Showing papers by "Fu-Ping Gao published in 2021"
TL;DR: The results indicate that ML models outperform empirical prediction formulations with lower prediction error and the predicted correlations between input and output variables using five ML models show great agreement with the physical explanation.
Abstract: Compression index Cc is an essential parameter in geotechnical design for which the effectiveness of correlation is still a challenge. This paper suggests a novel modelling approach using machine learning (ML) technique. The performance of five commonly used machine learning (ML) algorithms, i.e. back-propagation neural network (BPNN), extreme learning machine (ELM), support vector machine (SVM), random forest (RF) and evolutionary polynomial regression (EPR) in predicting Cc is comprehensively investigated. A database with a total number of 311 datasets including three input variables, i.e. initial void ratio e0, liquid limit water content wL, plasticity index Ip, and one output variable Cc is first established. Genetic algorithm (GA) is used to optimize the hyper-parameters in five ML algorithms, and the average prediction error for the 10-fold cross-validation (CV) sets is set as the fitness function in the GA for enhancing the robustness of ML models. The results indicate that ML models outperform empirical prediction formulations with lower prediction error. RF yields the lowest error followed by BPNN, ELM, EPR and SVM. If the ranges of input variables in the database are large enough, BPNN and RF models are recommended to predict Cc. Furthermore, if the distribution of input variables is continuous, RF model is the best one. Otherwise, EPR model is recommended if the ranges of input variables are small. The predicted correlations between input and output variables using five ML models show great agreement with the physical explanation.
70 citations
TL;DR: In this article, the wave-induced pore pressure response in a sandy seabed was physically simulated with a large wave flume, where the breaking wave was generated by superimposing a series of longer waves onto the preceding shorter waves at a specified location.
Abstract: Previous studies on wave-induced pore pressure in a porous seabed mainly focused on non-breaking regular waves, e.g., Airy linear waves or Stokes non-linear waves. In this study, breaking-wave induced pore pressure response in a sandy seabed was physically simulated with a large wave flume. The breaking-wave was generated by superimposing a series of longer waves onto the foregoing shorter waves at a specified location. Water surface elevations and the corresponding pore pressure in the process of wave breaking were measured simultaneously at three typical locations, i.e., at the rear, just at, and in front of the wave breaking location. Based on test results, characterization parameters are proposed for the wave surface elevations and the corresponding pore-pressures. Flume observations indicate that the wave height was greatly diminished during wave breaking, which further affected the pore-pressure responses. Moreover, the measured values of the characteristic time parameters for the breaking-wave induced pore-pressure are larger than those for the free surface elevation of breaking-waves. Under the action of incipient-breaking or broken waves, the measured values of the amplitude of transient pore-pressures are generally smaller than the predicted results with the analytical solution by Yamamoto et al. (1978) for non-breaking regular waves with equivalent values of characteristic wave height and wave period.
7 citations
6 citations
TL;DR: In this paper, a coupled flow-seepage-elastoplastic modeling approach was proposed that could realize the synchronous simulation of the pipe hydrodynamics, the seepage flow, and elastoplastastic behavior of the seabed soil beneath the pipe.
Abstract: The instability of a partially embedded pipeline under ocean currents involves complex fluid–pipe–soil interactions, which may induce two typical instability modes; i.e., the lateral instability of the pipe and the tunnel erosion of the underlying soil. In previous studies, such two instability modes were widely investigated, but separately. To reveal the competition mechanism between the lateral instability and the tunnel erosion, a coupled flow-seepage-elastoplastic modeling approach was proposed that could realize the synchronous simulation of the pipe hydrodynamics, the seepage flow, and elastoplastic behavior of the seabed soil beneath the pipe. The coupling algorithm was provided for flow-seepage-elastoplastic simulations. The proposed model was verified through experimental and numerical results. Based on the instability criteria for the lateral instability and tunnel erosion, the two instability modes and their corresponding critical flow velocities could be determined. The instability envelope for the flow–pipe–soil interaction was established eventually, and could be described by three parameters; i.e., the critical flow velocity (Ucr), the embedment-to-diameter ratio (e/D), and the non-dimensional submerged weight of the pipe (G). There existed a transition line on the envelope when switching from one instability mode to the other. If the flow velocity of ocean currents gets beyond the instability envelope, either tunnel erosion or lateral instability could be triggered. With increasing e/D or concurrently decreasing G, the lateral instability was more prone to being triggered than the tunnel erosion. The present analyses may provide a physical insight into the dual-mode competition mechanism for the current-induced instability of submarine pipelines.
6 citations
TL;DR: In this article, a non-Darcy flow model is proposed based on a Karush-Kuhn-Tucker (KKT) condition, where the primal constraint arises from the fact that the tensile behavior does not exist in a noncohesive seabed, while the dual condition arises from physical evidences that the porefluid velocity increases during liquefaction.
4 citations