loss every year worldwide. Thus, it is meaningful to predict the corrosion loss in different field environments. Generally, the corrosion prediction method includes three parts of work: the modelling of the corrosive environment, the calibration of the corrosion effects, and the establishment of the corrosion kinetics. This paper gives an overview of the existing methods as well as promising tools and technologies which can be used in corrosion prediction. The basic corrosion kinetic model is the power function model and it is accurate for short-term corrosion process. As for the long-term corrosion process, the general linear models are more appropriate as they consider the protective effect of the corrosion products. Most corrosion effect models correlate the environmental variables, which are characterized by the annual average value in most cases, with corrosion parameters by linear equations which is known as the dose-response function. Apart from these conventional methods, some mathematical and numerical methods are also appropriate for corrosion prediction. The corrosive environment can be described by statistical distributions, time-varying functions and even geographic information system (GIS), while the corrosion effect can be captured via response surface models and statistical learning methods.

In this work, the corrosion behavior and mechanical properties of 30CrMnSiA high-strength steel under a harsh marine atmosphere environment were systematically studied using accelerated test technology, along with corrosion kinetic analysis, microstructure and phase composition analysis, electrochemical measurements, and mechanical property tests. The influence of corrosion time on corrosion kinetics was characterized by the weight loss method. The corrosion layer and its product evolution were analyzed by SEM, EDS, XRD, and XPS. The corrosion behavior of steel was studied by a potentiodynamic polarization curve and EIS. Finally, the influence of corrosion on mechanical properties was studied by tensile and fatigue tests. The results show that 30CrMnSiA high strength steel has good corrosion resistance in a harsh marine atmosphere environment. Its corrosion behavior is cyclical: the outer rust layer exfoliated, the inner rust layer became the outer rust layer, and the matrix became inner rust due to the attack by the corrosive medium. The rust layer had a great protective effect on the matrix. The mechanical properties of 30CrMnSiA high-strength steel were reduced under the corrosive environment, and corrosion had a significant effect on its fatigue resistance.

Predicting the corrosion loss over time in different environments is challenging because the corrosion process develops over time and is influenced by multiple environmental factors simultaneously. Conventional regression analysis is not applicable because it has several limitations in dealing with multilevel structured data. In this paper, the hierarchical linear modelling method is employed instead. A two-level linear growth model is built to analyse the individual corrosion growth and the corrosion effects of the environment on carbon steel, zinc, and copper. Environmental factors that have important impact on the corrosion process are distinguished. By including the corrosion time and environmental factors as predictor variables, the model predictions are in good agreement with the experiments. The uncertainty of the corrosion data is quantitatively described and the confidence intervals are obtained. Then the long-term corrosion loss is predicted with power-linear corrosion kinetic model and corrosion-induced damage risks are analysed.

Corrosion in natural seawater is difficult to simulate in a laboratory due to the slow rate and complexity of the corrosion process which involves multiple influential factors. This paper aims to explore the quantitative effect of environmental factors on corrosion process and find the best experimental conditions which represent the actual environment and have the best acceleration effect. A new framework is followed in this paper which consists of three parts: design of experiments, outdoor and laboratory corrosion tests, and corrosion mechanism consistency confirmation. A L₆(3¹ × 2²) orthogonal experiment is designed in laboratory to study the effect of temperature, salinity, and dissolved oxygen on marine corrosion behavior of B10 copper alloy. In each test, H₂O₂ is added in seawater to accelerate the corrosion process. Outdoor exposure tests are also conducted in natural seawater. Results show that the corrosion process in laboratory and outdoor follows the same mechanism, in view of corrosion product and morphology, corrosion kinetics, as well as mechanical properties. With the help of quantitative analysis of the test results, a better acceleration condition can be designed.

There are obvious differences in the corrosion process of materials in different climatic regions, and it is of great significance to establish the corrosion process model considering the effects of environmental factors to improve environmental adaptability. However, many existing models do not take comprehensive environmental factors into account, and the selection of environmental factors in these corrosion models is not sufficiently investigated. Meanwhile, the comparison of models is also based on goodness-of-fit. In order to improve the atmospheric corrosion prediction model and use an improved method to judge the suitable model instead of traditional goodness-of-fit, this paper first studies the effects of various atmospheric environment factors on corrosion loss. Then a corrosion prediction model considering the influence of various atmospheric environment factors is proposed, and the fitting effect is compared with the previous model with experimental data. Considering that the best fitting cannot guarantee the best prediction, a new criterion of model selection is also proposed for corrosion model confirmation. Corrosion prediction results reveal that the proposed model is more effective than the existing models. Finally, the influence of the environmental factors on the model output is discussed by sensitivity analysis.