An extension approach to estimate soil corrosivity for buried pipelines

Highlights

• An extension theory based approach was developed to evaluate soil corrosivity.

• The matter element models were established to combine corrosivity levels, soil properties, and property values together.

• The contradiction problems among soil properties were solved through the proposed approach.

• The corrosivity level were quantitatively determined using eigenvalues of grade variables.

Abstract

Underground oil and gas pipelines are usually expected to be long-lasting. Corrosion of these steel pipelines may cause structural failures that significantly threaten life and cause environmental hazards. Therefore, developing a reliable approach to estimate soil corrosivity is important for designing a targeted anti-corrosion structure and performing risk assessment. In this study, an extension-based approach is proposed to evaluate soil corrosivity based on the following seven soil properties: redox potential, soil resistivity, pH, pipe-to-soil potential, water content, Chloride (Cl⁻) concentration, and salt content. Specifically, the soil was classified into five corrosivity levels, and the classic domain element, joint domain element, and element to be evaluated were established by the matter-element theory. Then, the corrosivity level was determined based on the maximum correlation degree of the multi-index to the five levels, and the final classification was obtained from the eigenvalues of the grade variables. Finally, the case study was examined to validate the application of the approach, and the results were compared to the method of buried metal specimens, which was used as a criterion. The present approach, which provided a more detailed classification, was demonstrated to be a superior choice for classifying soil corrosivity levels.

Introduction

Several factors may directly or indirectly contribute to the structural failure of buried oil and gas pipelines, such as third-party damage, corrosion, improper maintenance, construction defect/material failure, etc.; however, the most important is corrosion [1,2]. Corrosion deterioration in the surface of buried pipelines is mainly due to the direct contact of pipelines with an aggressive soil environment. As the service life of buried pipelines increases, perforation and rupture failures are increasingly significant, which may threaten life and properties. To clarify the external corrosive environment of buried pipelines and guide targeted anti-corrosion measures, assessing soil corrosivity in a feasible manner is critical.

The corrosion of metal pipelines in soil has been of increasing concern in the last decades. The study of soil corrosivity has a long and substantial history, and several methods have been developed to predict the corrosive environment impacting underground pipelines. In general, the weight loss of buried metal specimens method is deemed as the most classical, objective, and accurate method. The weight loss is converted into an annual corrosion rate to estimate the soil corrosivity, and is often used as a criterion to measure the reliability of other new and existing methods. However, it cannot be used for prediction because the metal specimens must be buried in the soil for an extended period. This method is significantly labor-intensive and time-consuming [[3], [4], [5], [6]]. Previous studies have also revealed that the physical and chemical properties of soils significantly affect the corrosion propagation of buried metals [[7], [8], [9], [10], [11]]. Based on several factors considered to evaluate soil corrosivity, the single index method and the multi-factors comprehensive method were presented. The single index method (e.g., soil corrosivity, redox potential, pH, water content, and salt content) has long been attempted by several countries [3,4,12]. The National Association of Corrosion Engineers (NACE) and the American Society for Testing and Materials (ASTM) suggest that soil resistivity determines the degradation process of buried pipelines [[13], [14], [15]]. Generally, soils with lower resistivity appear to be more corrosive. However, soil corrosivity is also related to water and salt contents; soils with higher water and salt contents appear to be more corrosive. A simple factor of soil resistivity often leads to misjudgment under certain field conditions [5,16]. In fact, no single soil property can determine soil corrosivity, and the interaction of several factors must be considered [12]. Therefore, comprehensive approaches that consider multi-factors are regarded as more reliable and realistic. Several scoring methods that include a significant number of factors are introduced to predict soil corrosivity. The Cast Iron Pipe Research Association, predecessor to the current organization, the Ductile Iron Pipe Association in 1964, recommends a 10-point scoring (10-P) method based on five soil properties (soil resistivity, pH, redox potential, sulfate concentration, and moisture content) to assess the soil corrosivity for cast iron pipes [[16], [17], [18], [19]]. The currently used scores are defined by the American Water Works Association (AWWA) [20,21]. Metalogic N.V. (Heverlee, Belgium<www.metalogic.be>) also proposed a scoring method involving 12 factors (e.g., soil resistivity, water content, pH, buffering capacity, sulfide, chloride) for evaluating soil corrosivity [17]. In addition, several other scoring methods that include a couple factors are also available [16,22,23]. The following four aspects are considered as the major drawbacks of a scoring method: 1) the nonlinear relationship among the factors cannot be processed for the additivity of the factor scores, 2) it neglects the fact that the contribution of different factors to soil corrosivity may significantly vary, and is essentially a weighted-average method, 3) it is costly and often impractical to collect all the data required, 4) it considers fixed factors regardless of availability.

Data mining techniques are considered to be more flexible and apply comprehensive scientific approaches to manage multi-factors for the assessment of soil corrosivity, such as the fuzzy based method, clustering based method, and artificial neural network [6,9,11,[15], [16], [17], [18], [19]]. However, these approaches are limited by the following three factors: 1) the species and quantities of soil properties are fixed and cannot be adjusted according to field conditions, 2) the contradictory problems presented by the interaction of multi-factors and the inconsistent effects on soil corrosion cannot be sufficiently addressed, 3) the classification of soil corrosivity is not highly specific.

Considering a new method for managing incompatible problems, the extension theory has been increasingly applied in the engineering field [[24], [25], [26], [27]]. The extension theory, which applies the matter-element theory and extension set theory as its theoretical framework, is a new mathematical tool to solve contradictory problems. A matter element that combines the object, object features, and feature value ranges, is the logical cell of the extension theory. The contradictory problem among the object features can be transformed into a compatible problem by establishing a matter element model. The effect of the object features on an object can be quantitatively determined by a correlation function calculation; its essence belongs to the principle of near selection of multi features. Therefore, the extension-based approach proposed in this study is a superior choice for assessing soil corrosivity [[24], [25], [26], [27]]. It is preferred for the following four reasons: 1) the species and quantities of soil properties are not restricted and can adjust according to specific conditions, 2) the contradictory problems among soil properties can be sufficiently addressed, 3) it can calculate the weight vector objectively, and 4) it can provide a more detailed classification of soil corrosivity through the eigenvalues of grade variables.

In this study, we aim to develop an extension-based approach for the assessment of soil corrosivity. The corrosivity level, soil properties, and property values are combined by establishing matter element models. The effect of soil properties on soil corrosivity is quantitatively presented by the correlation functions of the extension set theory. The weight vector is calculated by the entropy weight method. The final classification is determined by the eigenvalues of the grade variables. Finally, the approach is applied to the soil survey data. It is proved to be applicable and accurate for soil corrosivity classification.