Prediction of river flow using hybrid neuro-fuzzy models

The complex nature of hydrological phenomena, like rainfall and river flow, causes some limitations for some admired soft computing models in order to predict the phenomenon. Evolutionary algorithms (EA) are novel methods that used to cover the weaknesses of the classic training algorithms, such as trapping in local optima, poor performance in networks with large parameters, over-fitting, and etc. In this study, some evolutionary algorithms, including genetic algorithm (GA), ant colony optimization for continuous domain (ACO_R), and particle swarm optimization (PSO), have been used to train adaptive neuro-fuzzy inference system (ANFIS) in order to predict river flow. For this purpose, classic and hybrid ANFIS models were trained using river flow data obtained from upstream stations to predict 1-, 3-, 5-, and 7-day ahead river flow of downstream station. The best inputs were selected using correlation coefficient and a sensitivity analysis test (cosine amplitude). The results showed that PSO improved the performance of classic ANFIS in all the periods such that the averages of coefficient of determination, R², root mean square error, RMSE (m³/s), mean absolute relative error, MARE, and Nash-Sutcliffe efficiency coefficient (NSE) were improved up to 0.19, 0.30, 43.8, and 0.13%, respectively. Classic ANFIS was only capable to predict river flow in 1-day ahead while EA improved this ability to 5-day ahead. Cosine amplitude method was recognized as an appropriate sensitivity analysis method in order to select the best inputs.     

Experimental Studies on Permeability of Intact and Singly Jointed Meta-Sedimentary Rocks Under Confining Pressure

Three different types of permeability tests were conducted on 23 intact and singly jointed rock specimens, which were cored from rock blocks collected from a rock cavern under construction in Singapore. The studied rock types belong to inter-bedded meta-sandstone and meta-siltstone with very low porosity and high uniaxial compressive strength. The transient pulse water flow method was employed to measure the permeability of intact meta-sandstone under a confining pressure up to 30 MPa. It showed that the magnitude order of meta-sandstone’s intrinsic permeability is about 10^?18 m². The steady-state gas flow method was used to measure the permeability of both intact meta-siltstone and meta-sandstone in a triaxial cell under different confining pressures spanning from 2.5 to 10 MPa. The measured permeability of both rock types ranged from 10^?21 to 10^?20 m². The influence of a single natural joint on the permeability of both rock types was studied by using the steady-state water flow method under different confining pressures spanning from 1.25 to 5.0 MPa, including loading and unloading phases. The measured permeability of both jointed rocks ranged from 10^?13 to 10^?11 m², where the permeability of jointed meta-siltstone was usually slightly lower than that of jointed meta-sandstone. The permeability of jointed rocks decreases with increasing confining pressure, which can be well fitted by an empirical power law relationship between the permeability and confining pressure or effective pressure. The permeability of partly open cracked specimens is lower than that of open cracked specimens, but it is higher than that of the specimen with a dominant vein for the meta-sandstone under the same confining pressure. The permeability of open cracked rock specimens will partially recover during the unloading confining pressure process. The equivalent crack (joint) aperture is as narrow as a magnitude order of 10^?6 m (1 μm) in the rock specimens under confining pressures spanning from 1.25 to 5.0 MPa, which represent the typical ground stress conditions in the cavern. The in situ hydraulic conductivity measurements conducted in six boreholes by the injection test showed that the in situ permeability of rock mass varies between 10^?18 and 10^?11 m². The lower bound of the in situ permeability is larger than that of the present laboratory-tested intact rock specimens, while the upper bound of the in situ permeability is less than that of the present laboratory-tested jointed rock specimens. The in situ permeability test results were thus compatible with our present laboratory permeability results of both intact and jointed rock specimens.     

A study of soft computing models for prediction of longitudinal wave velocity

Genetic algorithm (GA) and support vector machine (SVM) optimization techniques are applied widely in the area of geophysics, civil, biology, mining, and geo-mechanics. Due to its versatility, it is being applied widely in almost every field of engineering. In this paper, the important features of GA and SVM are discussed as well as prediction of longitudinal wave velocity and its advantages over other conventional prediction methods. Longitudinal wave measurement is an indicator of peak particle velocity (PPV) during blasting and is an important parameter to be determined to minimize the damage caused by ground vibrations. The dynamic wave velocity and physico-mechanical properties of rock significantly affect the fracture propagation in rock. GA and SVM models are designed to predict the longitudinal wave velocity induced by ground vibrations. Chaos optimization algorithm has been used in SVM to find the optimal parameters of the model to increase the learning and prediction efficiency. GA model also has been developed and has used an objective function to be minimized. A parametric study for selecting the optimized parameters of GA model was done to select the best value. The mean absolute percentage error for the predicted wave velocity (V) value has been found to be the least (0.258 %) for GA as compared to values obtained by multivariate regression analysis (MVRA), artificial neural network (ANN), adaptive neuro-fuzzy inference system (ANFIS), and SVM.     

