The study of Statistics is widely discussed and applied in various areas of activity, both academically and in practice. Historically, the behavior of materials in geotechnics has been addressed in a deterministic manner, which can result in considerable errors due to uncertainties associated with the measurement of geotechnical parameters, the execution of tests, and the natural distribution of soils/tailings. The application of statistical analyses in mining geotechnics has seen increased recognition and importance in recent years; however, it still reflects a constrained approach, specially concerning the practical use of the conducted studies (Ferreira, 2022).
In the mining sector, soil studies are conducted using deterministic tests at specific points within the structure, which may result in erroneous estimates for unsampled areas, as the materials are often heterogeneous and subjected to varying levels of stress, moisture, particle size distribution, and other conditions. While efforts are consistently made to obtain a significant distribution of samples across the structure, there is growing interest in statistically analyzing the distribution and correlation of the physical and geomechanical geotechnical parameters that characterize the behavior of piles and dams, to enhance the verification and assurance of construction safety and compliance.
Authors in the field discuss different applications of statistics in geotechnics in the mining area, highlighting the importance of performed analyses and the continuation of these studies. Azevedo (2019) addresses the study of outliers to identify discrepant readings in piezometers within dam monitoring. In a case study, the author compares Statistical Process Control methodologies based on a normal distribution with the Box Plot, applied to the dataset, and concludes that the latter is the more suitable method for the given objective. The author highlights that the application of the statistical solution in the interpretation of piezometric readings is configured as a tool for control and reliability of the measured values.
Badaró (2022) discusses the use of statistical analyses to compare the performance of different methods of obtaining moisture for the quality assurance and quality control of filtered tailings disposal structures. Moisture results obtained from the pan drying method and the infrared scale are compared, establishing a linear relationship between both and the oven method. As per the author, the data generated during technological control is, by necessity, outcomes related to a particular control parameter. Therefore, organizing, describing, interpreting, and analyzing the set of data is a duty of statistics, which represents an important tool for decision-making.
Ferreira (2022) discusses the comparison of deterministic and probabilistic analyses for determining the safety factor in dam stability studies. The author emphasizes that, unlike manufactured materials used in civil construction, soil lacks exact parameters and is unlikely to be accurately represented by a single value. The purpose of probabilistic analyses is to account for this variability in the evaluations. When comparing the results of the analyses, it was observed that the probabilistic analysis yielded more representative data than the deterministic analysis, better accounting for the existing uncertainties.
Ferreira (2022) points out that one of the primary obstacles in utilizing statistics in geotechnics within the mining sector is the large volume of data that must be analyzed. The absence of centralized data in an intelligent environment further complicates the execution of automated statistical analyses, which directly affects the practical application of these results. Given the acknowledged importance of statistical applications in geotechnics within the literature, a statistical module is currently under development in Geolabor. Its purpose is to enable the automation of analyses tailored to user contexts, thereby aiding critical decision-making during construction projects.
Among the various possibilities of analysis that will be implemented in the Geolabor statistical module, the following stand out: Statistical visualization of the geotechnical parameters obtained from the conducted tests; Identification of points that, statistically, would not comply with the technical specifications established in the project, even when, deterministically, they comply with the established criteria; Implementation of regressions, allowing the interpretation of the variation of a geotechnical parameter in relation to others and of correlations between the analyzed geotechnical parameters; Correlation between geotechnical parameters of the QA/QC with the parameters obtained in the geotechnical investigations; and the prediction of geotechnical parameters at unsampled locations within the structure, framed within a confidence interval.
Among the various analyses to be implemented, the following are particularly noteworthy for their impact on geotechnics: The possibility of comparing estimated test results with those obtained, allowing the quality of the results to be assessed; Understanding the spatial distribution of the geotechnical parameters of the structure, along its entire length; Performing analyses of the correlations between the geotechnical parameters and the construction process, allowing a better understanding of the behavior of the structure; Obtaining test results taking into account the uncertainties of laboratory measurements, ensuring the quality of the tests; Understanding the probability of success of the QA/QC process for different contexts.
The forthcoming implementations in Geolabor will focus on notifying users of possible areas requiring attention and proposing more in-depth critical analyses in particular locations or tests during the construction phase. Statistical studies do not replace test results or the need for ongoing testing during construction; instead, they serve as tools to support critical decision-making, which is the primary function of Geolabor as a statistical tool in geotechnics.
Written by: Bianca Lacerda
References
AZEVEDO, André Gonçalves de. Análise estatística para identificação de leituras discrepantes de piezometria: estudo de caso em barragem de terra. 2019). Available at: < monografias.ufop.br/ > . Accessed on: October 04, 2023
BADARÓ, João Victor Oliveira. Análise estatística como ferramenta de tomada de decisões: estudo de caso na determinação da umidade no controle tecnológico de uma estrutura de disposição de rejeito filtrado. 2022. Available at: < monografias.ufop.br/ > . Accessed on: October 04, 2023
FERREIRA, Paula Marinho. Avaliação da confiabilidade do fator de segurança obtido por análise probabilística em relação à análise determinística em barragens de mineração. 2022. Available at: < monografias.ufop.br/ > . Accessed on: October 04, 2023
