High Coefficient Of Determination Research Articles

A Novel Technique in Determining Mud Cake Permeability in SiO2 Nanoparticles and KCl Salt Water Based Drilling Fluid using Deep Learning Algorithm

The permeability of the mud cake formed at the formation-wellbore interface is an important factor in the designing of water-based drilling fluids. This study presents a novel approach to utilizing experimental thixotropic and rheological parameters of polymeric water-based drilling fluids having varying concentrations of SiO2 nanoparticles and KCl salt. A fully connected feed-forward multi-layered neural network, more commonly known as a Multilayer Perceptron (MLP) was developed to predict the mud cake permeability using input parameters such as SiO2 & KCl concentration, differential pressure, temperature, mud cake thickness, API LPLT and HPHT filter loss volume and spurt loss volume. The results suggested that the developed Multilayer Perceptron model effectively determined the mud cake permeability based on the input parameters of the WBDF mentioned above. The model converged on the global minima, minimizing the loss function using the Gradient descent algorithm. A higher Coefficient of Determination (R2) value i.e., 0.8781, and a lesser Root Mean Square Error (RMSE) value i.e., 0.04378 indicates the higher accuracy of the model. Pearson’s Coefficient of Correlation obtained via the heatmap indicates that mud cake permeability is strongly influenced by the differential pressure followed by filter loss volume, spurt loss volume, mud cake thickness, and temperature. Previous similar studies have focused on using machine learning algorithms, this study utilized a robust deep learning algorithm i.e., Multilayer Perceptron (MLP) neural network to simultaneously model the combined effects of SiO2 nanoparticles and KCl salt concentrations on mud cake permeability, offering an unprecedented level of accuracy in predicting key WBDF performance parameters

A novel spine tester TO GO.

Often after large animal experiments in spinal research, the question arises-histology or biomechanics? While biomechanics are essential for informed decisions on the functionality of the therapy being studied, scientists often choose histological analysis alone. For biomechanical testing, for example, flexibility, specimens must be shipped to institutions with special testing equipment, as spine testers are complex and immobile. The specimens must usually be shipped frozen, and, thus, biological and histological investigations are not possible anymore. To allow both biomechanical and biological investigations with the same specimen and, thus, to reduce the number of required animals, the aim of the study was to develop a spine tester that can be shipped worldwide to test on-site. The "Spine Tester TO GO" was designed consisting of a frame with three motors that initiate pure moments and rotate the specimen in three motion planes. A load cell and an optical motion tracking system controlled the applied loads and measured range of motion (ROM) and neutral zone (NZ). As a proof of concept, the new machine was validated and compared under real experimental conditions with an existing testing machine already validated employing fresh bovine tail discs CY34 (n = 10). The new spine tester measured reasonable ROM and NZ from hysteresis curves, and the ROM of the two testing machines formed a high coefficient of determination R 2 = 0.986. However, higher ROM results of the new testing machine might be explained by the lower friction of the air bearings, which allowed more translational motion. The spine tester TO GO now opens up new opportunities for on-site flexibility tests and contributes hereby to the 3R principle by limiting the number of experimental animals needed to obtain full characterization of spine units at the macroscopic, biomechanical, biochemical, and histological level.

Comparative analysis of cloud properties over drought- and flood-prone regions of western India using machine learning techniques

ABSTRACT Cloud properties are pivotal in analyzing rainfall patterns globally, especially in monsoon-dependent countries such as India. The impact of climate change becomes more important in regions susceptible to hydrometeorological events due to different monsoon regimes. To examine regional heterogeneity of cloud properties, this study investigates long-term trends and predictive capabilities for cloud properties in drought-prone and flood-prone regions of western India, utilizing satellite data and employing various machine learning (ML) models to comprehend intricate data patterns and enhance predictive accuracy. The results show higher mean and variability in cloud parameters over the flood-prone area due to favorable rain conditions, reflecting higher cloud microphysical and optical properties. These parameters negatively correlate with some cloud macrophysical properties along with the aerosol property in the drought-prone area. Additionally, a moderate correlation exists between certain cloud characteristics of one region and another. Employing ML algorithms for regression analysis and comparing them for cloud effective radius across regions shows promising results, with random forest (RF) demonstrating high coefficient of determination (0.86, 0.93) and low root mean squared error (0.76, 1.15) due to its robustness and high accuracy. This research enhances the understanding of regional heterogeneity in India and shows that ML can be helpful in predicting future cloud dynamics and climate variables identifying the most suitable model.

