Prediction Of Liquefaction Research Articles

Prediction of Wave-induced Liquefaction using Artificial Neural Network and Wide Genetic Algorithm

Prediction of Wave-induced Liquefaction using Artificial Neural Network and Wide Genetic Algorithm

Prediction of Wave-induced Liquefaction using Artificial Neural Network and Wide Genetic Algorithm


 The hassle of analytical and numerical solution for liquefaction modeling, repetitive laboratory testing and expensive field observations, have opened opportunities to develop simple, practical, inexpensive and valid prediction of wave-induced liquefaction. In this study, Artificial Neural Network (ANN) regression modeling is used to predict the depth of liquefaction. Despite of using Back Propagation (BP) to train ANN, a modified Genetic Algorithm (called as Wide GA, WGA) is used as ANN training method to improve ANN prediction accuracy and to overcome BP weaknesses such as premature convergence and local optimum. WGA also aim to avoid conventional GA weaknesses such as low population diversity and narrow search coverage. Key WGA operations are Wide Tournament Selection, Multi-Parent BLX-? Crossover, Aggregate Mate Pool Mutation and Direct Fresh Mutation-Crossover. ANN prediction accuracy measured by Median APE (MdAPE). Global optimum solution of WGA is best ANN connections weights configuration with smallest MdAPE.

Liquefaction prediction using CPT data by triangular chart identification

Cone Penetration Test (CPT) is a common in-situ test renowned in liquefaction prediction. Description, deterministic and probabilistic models are used for predicting the occurrence of soil liquefaction. In this paper, new format of classification charts, i.e. triangular are proposed to soil liquefaction identification based on CPT data, including field investigation records for 18 earthquakes. Triangular charts containing normalised cyclic stress ratio for earthquake magnitude 7.5 (CSR 7.5), corrected for overburden stress cone tip resistance (q c1) and friction ratio (R f ). This study addresses capability of new triangular charts to predict liquefaction potential based on specified variables. The performance of the proposed approach was compared within a quantitative validation framework (F-score method) over available methods. Results indicate that the new approach is appreciable in prediction of the liquefaction which can be realised in practice.

Experimental and conceptual evidence about the limitations of shear wave velocity to predict liquefaction

Based on the liquefaction performance of sites with seismic activity, the normalized shear wave velocity, Vs1, has been proposed as a field parameter for liquefaction prediction. Because shear wave velocity, Vs, can be measured in the field with less effort and difficulty than other field tests, its use by practitioners is highly attractive. However, considering that its measurement is associated with small strain levels, of the order of 10−4–10−3%, Vs reflects the elastic stiffness of a granular material, hence, it is mainly affected by soil type, confining pressure and soil density, but it is insensitive to factors such as overconsolidation and pre-shaking, which have a strong influence on the liquefaction resistance. Therefore, without taking account of the important factors mentioned above, the correlation between shear wave velocity and liquefaction resistance is weak.In this paper, laboratory test results are presented in order to demonstrate the significant way in which OCR (overconsolidation ratio) affects both shear wave velocity and liquefaction resistance. While Vs is insensitive to OCR, the liquefaction resistance increases significantly with OCR. In addition, the experimental results also confirm that Vs correlates linearly with void ratio, regardless of the maximum and minimum void ratios, which means that Vs is unable to give information about the relative density. Therefore, if shear wave velocity is used to predict liquefaction potential, it is recommended that the limitations presented in this paper be taken into account.

Liquefaction Analyses of Reclaimed Ground and Levee Considering the Damage by the 2011 Great Tohoku Earthquake and Lessons

We have reviewed liquefaction-induced damage to reclaimed ground, levees and industrial complexes, caused by the 2011 off the Pacific Coast of Tohoku Earthquake. Then, we reconsidered the method of the prediction of liquefaction and estimated liquefaction at Urayasu housing area using a new liquefaction prediction method based on dynamic numerical liquefaction analysis of the ground. We also clarified the feature of the damage of levee due to the 2011 Great Tohoku Earthquake based on the dynamic liquefaction analysis and listed the lessons taught by the analyses and the damages due to the 2011 Tohoku Great Earthquake are shown.