A new developed approach for the prediction of ground vibration using a hybrid PSO-optimized ANFIS-based model

Ground vibration is one of the common environmental effects of blasting operation in mining industry, and it may cause damage to the nearby structures and the surrounding residents. So, precise estimation of blast-produced ground vibration is necessary to identify blast-safety area and also to minimize environmental effects. In this research, a hybrid of adaptive neuro-fuzzy inference system (ANFIS) optimized by particle swarm optimization (PSO) was proposed to predict blast-produced ground vibration in Pengerang granite quarry, Malaysia. For this goal, 81 blasting were investigated, and the values of peak particle velocity, distance from the blast-face and maximum charge per delay were precisely measured. To demonstrate the performance of the hybrid PSO–ANFIS, ANFIS, and United States Bureau of Mines empirical models were also developed. Comparison of the predictive models was demonstrated that the PSO–ANFIS model [with root-mean-square error (RMSE) 0.48 and coefficient of determination (R ²) of 0.984] performed better than the ANFIS with RMSE of?1.61 and R ² of 0.965. The mentioned results prove the superiority of the newly developed PSO–ANFIS model in estimating blast-produced ground vibrations.     

Analysis of soil erosion characteristics in small watersheds with particle swarm optimization, support vector machine, and artificial neuronal networks

Sand production by soil erosion in small watershed is a complex physical process. There are few physical models suitable to describe the characteristics of the intense erosion in domestic loess plateau. Introducing support vector machine (SVM) oriented to small sample data and possessing good extension property can be an effective approach to predict soil erosion because SVM has been applied in hydrological prediction to some extent. But there are no effective methods to select the rational parameters for SVM, which seriously limited the practical application of SVM. This paper explored the application of intelligence-based particle swarm optimization (PSO) algorithm in automatic selection of parameters for SVM, and proposed a prediction model by linking PSO and SVM for small sample data analysis. This method utilized the high efficiency optimization property and swarm paralleling property of PSO algorithm and the relatively strong learning and extending capacity of SVM. For an example of Huangfuchuan small watershed, its intensive fragmentation and intense erosion earn itself the name of “worst erosion in the world”. Using four characteristics selection algorithms of correlation feature selection, the primary affecting factors for soil erosion in this small watershed were determined to be the channel density, ravine area, sand rock proportion, and the total vegetation coverage. Based on the proposed PSO–SVM algorithm, the soil erosion modulus in the small watershed was predicted. The accuracy of the simulation and prediction was good, and the average error was 3.85%. The SVM predicting model was based on the monitoring data of sand production. The construction of the SVM erosion modulus prediction model for the small watershed comprehensively reflected the complex mechanism of soil erosion and sand production. It had certain advantage and relatively high practical value in small sample prediction in the discipline of soil erosion.     

Applicability of Statistical Learning Algorithms for Spatial Variability of Rock Depth

Two algorithms are outlined, each of which has interesting features for modeling of spatial variability of rock depth. In this paper, reduced level of rock at Bangalore, India, is arrived from the 652 boreholes data in the area covering 220 sq⋅km. Support vector machine (SVM) and relevance vector machine (RVM) have been utilized to predict the reduced level of rock in the subsurface of Bangalore and to study the spatial variability of the rock depth. The support vector machine (SVM) that is firmly based on the theory of statistical learning theory uses regression technique by introducing ε-insensitive loss function has been adopted. RVM is a probabilistic model similar to the widespread SVM, but where the training takes place in a Bayesian framework. Prediction results show the ability of learning machine to build accurate models for spatial variability of rock depth with strong predictive capabilities. The paper also highlights the capability of RVM over the SVM model.     

