Joaquim Agostinho Barbosa Tinoco

Artificial intelligence applications in transportation geotechnics

This paper presents a brief overview of artificial intelligence applications in transportation geotechnics, highlighting new approaches and current research directions, including issues related to data mining interpretability and prediction capacities. Several practical applications to earthworks, including the compaction management and quality control aspects of embankments, as well as pavement evaluation, design and management, and the mechanical behaviour of jet grouting material, are presented to illustrate the advantages of using data mining, including artificial neural networks, support vector machines, and evolutionary computation techniques in this domain. This study also propose a novel simplified compaction table for reusing geomaterials and compaction management in embankments and applied one- and two-dimensional advanced sensitivity analyses to better interpret the proposed data-driven models for the prediction of the deformability modulus of jet grouting field samples. These applications show the capabilities of data mining models to address complex problems in transportation geotechnics involving highly nonlinear relationships of data and optimisation needs.

A data mining approach for Jet Grouting Uniaxial Compressive Strength prediction

Jet Grouting (JG) is a Geotechnical Engineering technique that is characterized by a great versatility, being the best solution for several soil treatment improvement problems. However, JG lacks design rules and quality control. As the result, the main JG works are planned from empirical rules that are often too conservative. The development of rational models to simulate the effect of the different parameters involved in the JG process is of primary importance in order to satisfy the binomial safety-economy that is required in any engineering project. In this work, three data mining models, i.e. Artificial Neural Networks (ANN), Support Vector Machines (SVM) and Functional Networks (FN), were adapted to predict the Uniaxial Compressive Strength (UCS) of JG laboratory formulations. A comparative study was held, by using a dataset used that was obtained from several studies previously accomplished in University of Minho. We show that the novel data-driven models are able to learn with high accuracy the complex relationships between the UCS of JG laboratory formulations and its contributing factors.

Using data mining techniques to predict deformability properties of jet grouting laboratory formulations over time

Jet Grouting (JG) technology is one of the most used softsoil improvements methods. When compared with other methods, JG is more versatile, since it can be applied to several soil types (ranging from coarse to fine-grained soils) and create elements with different geometric shapes (e.g. columns, panels). In geotechnical works where the serviceability limit state design criteria is required, deformability properties of the improved soil need to be quantified. However, due to the heterogeneity of the soils and the high number of variables involved in the JG process, such design is a very complex and hard task. Thus, in order to achieve a more rational design of JG technology, this paper proposes and compares three data mining techniques in order to estimate the different moduli that can be defined in an unconfined compressed test of JG Laboratory Formulations (JGLF). In particular, we analyze and discuss the predictive capabilities of Artificial Neural Networks, Support Vector Machines or Functional Networks. Furthermore, the key parameters in modulus estimation are identified by performing a 1-D sensitivity analysis procedure. We also analyze the effect of such variables in JGLF behavior.

Jet grouting column diameter prediction based on a data-driven approach

This study takes advantage of the high learning capabilities of data mining (DM) techniques towards to the development of a novel approach for jet grouting (JG) column diameter prediction. The high number of variables involved in JG technology as well as the complex phenomena related with the injection process make JG column diameter (D) prediction a difficult task. Therefore, in order to overcome it, the flexible learning capabilities of DM techniques were applied as an alternative approach of the traditional tools. The achieved results show that both artificial neural network and support vector machine algorithms can be trained to accurately predict D built in different soil types of clayey nature and using different JG systems. In both cases a coefficient of correlation () very close to the unity was achieved. For models training, a set of eight input variables were considered. Among them, the rod withdrawal speed, flow rate of the grout slurry and the JG system were identified as the most relevant ones, although the grout pressure and the dynamic impact of the grout also revealed an important influence on D prediction. Moreover, additionally to the identification of the key model variables, it was also measured their effects on D prediction based on a data-based sensitivity analysis. These achievements represent a novel contribution for JG technology, mainly at the design level. Furthermore, the obtained results also underline the potential and contribution of DM to solving complex problem in geotechnical engineering.