Simplified Liquefaction Prediction and Assessment Method Considering Waveforms and Durations of Earthquakes

AbstractThis paper presents a new simplified liquefaction prediction and assessment method that is capable of considering the influence of the waveforms and durations of earthquakes. The concept of...

Assessment of seismic liquefaction potential based on Bayesian network constructed from domain knowledge and history data

Prediction of seismic liquefaction is difficult due to the uncertainties and complexity of multiple related factors. Bayesian network is a just right effective tool to deal the problem because of merging multiple source information and domain knowledge in a consistent system, reflecting and analyzing the interdependent uncertain relationships between variables. This paper used two ways to construct generic Bayesian network models with twelve significant factors of seismic liquefaction, of which the first model is constructed only by interpretive structural modeling and causal mapping approach for incomplete data contained huge missing values. Another one is constructed by combining K2 algorithm and domain knowledge for complete data. Compared with artificial neural network and support vector machine using 5-fold cross-validation, the two Bayesian network models provided a better performance, and the second Bayesian network model is slightly better than the first one. This paper also offers a sensitivity analysis of the input factors. In the twelve variables, standard penetration test number, soil type, vertical effective stress, depth of soil deposit, and peak ground acceleration have more significant influences on seismic liquefaction than others. Our results suggest that the Bayesian network is useful for prediction of seismic liquefaction and is simple to perform in practice.

Calibration of an Advanced Constitutive Model for Babolsar Sand Accompanied by Liquefaction Analysis

In our research, we apply numerical modeling for prediction of liquefaction of sands during and after dynamic loading. In numerical modeling, to properly simulate the generation, redistribution, and dissipation of excess pore water pressure during and after dynamic loading, it is important to use a suitable constitutive model for soil. In this article, Dafalias and Manzari’s model [2004] (a critical state bounding surface plasticity model) was used to model the behavior of saturated sand due to relatively simple of formulations and a unique set of input parameters for a wide range of initial stress and void ratio. The attention in this article is on Babolsar sand. After calibration model parameters for Babolsar sand, the analysis of liquefaction using the modeling of a centrifuge test and predictions of model was carried out. The results indicate a reasonable performance of the model for prediction of behavior of types of sands. Also, Babolsar sand has more prone to dilatancy than Nevada and Toyoura sands.

Metaheuristic optimization within machine learning-based classification system for early warnings related to geotechnical problems

Abstract This study proposes a novel classification system integrating swarm and metaheuristic intelligence, i.e., a smart firefly algorithm (SFA), with a least squares support vector machine (LSSVM). Benchmark functions were used to validate the optimization performance of the SFA. The experimental results showed that the SFA obtained 100% success rate in searching the optimum for most benchmark functions. The SFA was then integrated with the LSSVM to create a metaheuristic optimized classification model. A graphical user interface was developed for the proposed classification system to assist engineers and researchers in executing advanced machine learning tasks. The system was applied to several geotechnical engineering problems that involved measuring the groutability of sandy silt soil, monitoring seismic hazards in coal mines, predicting postearthquake soil liquefaction, and determining the propensity of slope collapse. The prediction problems in these studies were complex because they were dependent on various physical factors, and such factors exhibited highly nonlinear relations. The analytical results revealed that the metaheuristic optimization within machine learning-based classification system exhibited a groutability prediction accuracy of 95.42%, seismic prediction accuracy of 93.96%, soil liquefaction prediction accuracy of 95.18%, and soil collapse prediction accuracy of 95.45%. Hence, the proposed system is a promising tool to provide decision-makers with timely warnings of geotechnical hazards.

Application of genetic algorithm-based support vector machines for prediction of soil liquefaction

This paper presents a hybrid genetic algorithm (GA) and support vector machine (SVM) techniques to predict the potential of soil liquefaction. GA is employed in selecting the optimal values of the kernel function and the penalty parameter in SVM model to improve the forecasting accuracy. The database used in this study includes 109 CPT-based field observations from five major earthquakes between 1964 and 1983. Several important parameters, including the cone resistance, total vertical stress, effective vertical stress, mean grain size, normalized peak horizontal acceleration at ground surface, cyclic stress ratio, and earthquake magnitude, were used as the input parameters, while the potential of soil liquefaction was the output parameter. The predictions from the GA-SVM model were compared with those from three methods: grid search (GS) method, artificial neural network (ANN) model, and C4.5 decision tree approach. The overall classification success rates for the entire dataset predicted by GA-SVM, ANN, C4.5 decision tree, and GS-SVM models are 97.25, 97.2, 96.3, and 92.66 %, respectively. The study concluded that the proposed GA-SVM model improves the classification accuracy and is a feasible method in predicting soil liquefaction.