Monthly reservoir inflow forecasting using a new hybrid SARIMA genetic programming approach

Forecasting reservoir inflow is one of the most important components of water resources and hydroelectric systems operation management. Seasonal autoregressive integrated moving average (SARIMA) models have been frequently used for predicting river flow. SARIMA models are linear and do not consider the random component of statistical data. To overcome this shortcoming, monthly inflow is predicted in this study based on a combination of seasonal autoregressive integrated moving average (SARIMA) and gene expression programming (GEP) models, which is a new hybrid method (SARIMA–GEP). To this end, a four-step process is employed. First, the monthly inflow datasets are pre-processed. Second, the datasets are modelled linearly with SARIMA and in the third stage, the non-linearity of residual series caused by linear modelling is evaluated. After confirming the non-linearity, the residuals are modelled in the fourth step using a gene expression programming (GEP) method. The proposed hybrid model is employed to predict the monthly inflow to the Jamishan Dam in west Iran. Thirty years’ worth of site measurements of monthly reservoir dam inflow with extreme seasonal variations are used. The results of this hybrid model (SARIMA–GEP) are compared with SARIMA, GEP, artificial neural network (ANN) and SARIMA–ANN models. The results indicate that the SARIMA–GEP model (R ²=78.8, VAF =78.8, RMSE =0.89, MAPE =43.4, CRM =0.053) outperforms SARIMA and GEP and SARIMA–ANN (R ²=68.3, VAF =66.4, RMSE =1.12, MAPE =56.6, CRM =0.032) displays better performance than the SARIMA and ANN models. A comparison of the two hybrid models indicates the superiority of SARIMA–GEP over the SARIMA–ANN model.     

Predicting longitudinal dispersion coefficient using ANN with metaheuristic training algorithms

A reliable prediction of dispersion coefficient can provide valuable information for environmental scientists and river engineers as well. The main objective of this study is to apply intelligence techniques for predicting longitudinal dispersion coefficient in rivers. In this regard, artificial neural network (ANN) models were developed. Four different metaheuristic algorithms including genetic algorithm (GA), imperialist competitive algorithm (ICA), bee algorithm (BA) and cuckoo search (CS) algorithm were employed to train the ANN models. The results obtained through the optimization algorithms were compared with the Levenberg–Marquardt (LM) algorithm (conventional algorithm for training ANN). Overall, a relatively high correlation between measured and predicted values of dispersion coefficient was observed when the ANN models trained with the optimization algorithms. This study demonstrates that the metaheuristic algorithms can be successfully applied to make an improvement on the performance of the conventional ANN models. Also, the CS, ICA and BA algorithms remarkably outperform the GA and LM algorithms to train the ANN model. The results show superiority of the performance of the proposed model over the previous equations in terms of DR, R ² and RMSE.     

A derivative-free methodology with local and global search for the constrained joint optimization of well locations and controls

In oil field development, the optimal location for a new well depends on how it is to be operated. Thus, it is generally suboptimal to treat the well location and well control optimization problems separately. Rather, they should be considered simultaneously as a joint problem. In this work, we present noninvasive, derivative-free, easily parallelizable procedures to solve this joint optimization problem. Specifically, we consider Particle Swarm Optimization (PSO), a global stochastic search algorithm; Mesh Adaptive Direct Search (MADS), a local search procedure; and a hybrid PSO–MADS technique that combines the advantages of both methods. Nonlinear constraints are handled through use of filter-based treatments that seek to minimize both the objective function and constraint violation. We also introduce a formulation to determine the optimal number of wells, in addition to their locations and controls, by associating a binary variable (drill/do not drill) with each well. Example cases of varying complexity, which include bound constraints, nonlinear constraints, and the determination of the number of wells, are presented. The PSO–MADS hybrid procedure is shown to consistently outperform both stand-alone PSO and MADS when solving the joint problem. The joint approach is also observed to provide superior performance relative to a sequential procedure.     

Intelligent modelling of clay compressibility using hybrid meta-heuristic and machine learning algorithms

Compression index Ccis 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 thefitness 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.     

基于遗传算法的优化搜索技术

遗传算法是借鉴生物界自然选择和自然遗传机制的随机化搜索算法,其主要特点是搜索不依赖于梯度信息,能自动获取和积累有关搜索空间的知识,并自适应地控制搜索过程,从而得到最优解或准最优解的通用搜索算法。它尤其适合于组合优化、机器学习、自适应控制、规划设计和人工智能等领域。     

Estimation of air-overpressure produced by blasting operation through a neuro-genetic technique