Jet grouting mechanicals properties prediction using data mining techniques

The authors wish to thank to Portuguese Foundation for Science and Technology (FCT) the support given through the doctoral grant SFRH/BD/45781/2008

Prediction of Soil-Cement Mechanical Properties Using EC2 Approach

For a better design of Jet Grouting (JG) and Cutter Soil Mixing (CSM) technologies, a set of laboratory formulations are usually prepared aiming to give a first idea of the mechanical behavior of the final mixture. However, these formulations can represent an important cost to the project. Therefore, aiming to reduce such cost, in the present work the analytical expressions proposed by Eurocode 2 for strength and stiffness prediction of concrete were adapted to soil-cement laboratory formulations for JG and CSM projects. It is shown that these expressions can be successful applied in mechanical properties prediction over time of soft soil stabilized with cement for a wide range of cement content, water cement ratios and soil types.

Application of Data Mining Techniques in the Estimation of Mechanical Properties of Jet Grouting Laboratory Formulations over Time

Sometimes, the soil foundation is inadequate for constructions purpose (soft-soils). In these cases there is need to improve its mechanical and physical properties. For this purpose, there are several geotechnical techniques where Jet Grouting (JG) is highlighted. In many geotechnical structures, advance design incorporates the ultimate limit state (ULS) and the serviceability limit state (SLS) design criteria, for which uniaxial compressive strength and deformability properties of the improved soils are needed. In this paper, three Data Mining models, i.e. Artificial Neural Networks (ANN), Support Vector Machines (SVM) and Functional Networks (FN), were used to estimate the tangent elastic Young modulus at 50% of the maximum stress applied (E tg50%) of JG laboratory formulations over time. A sensitivity analysis procedure was also applied in order to understand the influence of each parameter in E tg50% estimation. It is shown that the data driven model is able to learn the complex relationship between E tg50% and its contributing factors. The obtained results, namely the relative importance of each parameter, were compared with the predictive models of elastic Young modulus at very small strain (E 0) as well as the uniaxial compressive strength (Q u ). The obtained results can help to understand the behavior of soil-cement mixtures over time and reduce the costs with laboratory formulations.

Stability Condition Identification of Rock and Soil Cutting Slopes Based on Soft Computing

AbstractFor transportation infrastructure, one of the greatest challenges today is to keep large-scale transportation networks, such as railway networks, operational under all conditions. This task...

An application of Markov chains to predict the evolution of performance indicators based on pavement historical data

One of the main awareness for a road infrastructures manager is to increase its efficiency under limited resources. Pavement Management Systems aim, at last, to support road administrations in the decision-making process regarding its management policy and long-term strategies for maintenance and rehabilitation activities. While several road administrations are putting efforts in developing optimisation methodologies to enhance their decision making process, many still lack of data that allows the development of reliable prediction models for pavement performance. This is a key aspect to develop and test decision-making methodologies. Although there are several prediction models available in the literature, their practical applications are often limited to the very specific network from which data were retrieved at first and to a specific performance indicator (PI). This paper presents a practical application of a Markov model to predict the evolution of five PIs – cracking, skid resistance, bearing capacity, longitudinal evenness and transverse evenness – and consequent combined PIs, using historical data from an extensive pavement database. The conversion for PIs is made through a standardisation procedure proposed by an European COST Action, which may be considered a reference classification system for road administrations. The presented model is intended to be an useful input for researchers and administrations willing to develop and test different optimisation approaches.

Advanced tools and techniques to add value to soil stabilization practice

The aim of this paper is to demonstrate the advanced tools and techniques used for adding value to the soil stabilization practice. The tools presented involve advanced laboratory tests and modeling using codes and soft computing to evaluate the mechanical behavior of stabilized soils with cement, ranging from short-term to long-term behavior. More precisely, these tools are able to: 1. Predict the mechanical behavior of the stabilized soils over time from data obtained in the early ages saving time in laboratory tests; 2. Predict the mechanical behavior of the stabilized soils over time based on basic parameters of soil type and binder using historical accurate data, avoiding mechanical laboratory tests. 3. Incorporate the serviceability limit state concept in a novel proposal to estimate the design modulus in function of the uniaxial compressive strength and the strain level, making more economic and sustainable geotechnical solutions.