電子地盤モデルを用いた浦安市域の東日本大震災時の液状化判定

電子地盤モデルを用いた浦安市域の東日本大震災時の液状化判定

A Bayesian network approach for predicting seismic liquefaction based on interpretive structural modeling

ABSTRACTThe Bayesian network (BN) is a type of graphical network based on probabilistic inference that has been gradually applied to assessment of seismic liquefaction potential. However, how to construct a robust BN remains underexplored in this field. This paper aims to present an efficient hybrid approach combining domain knowledge and data to construct a BN that facilitates the integration of multiple factors and the quantification of uncertainties within a network model to assess seismic liquefaction. Initially, only using given domain knowledge, a naive network model can be constructed using interpretive structural modeling. Thereafter, some effective information about the naive model is provided to construct a robust model using structural learning of BN from historic data. Finally, the returning predictive results and the predictive results are compared to other methods including non-probabilistic and probabilistic models for seismic liquefaction using the metrics of the overall accuracy, the area under the curve of receiver operating characteristic, prediction, recall and F1 score. The methodology proposed in this paper achieved better performance, and we discussed the power and value of the proposed approach at the end of this paper, which suggest that BN is a good alternative tool for seismic liquefaction prediction.

Determination of seismic liquefaction potential of soil based on strain energy concept

In the present study, minimax probability machine regression (MPMR) and extreme learning machine (ELM) have been adopted for prediction of seismic liquefaction of soil based on strain energy. Initial effective mean confining pressure ( $$ \sigma_{\text{mean}}^{\prime} $$ ), initial relative density after consolidation (D r), percentage of fines content (FC), coefficient of uniformity (C u), and mean grain size (D 50) have been taken as inputs of MPMR and ELM models. MPMR and ELM have been used as regression techniques. The performances of MPMR and ELM have been compared with the artificial neural network. A sensitivity analysis has been carried out to determine the effect of each input. The experimental results demonstrate that proposed methods are robust models for determination seismic liquefaction potential of soil based on strain energy.

Technical Notoe: Prediction of Static Liquefaction by Nor Sand Constitutive Model

Abstract The paper gives a short description of unstable behaviour of saturated sand under undrained monotonic loading. Constitutive model Nor Sand capable to describe static liquefaction is presented. The model is based on critical state soil mechanics and assumes associated flow rule. Hardening law incorporates the state parameter proposed earlier by Been and Jefferies. Results of numerical simulations of undrained element tests have been presented and discussed.

Deformation Behaviour on the Simulation of Liquefaction in Particle Geomaterials

Liquefaction has recently been shown to occur in constitutive models at a critical value for the hardening modulus. Thus, the nearness of a soil element to liquefying at a given instant can be determined by finding the difference between its hardening modulus and its critical hardening modulus. The constitutive functions of the pore medium were described as follows: the evolution of the constitutive effective stress with imposed solid matrix deformation; the intrinsic mass densities with intrinsic pressure on all three phase; and the relative flow vector with intrinsic pressure for the water and air phase. The revelation has been applied herein to map the progression of static liquefaction in large scale boundary value problems under monotonic loading conditions. Simulations are presented to demonstrate how the proposed criterion can be applied to real-world situations. In addition, the influences of the loading rate and the mesh size on the liquefaction prediction are examined. The methodology proposed herein provides a powerful means of assessing liquefaction risk based on solid mechanical theory rather than empiricism. Results of the mesh size effect showed that the axial strains with the onset of the deformation band were different, if the mesh sizes were not the same even in the same sample.