Blasting operations usually produce significant environmental problems which may cause severe damage to the nearby areas. Air-overpressure (AOp) is one of the most important environmental impacts of blasting operations which needs to be predicted and subsequently controlled to minimize the potential risk of damage. In order to solve AOp problem in Hulu Langat granite quarry site, Malaysia, three non-linear methods namely empirical, artificial neural network (ANN) and a hybrid model of genetic algorithm (GA)–ANN were developed in this study. To do this, 76 blasting operations were investigated and relevant blasting parameters were measured in the site. The most influential parameters on AOp namely maximum charge per delay and the distance from the blast-face were considered as model inputs or predictors. Using the five randomly selected datasets and considering the modeling procedure of each method, 15 models were constructed for all predictive techniques. Several performance indices including coefficient of determination (R ²), root mean square error and variance account for were utilized to check the performance capacity of the predictive methods. Considering these performance indices and using simple ranking method, the best models for AOp prediction were selected. It was found that the GA–ANN technique can provide higher performance capacity in predicting AOp compared to other predictive methods. This is due to the fact that the GA–ANN model can optimize the weights and biases of the network connection for training by ANN. In this study, GA–ANN is introduced as superior model for solving AOp problem in Hulu Langat site.     

Interval prediction of landslide displacement with dual-output least squares support vector machine and particle swarm optimization algorithms

For landslide displacement, interval predictions are generally more realistic and reliable compared with traditional point predictions. This paper presents a new interval prediction method for landslide displacement integrating dual-output least squares support vector machine (DO-LSSVM) and particle swarm optimization (PSO) algorithms. In this new method, the PSO algorithm is employed to optimize coefficients of the least squares support vector machine (LSSVM) model for obtaining point prediction results, and the interval prediction of the landslide displacement is made based on the dual-outputs obtained from the DO-LSSVM model. To assess the rationality of the predictions, three performance evaluation indicators, including the prediction interval coverage probability (PICP), normalized mean prediction interval width (NMPIW), and coverage width-based criterion (CWC), are established. Case studies of the Tanjiahe landslide and the Baishuihe landslide in the Three Gorges Reservoir region are then used to demonstrate the effectiveness of the proposed method in predicting the landslide displacement interval. The case study results demonstrate that this new method has the best overall performance compared with other existing methods, and this new method can provide accurate and reliable results for the medium- to long-term interval prediction of landslide displacement.

     

基于遗传支持向量机的地下工程裂隙岩体注浆量预测

提出了地下工程裂隙岩体注浆量预测的遗传支持向量机方法,通过支持向量机对实际注浆数据样本进行学习,建立注浆量及其影响因素之间的非线性映射关系,基于这种关系实现注浆量的预测。模型建立过程中,考虑到支持向量机惩罚因子和核参数对预测精度的影响,以预测误差为适应度,采用遗传算法对最佳参数进行搜索。结果表明,本文方法计算快速,预测精度高,是一种注浆量预测的好方法。     

A model for coupled mechanical and hydraulic behaviour of a rock joint

Constitutive laws for rock joints should be able to reproduce the fundamental mechanical behaviour of real joints, such as dilation under shear and strain softening due to surface asperity degradation. In this work, we extend the model of Plesha to include hydraulic behaviour. During shearing, the joint can experience dilation, leading to an initial increase in its permeability. Experiments have shown that the rate of increase of the permeability slows down as shearing proceeds, and, at later stages, the permeability could decrease again. The above behaviour is attributed to gouge production. The stress–strain relationship of the joint is formulated by appeal to classical theories of interface plasticity. It is shown that the parameters of the model can be estimated from the Barton–Bandis empirical coefficients; the Joint Roughness Coefficient (JRC) and the Joint Compresive strength (JSC). We further assume that gouge production is also related to the plastic work of the shear stresses, which enables the derivation of a relationship between the permeability of the joint and its mechanical aperture. The model is implemented in a finite element code (FRACON) developed by the authors for the simulation of the coupled thermal–hydraulic–mechanical behaviour of jointed rock masses. Typical laboratory experiments are simulated with the FRACON code in order to illustrate the trends predicted in the proposed model. © 1998 by John Wiley & Sons. Ltd.     

Hybrid differential evolution and particle swarm optimization for optimal well placement

