Artificial intelligence models to generate visualized bedrock level: a case study in Sweden
- PDF / 4,307,677 Bytes
- 20 Pages / 595.276 x 790.866 pts Page_size
- 32 Downloads / 190 Views
ORIGINAL ARTICLE
Artificial intelligence models to generate visualized bedrock level: a case study in Sweden Abbas Abbaszadeh Shahri1,3 · Stefan Larsson2 · Crister Renkel3 Received: 8 February 2020 / Accepted: 2 April 2020 © Springer Nature Switzerland AG 2020
Abstract Assessment of the spatial distribution of bedrock level (BL) as the lower boundary of soil layers is associated with many uncertainties. Increasing our knowledge about the spatial variability of BL through high resolution and more accurate predictive models is an important challenge for the design of safe and economical geostructures. In this paper, the efficiency and predictability of different artificial intelligence (AI)-based models in generating improved 3D spatial distributions of the BL for an area in Stockholm, Sweden, were explored. Multilayer percepterons, generalized feed-forward neural network (GFFN), radial based function, and support vector regression (SVR) were developed and compared to ordinary kriging geostatistical technique. Analysis of the improvement in progress using confusion matrixes showed that the GFFN and SVR provided closer results to realities. The ranking of performance accuracy using different statistical errors and precision–recall curves also demonstrated the superiority and robustness of the GFFN and SVR compared to the other models. The results indicated that in the absence of measured data the AI models are flexible and efficient tools in creating more accurate spatial 3D models. Analyses of confidence intervals and prediction intervals confirmed that the developed AI models can overcome the associated uncertainties and provide appropriate prediction at any point in the subsurface of the study area. Keywords Artificial intelligence · Bedrock level model · 3D spatial distribution · Predictability level · Sweden
Introduction Accurate prediction of spatial bedrock level (BL) can reveal complex patterns at different scales and can provide important information on designing and constructing civil engineering projects. However, observed BL data have inherent measurement errors and may contain several outliers. This implies that spatial modeling of BL is a difficult task that, due to associated uncertainties and non-uniformly distributed datasets (Baecher 1986), can increase the costs and the risks of a project (Clarke et al. 2009; Mey et al. 2015). Referring to Fig. 1, producing highly accurate models to predict the BL variation and its associated uncertainties is of
* Abbas Abbaszadeh Shahri [email protected] 1
Department of Civil Engineering, Islamic Azad University, Roudehen Branch, Tehran, Iran
2
Division of Soil and Rock Mechanics, KTH Royal Institute of Technology, Stockholm, Sweden
3
Johan Lundberg AB, Uppsala, Sweden
great importance to avoid inconvenience (Wang 2019; Yan et al. 2018; Manandhar et al. 2016; Abbaszadeh Shahri et al. 2011; Cerato and Lutenegger 2007). Collecting geotechnical data in large projects reduces the associated uncertainties with BL modeling. However, the applicability of s
Data Loading...