砂基液化的因素筛选及预测模型

为降低数据维数，简化数据运算，我们采用因子分析和判别分析相结合的方法，运用方差累计贡献率在85%以上的前k个主成分代替原始砂基液化的有关因素，对砂基液化因素进行分析，这种方法并没有缩减样本量，只是对原始数据进行了浓缩和综合，通过对得到的因子得分数据进行判别分析，可得到一组判别结果。另外，利用因子分析的提取方法得到变量的共同度，变量共同度高的表示变量中的大部分信息均能够被因子所提取，选出变量共同度较高的对应的变量，利用这些变量再次进行判别分析，对两次判别分析得到的结果与原结果进行汇总对比，分析误判率。结果表明，这两种方法的结合在一定程度上用于筛选砂基液化的主要因素以及预测砂基液化可行性强，效果较好。 In order to reduce data dimension, simplify data operation, we adopted the method combining the factor analysis and discriminant analysis, and applied the cumulative variance contribution rate of k in front of more than 85% of the principal components instead of the original related factors of sand liquefaction to analyze, this method didn’t reduce sample size, just made the raw data enrich- ment and comprehensive, did the discriminant analysis based on the factor score data, a set of discriminant results can be obtained. In addition, the extraction methods of principle component analysis were used to get the variable joint degrees, high variable joint degrees indicated the most information can be extracted by factor, then found the corresponding variable, did the discriminant analysis using these variables again, the two discriminant analysis results were compared with the original results and analyzed the misjudgment rate. Results show that the combination of the two methods has strong feasibility in filtering the main factors of sandlique faction and the prediction of sand liquefaction to some extent, and the effect is better.

Wave Prediction Using Neural Networks at New Mangalore Port along West Coast of India

Estimation of wave parameters is of great importance in coastal activities such as design studies for harbour, inshore and offshore structures, coastal erosion, sediment transport and wave energy estimation. Traditional methods like semi empirical formulations and numerical models have disadvantages of excessive data requirement, time consumption and are tedious to carry out.Artificial Intelligent methods like ANN (Artificial Neural Network) are powerful and flexible modeling tool is being widely used in coastal engineering field since last two decades for solving various problems related to time series forecasting of waves and tides; prediction of sea-bed liquefaction and scour depth; and estimation of design parameters of coastal engineering structures. It imbibes the qualities of exploiting non-linearity, adaptability to adjust free parameters by mapping input output data sets using various learning algorithms and fault tolerance.Using input and output information, intelligent systems enabled to estimate the hidden law behind the information and to find their relations among them. In the present study attempt has been made to predict waves at New Mangalore Port Trust (NMPT) located along the west coast of India using Feed Forward Back Propagation (FFBP) with LM algorithm and a recurrent network called Non-linear Auto Regressive with exogenous input(NARX) network. Field data of NMPT has been used to train and test the network performance, which are measured in terms of mean square error (mse) and correlation coefficient (r). Effect of network architecture on the performance of model has been studied.Correlation coefficient is found to be 0.94 in case of NARX predictions indicating better performance than FFBP network whose ‘r’ value is 0.9. It was found that for time series prediction NARX network outperform FFBP network not only in terms of accuracy but also in terms of time required for computation.

Prediction of Soil Liquefaction by Using UBC3D-PLM Model in PLAXIS

Prediction of Soil Liquefaction by Using UBC3D-PLM Model in PLAXIS

大阪・和歌山平野の液状化危険度と地下水位低下による対策効果の予測

大阪・和歌山平野の液状化危険度と地下水位低下による対策効果の予測

A modification to dense sand dynamic simulation capability of Pastor–Zienkiewicz–Chan model

A modification to the nonlinear Pastor–Zienkiewicz–Chan (PZC) constitutive model without any change in the number of model parameters is introduced in order to simulate stiffness degradation of dense sands at dynamic loading. The PZC model is based on generalized plasticity and was verified by good prediction of liquefaction and undrained behavior of saturated sand. The PZC is a robust model that can predict drained dynamic behavior of sands, especially stiffness increase in loose sand at reloading of dynamic loading. Yet, this model does not show stiffness degradation of dense sand at reloading. The modification is made through modifying the stress memory factor, HDM, which is multiplied by the plastic modulus, HL. This modification does not influence reloading behavior of loose sand. The modified PZC model is verified via results of drained cyclic tests. Two cyclic triaxial tests on loose and dense specimens, along with two cyclic plane strain tests on dense sand are utilized for validation. The model simulation shows that the modified PZC model is able to predict the stiffness degradation of dense sand at reloading well.