There is no gainsaying that determining the optimal number, type, and location of hydrocarbon reservoir wells is a very important aspect of field development planning. The reason behind this fact is not farfetched—the objective of any field development exercise is to maximize the total hydrocarbon recovery, which for all intents and purposes, can be measured by an economic criterion such as the net present value of the reservoir during its estimated operational life-cycle. Since the cost of drilling and completion of wells can be significantly high (millions of dollars), there is need for some form of operational and economic justification of potential well configuration, so that the ultimate purpose of maximizing production and asset value is not defeated in the long run. The problem, however, is that well optimization problems are by no means trivial. Inherent drawbacks include the associated computational cost of evaluating the objective function, the high dimensionality of the search space, and the effects of a continuous range of geological uncertainty. In this paper, the differential evolution (DE) and the particle swarm optimization (PSO) algorithms are applied to well placement problems. The results emanating from both algorithms are compared with results obtained by applying a third algorithm called hybrid particle swarm differential evolution (HPSDE)—a product of the hybridization of DE and PSO algorithms. Three cases involving the placement of vertical wells in 2-D and 3-D reservoir models are considered. In two of the three cases, a max-mean objective robust optimization was performed to address geological uncertainty arising from the mismatch between real physical reservoir and the reservoir model. We demonstrate that the performance of DE and PSO algorithms is dependent on the total number of function evaluations performed; importantly, we show that in all cases, HPSDE algorithm outperforms both DE and PSO algorithms. Based on the evidence of these findings, we hold the view that hybridized metaheuristic optimization algorithms (such as HPSDE) are applicable in this problem domain and could be potentially useful in other reservoir engineering problems.     

Prediction of Enhanced Soil–Anchored Geogrid Interactions in Direct Shear Mode Using Gene Expression Programming

Design of reinforced soil structures is greatly influenced by soil–geosynthetic interactions at interface which is normally assessed by costly and time consuming laboratory tests. In present research, using the results of large-scale direct shear tests conducted on soil–anchored geogrid samples a model for predicting Enhanced Interaction Coefficient (EIC) is proposed enabling researchers/engineers easily, quickly and at no cost to estimate soil–geosynthetic interactions. In this regard well and poorly graded sands, anchors of three different size and anchorage lengths from the shear surface together with normal pressures of 12.5, 25 and 50 kPa were used. Artificial Intelligence (AI) called the Gene Expression Programming (GEP) was adopted to develop the model. Input variables included coefficients of curvature and uniformity, normal pressure, effective grain size, anchor base and surface area, anchorage length and the output variable was EIC. Contributions of input variables were evaluated using sensitivity analysis. Excellent correlation between the GEP-based model and the experimental results were achieved showing that the proposed model is well capable of effectively estimating soil–anchored geogrid enhanced interaction coefficient. Sensitivity analysis for parameter importance shows that the most influential variables are normal pressure (σ_n) and anchorage length (L) and the least effective parameters are average particle size (D₅₀) and anchor base area (A_b).

     

Hybrid finite-discrete element modelling of dynamic fracture and resultant fragment casting and muck-piling by rock blast

A state-of-the-art review is conducted to highlight the fracture mechanism in rock blast and advantages and limitations of various methods in modelling it. A hybrid finite-discrete element method (FEM-DEM) is implemented to simulate rock fracture and resultant fragment muck-piling in various blasting scenarios. The modelled crushed, cracked and long radial crack zones are compared with those in literatures to calibrate the hybrid FEM-DEM. Moreover, the hybrid modelling reproduces the rock fragmentation process during blasting. It is concluded that the hybrid FEM-DEM is superior to continuous and discontinuous methods in terms of modelling dynamic fracture of rock under blast-induced impact load.     

基于岩石颗粒排列方式的低渗透储层应力敏感性分析

以颗粒堆积模型为基础,考虑了低渗透岩心颗粒不同排列方式和不同变形方式,建立了毛管束模型,并通过颗粒Hertz接触变形原理对毛管变形量进行计算,研究毛管和多孔介质应力敏感性定量表征关系,通过有效毛管分数和毛管变形规律探讨了低渗透储层应力敏感性的作用机制。研究表明,低渗透储层的应力敏感性主要表现为渗透率的应力敏感性,相比于渗透率应力敏感性,孔隙度应力敏感性较弱;低渗透储层应力敏感性与岩石颗粒排列方式、颗粒变形方式、岩石微观孔隙结构、固液界面作用力和启动压力梯度效应等密切相关;考虑有效毛管分数和毛管变形量的多孔介质应力敏感性量化模型可从应力敏感性微观作用机制角度解释低渗透储层应力敏感性。     

Numerical determination of the effective permeability coefficient of soil–rock mixtures using the numerical manifold method

The present study extends the numerical manifold method to investigate the effective permeability coefficient (k_eff) of soil–rock mixtures. The influence of rock content, rock size, rock shape, and rock blocks' major axis direction on k_eff is studied. The results show the following: (1) k_eff decreases as the rock content increases; (2) the influence of rock size on k_eff can be neglected if other parameters are fixed; (3) the values of k_eff are nearly the same if rock blocks are in circular or regular hexagon shapes; and (4) the major axis direction of rock blocks has some influence on k_eff.