Numerical modeling and neural networks to identify model parameters from piezocone tests: II. Multi‐parameter identification from piezocone data

This paper completes the study presented in the accompanying paper, and demonstrates a numerical algorithm for parameter prediction from the piezocone test (CPTU) data. This part deals with a development of neural network (NN) models which are able to map multi‐variable input data onto typical geotechnical characteristics and constitutive parameters of the modified Cam clay model, which has been applied in this study. The identification procedure is designed for the coupled hydro‐mechanical boundary value problem in normally‐and lightly overconsolidated clayey soils including partially drained conditions that may appear during cone penetration. The NN models are trained with pseudo‐experimental measurements derived with the aid of the numerical model of the piezocone test, presented in the accompanying paper. Different input configurations containing CPTU measurements and some complementary data are studied with respect to the accuracy of predicted parameter values. Finally, the performance of the developed NN predictors is tested with field CPTU data which are derived from three well‐documented characterization sites, and the obtained predictions are compared with benchmark laboratory results. Copyright © 2011 John Wiley & Sons, Ltd.     

基于MPSO的RBF耦合算法的桩基动测参数辨识

变异粒子群优化算法（MPSO）是一种基于群体智能的改进全局优化技术,其优势在于减小陷入局部极值的机率,增加全局搜索能力。将变异粒子群算法与径向基函数（RBF）神经网络结构进行结合,建立了变异粒子群神经网络（MPSO-RBF）耦合算法,充分发挥了MPSO算法的全局寻优能力和RBF算法的局部搜索优势。数值计算结果表明,所建立的方法能够对桩基动测进行多参数的识别和非线性优化问题的求解,具有良好全局收敛能力,是一种行之有效的智能算法。     

Optimization framework for calibration of constitutive models enhanced by neural networks

A two‐level procedure designed for the estimation of constitutive model parameters is presented in this paper. The neural network (NN) approach at the first level is applied to achieve the first approximation of parameters. This technique is used to avoid potential pitfalls related to the conventional gradient‐based optimization techniques, considered here as a corrector that improves predicted parameters. The feed‐forward NN (FFNN) and the modified Gauss–Newton algorithms are briefly presented. The proposed framework is verified for the elasto‐plastic modified Cam Clay model that can be calibrated based on standard triaxial laboratory tests, i.e. the isotropic consolidation test and the drained compression test. Two different formulations of the input data to the NN, enhanced by a dimensional reduction of experimental data using principal component analysis, are presented. The determination of model characteristics is demonstrated, first on numerical pseudo‐experiments and then on the experimental data. The efficiency of the proposed approach by means of accuracy and computational effort is also discussed. Copyright © 2008 John Wiley & Sons, Ltd.     

Simulation of long‐term consolidation behavior of soft sensitive clay using an elasto‐viscoplastic constitutive model

The phenomenon of excess pore water pressure increase or stagnation and continuing large ground deformation in soft sensitive clay following the completion of construction of embankment is simulated for a case study at Saint Alban, Quebec, Canada. The present model employs an updated Lagrangian finite element framework and is combined with an automatic time increment selection scheme. The simulation based on an elasto‐viscoplastic constitutive model considers soil‐structure degradation effect. It is shown that without consideration for the microstructural degradation effect, it is not possible to reproduce the field responses of soft sensitive clay even during the construction of the embankment. When the soil‐structure degradation effect is considered, the present model can offer reasonably accurate prediction for the consolidation behavior of soft sensitive clay, including the so‐called anomalous pore water pressure generation and continuing large deformation even after the end of construction, which has been posing numerous uncertainties on the long‐term performance of earth structures. Copyright © 2013 John Wiley & Sons, Ltd.     

软土物理力学性质指标与微结构参数的灰色关联-神经网络模型

通过对广州市南沙地区大量软土物理力学试验和微结构分析,获取了40组软土试样的物理力学性质指标和微观结构参数。综合运用灰色关联分析的数据分析能力和人工神经网络的非线性映射功能,建立了软土物理力学性质指标与微结构参数的灰色关联-径向基神经网络模型。该模型利用灰色关联分析方法对数据进行预处理,提取重要因子作为网络的输入,而径向基神经网络充分利用样本数据信息,自适应确定隐含层节点个数、径向基函数中心、宽度以及网络的权系数。克服了传统RBF网络隐层节点数为样本个数,当数据较多时导致网络结构庞大、学习速度慢的缺点。通过模型A和模型B的实例研究表明,该方法简化了网络结构,提高了训练速度和预测精度,为软土物理力学性质与微结构参数关系的定量研究提供了一条有效途径。     

Non‐local damage modelling of concrete: a procedure for the determination of model parameters

This paper presents a procedure for the determination of parameters of non‐local damage models. This is to assure a consistent response of a non‐local damage model, as choice of the internal length and other parameters of the model are varied. Correlations between the internal length and other parameters governing the local constitutive behaviour of the model are addressed and exploited. Focus is put on the relationship between the internal length of the non‐local model and the width of the fracture process zone. Numerical examples are used to demonstrate the rigour of the proposed method. Copyright © 2006 John Wiley & Sons, Ltd.     

