Comparative analysis of seven machine learning algorithms and five empirical models to estimate soil thermal conductivity

Zhao, Tianyue; Liu, Shuchao; Xu, Junzeng; He, Hailong; Wang, Dong; Horton, Robert; Liu, Gang

doi:10.1016/j.agrformet.2022.109080

Cited by 31 publications

(20 citation statements)

References 61 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As shown in Fig. 6 a, most of the symbols in the model are located below the 1:1 line, implying that the calculated values are smaller than the measured value, which is consistent with Zhao et al 51 . The Li et al (2008) model (RMSE/PBIAS/NSE = 0.66/0.60/− 0.32) and the Tien et al (2005) model (RMSE/PBIAS/NSE = 0.48/0.30/0.30) modified from the Campbell (1985) model mainly developed for fine-textured soils and frozen soils perform the worst that their model calculations severely underestimated the actual soil thermal conductivity based on results in Fig.…”

Section: Resultssupporting

confidence: 89%

See 1 more Smart Citation

A new model to predict soil thermal conductivity

Xiong

Feng

Jin

et al. 2023

Sci Rep

View full text Add to dashboard Cite

Thermal conductivity is a basic parameter of soil heat transferring, playing an important role in many fields including groundwater withdrawal, ground source heat pump, and heat storage in soils. However, it usually requires a lot of time and efforts to obtain soil thermal conductivity. To conveniently obtain accurate soil thermal conductivity, a new model describes the relationship between soil thermal conductivity (λ) and degree of saturation (Sr) was proposed in this study. Dry soil thermal conductivity (λdry) and saturated soil thermal conductivity (λsat) were described using a linear expression and a geometric mean model, respectively. A quadratic function with one constant was added to calculate λ beyond the lower λdry and upper λsat limit conditions. The proposed model is compared with five other frequently used models and measured data for 51 soil samples ranging from sand to silty clay loam. Results show that the proposed model match the measured data well. The proposed model can be used to determine soil thermal conductivity of a variety of soil textures over a wide range of water content.

show abstract

Section: Resultssupporting

confidence: 89%

“…Previous studies show that values calculated by the Campbell (1985) model is much smaller than the measured values 42 – 44 , 51 . Enhancing the Campbell (1985) model calculations requires calibration for different soils to obtain the model parameters which increases the complexity of the model.…”

Section: Methodsmentioning

confidence: 77%

A new model to predict soil thermal conductivity

Xiong

Feng

Jin

et al. 2023

Sci Rep

View full text Add to dashboard Cite

show abstract

“…[ 71 ]). The feature importance analysis scores of RF are also shown to be moderately more accurate than those of GBDT [ 72 , 73 ].…”

Section: Resultsmentioning

confidence: 99%

Enhanced wave overtopping simulation at vertical breakwaters using machine learning algorithms

2023

View full text Add to dashboard Cite

Accurate prediction of wave overtopping at sea defences remains central to the protection of lives, livelihoods, and infrastructural assets in coastal zones. In addressing the increased risks of rising sea levels and more frequent storm surges, robust assessment and prediction methods for overtopping prediction are increasingly important. Methods for predicting overtopping have typically relied on empirical relations based on physical modelling and numerical simulation data. In recent years, with advances in computational efficiency, data-driven techniques including advanced Machine Learning (ML) methods have become more readily applicable. However, the methodological appropriateness and performance evaluation of ML techniques for predicting wave overtopping at vertical seawalls has not been extensively studied. This study examines the predictive performance of four ML techniques, namely Random Forest (RF), Gradient Boosted Decision Trees (GBDT), Support Vector Machines—Regression (SVR), and Artificial Neural Network (ANN) for overtopping discharge at vertical seawalls. The ML models are developed using data from the EurOtop (2018) database. Hyperparameter tuning is performed to curtail algorithms to the intrinsic features of the dataset. Feature Transformation and advanced Feature Selection methods are adopted to reduce data redundancy and overfitting. Comprehensive statistical analysis shows superior performance of the RF method, followed in turn by the GBDT, SVR, and ANN models, respectively. In addition to this, Decision Tree (DT) based methods such as GBDT and RF are shown to be more computationally efficient than SVR and ANN, with GBDT performing simulations more rapidly that other methods. This study shows that ML approaches can be adopted as a reliable and computationally effective method for evaluating wave overtopping at vertical seawalls across a wide range of hydrodynamic and structural conditions.

show abstract

“…When the target temperature was lower than 350 °C, the oil content of the OBDC residue decreased significantly, and the oil content declined gradually after 350 °C, which is consistent with the results of the thermogravimetric experiment. In addition, with the continuous volatilization of the liquid phase of OBDC, the thermal conductivity of the oil-containing solid phase also decreased, , which caused the oil content of the residue to stabilize.…”

Section: Resultsmentioning

confidence: 99%

Study on Pyrolysis of Shale Gas Oil-Based Drilling Cuttings: Kinetics, Process Parameters, and Product Yield

Liu

Xiao

Dai³

et al. 2023

ACS Omega

View full text Add to dashboard Cite

The main reaction range (350–550 °C) of oil-based drilling cutting (OBDC) pyrolysis was studied by a thermogravimetric analyzer and a vacuum tube furnace. The average activation energies calculated by four model-free methods were 185.5 kJ/mol (FM), 184.16 kJ/mol (FWO), 166.17 kJ/mol (KAS), and 176.03 kJ/mol (Starink). The reaction mechanism was predicted by the Criado (Z-master plot) method. It is found that a high heating rate is helpful to predict the reaction mechanism, but it cannot be described by a single reaction model. Under the conditions of target temperature higher than 350 °C, residence time higher than 50 min, laying thickness less than 20 mm, and heating rate lower than 15 °C, the residual oil content is lower than 0.3% and the recovery rate of mineral oil is higher than 98.43%. Solid phase products accounted for more than 70%, reached the maximum 17.04% at 450 °C, and then decreased to 15.87% at 500 °C. Aromatic hydrocarbons, as coking precursors, are transformed from a low ring to a high ring. Recycled mineral oil can reconfigure oil-based mud (OBM). The research results can provide a theoretical basis for process optimization.

show abstract

Comparative analysis of seven machine learning algorithms and five empirical models to estimate soil thermal conductivity

Cited by 31 publications

References 61 publications

A new model to predict soil thermal conductivity

A new model to predict soil thermal conductivity

Enhanced wave overtopping simulation at vertical breakwaters using machine learning algorithms

Study on Pyrolysis of Shale Gas Oil-Based Drilling Cuttings: Kinetics, Process Parameters, and Product Yield

Contact Info

Product

Resources

About