Erosion void condition prediction models for buried linear assets

Sewer pipelines are a major component of any infrastructure. In fact, they are subject to deterioration through their service lives. Although researchers are pointing out the consequences of erosion voids in sewer pipelines, the literature lacks the evaluation and prediction of such a defect. This research's main goal is to propose a model that can predict the condition of the erosion voids present outside sewer pipelines considering the fuzzy expert system. The methodology relies on different factors that are expected to contribute to the voids' development and severity. Fuzzy membership functions are constructed for five identified factors. The aggregated index representing the condition of the studied defect is achieved by considering the relative weights of the factors collected from 32 experts. The model is implemented on two case studies. The first case study supplied average values of accuracy, precision, and true positive rate (TPR) of 83%, 80%, and 76%, respectively. The second case is used to associate the soil intrusion defect with the erosion voids condition. Accordingly, it concludes that higher soil intrusion percentages are observed in poor and critical erosion voids conditions. In addition, a regression analysis model is developed to study the relation between the structural grade of sewer pipelines and erosion voids conditions. The regression model produced a coefficient of determination (R2) of 70.0%, minimal errors, and a high average validity percentage (AVP). This research shall help decision-makers in studying erosion voids defect to avoid major rehabilitation costs and sinkholes.