基于偏最小二乘与神经网络耦合的储层参数预测

将偏最小二乘回归（PLS）与神经网络（NN）耦合，建立了储层参数预报模型。利用偏最小二乘对影响储层参数的诸多因素进行分析，提取对因变量影响强的成分，从而克服了变量间的多重相关性问题，降低了神经网络的输入维数；同时，利用神经网络建模可以较好地解决非线性的储层参数预测问题。计算实例表明，本耦合模型的拟合和预报精度优于独立使用神经网络模型的精度。     

A self‐consistent approach for micro–macro modeling of elastic–plastic deformation in polycrystalline geomaterials

This paper is devoted to multi‐scale modeling of elastic–plastic deformation of a class of geomaterials with a polycrystalline microstructure. We have extended and improved the simplified polycrystalline model presented in [Zeng T. et al., 2014. Mech. Mater. 69 (1):132–145]. A rigorous and fully consistent self‐consistent (SC) scheme is proposed to describe the interaction among plastic mineral grains. We have also deeply discussed the numerical issues related to the numerical implementation of the proposed micromechanical model. The efficiency of the proposed model and the related numerical procedure is evaluated in several representative cases. We have compared the numerical results respectively obtained from the fully SC model and two simplified ones. It is found that the SC model produces a softer stress–strain response than that of the simplified models. The comparisons between the estimation of overall behavior of a granite in different loading conditions and experimental data are also conducted. Copyright © 2015 John Wiley & Sons, Ltd.     

A constitutive model for the dynamic and high‐pressure behaviour of a propellant‐like material: Part I: Experimental background and general structure of the model

This paper is the first part of a work that aims at developing a mechanical model for the behaviour of propellant‐like materials under high confining pressure and strain rate. The behaviour of a typical material is investigated experimentally. Several microstructural deformation processes are identified and correlated with loading conditions. The resulting behaviour is complex, non‐linear, and characterized by multiple couplings. The general structure of a relevant model is sought using a thermodynamic framework. A viscoelastic‐viscoplastic‐compaction model structure is derived under suitable simplifying assumptions, in the framework of finite, though moderate, strains. Model development, identification and numerical applications are given in the companion paper. Copyright © 2001 John Wiley & Sons, Ltd.     

A constitutive model for the dynamic and high‐pressure behaviour of a propellant‐like material: Part II: Model development and applications

This paper presents the second part of a work that aims at developing a mechanical model for the behaviour of propellant‐like materials under high confining pressure and strain rate. The model is to be employed to determine the temperature rise due to mechanical dissipation during a dynamic penetration event. Using the kinematical and thermodynamic background derived in the first part, a viscoelastic–viscoplastic–compaction model is put forward and identified. Viscoelasticity and compaction refer to the elastomeric nature of the material, while viscoplasticity, including implicit damage through dilatancy, reflects its granular nature. Some numerical exercises are performed, in view of determining the major model forces and weaknesses, and of assessing the numerical algorithm robustness. A penetration event is preliminarily simulated, and a temperature rise field predicted. Necessary model improvements are finally discussed, together with several ways of research for a longer term approach. Copyright © 2001 John Wiley & Sons, Ltd.     

A depositional model for a wave‐dominated open‐coast tidal flat,based on analyses of the Cambrian–Ordovician Lagarto and Palmares formations,north‐eastern Brazil

Open‐coast tidal flats are hybrid depositional systems resulting from the interaction of waves and tides. Modern examples have been recognized, but few cases have been described in ancient rock successions. An example of an ancient open‐coast tidal flat, the depositional architecture of the Lagarto and Palmares formations (Cambrian–Ordovician of the Sergipano Belt, north‐eastern Brazil) is presented here. Detailed field analyses of outcrops allowed the development of a conceptual architectural model for a coastal depositional environment that is substantially different from classical wave‐dominated or tide‐dominated coastal models. This architectural model is dominated by storm wave, low orbital velocity wave and tidal current beds, which vary in their characteristics and distribution. In a landward direction, the storm deposits decrease in abundance, dimension (thickness and spacing) and grain size, and vary from accretionary through scour and drape to anisotropic hummocky cross‐stratification beds. Low orbital wave deposits are more common in the medium and upper portion of the tidal flat. Tidal deposits, which are characterized by mudstone interbedded with sandstone strata, are dominant in the landward portion of the tidal flat. Hummocky cross‐stratification beds in the rock record are believed, in general, to represent storm deposits in palaeoenvironments below the fair‐weather wave base. However, in this model of an open‐coast tidal flat, hummocky cross‐stratification beds were found in very shallow waters above the fair‐weather wave base. Indeed, this depositional environment was characterized by: (i) fair‐weather waves and tides that lacked sufficient energy to rework the storm deposits; (ii) an absence of biological communities that could disrupt the storm deposits; and (iii) high aggradation rates linked to an active foreland basin, which contributed definitively to the rapid burial and preservation of these hummocky cross‐stratification deposits.