LJF of thin-walled gapped uni-planar K-type tubular steel joints with collar under brace balanced axial loads

This paper investigates the efficacy of the collar plate in improving the Local Joint Flexibility (LJF) of thin-walled circular hollow section (CHS) K-joints under axial loads. Initially, a finite element model (FEM) was created and verified against data from 14 available experimental tests. Subsequently, an extensive series of 136 unreinforced and reinforced K-joints was simulated to analyze the influence of parameters such as collar plate dimensions (η and λ), brace angle (θ), and joint geometry (β, ξ, and γ) on the LJF factor (fLJF) and the fLJF ratio (χ). The FEMs incorporated plate-to-member contact and weld modeling. Findings indicate that the utilization of the plate reduces the fLJF by 73 %. Additionally, it was observed that plate length significantly affects the fLJF compared to plate thickness. While the impact of γ, θ, and ξ on χ is marginal, β, η and λ emerge as a noteworthy determinant of χ. Despite the pivotal role of fLJF in joint behavior, there exists a gap in studies and equations specifically addressing fLJF in collar-plate reinforced K-joints under axial loads. To address this gap, leveraging the FE results, a design equation is proposed. This equation is subsequently validated with high coefficients of determination and standards set by the UK Department of Energy.

Reference Ranges and Z-Score Equations for 19 Fetal Cardiac Biometry Structures From 18 to 34 Weeks' Gestation.

To determine equations for calculating the Z-scores of fetal cardiac structures between 18+0 and 34+6 weeks of gestation, create percentile reference tables and curves for the structures, and assess the intra- and inter-observer reproducibility of the measurements. A cross-sectional study was conducted involving 340 normal fetuses from singleton pregnancies between 18 and 34 weeks of gestational age (GA). Nineteen cardiac structures were evaluated: diameters of the mitral, tricuspid, aortic, and pulmonary valve annuli; length, diameter, and area of the left and right ventricles; cardiac area and circumference; and diameters of the ascending aorta, aortic isthmus, main pulmonary artery, right pulmonary artery, left pulmonary artery, and ductus arteriosus. Regression analysis was performed to determine the equations for the mean and standard deviation of all structures using GA, biparietal diameter (BPD), and femur length (FL) as independent variables. All equations had high coefficients of determination (R2). The best performance was achieved using the GA (R2 .819-.944), followed by FL (R2 .813-.937) and BPD (R2 .792-.934). The structure that demonstrated the highest R2 was the cardiac circumference and the smallest was the ductus arteriosus. Reference tables of percentiles 1, 5, 10, 50, 90, 95, and 99, and reference curves of Z-scores were created for all 19 cardiac structures, depending on the GA. All measurements demonstrated good and excellent reproducibility with an inter-observer intraclass correlation coefficient (ICC) of 0.774-0.972 and intra-observer ICC of 0.938-0.993. Equations were produced to calculate Z-scores as well as percentile tables and curves for 19 fetal heart structures. All the measurements demonstrated good reproducibility.

Evaluation of some anions in groundwater in Riyadh, Saudi Arabia, and human health risk assessment of nitrate and fluoride.

Groundwater is a vital source of water for human and agricultural use in many parts of the world. The purpose of this research was to establish the quality of groundwater in Riyadh, Saudi Arabia, as well as the human health concerns associated with it. We collected and examined groundwater samples for pH, EC, TDS, CaCO3, fluoride (F-), chloride (Cl-), sulfate (SO42-), and nitrate (NO3-). The ion chromatography conductometric detection method was constructed to determine fluoride, chloride, sulfate, and nitrate in groundwater. The suggested method worked well for the anions that were being studied; it had a high coefficient of determination (r2 > 0.998) and average recoveries for all analytes that were between 97.5% and 99.0%, with a range of error of 0.77 to 2.37%. Fluoride concentrations were detected between 0.001 and 0.14mg/L, which are within the acceptable limit by several organizations. Chloride was measured in the range of 17.1 to 966.5mg/L, with some samples above the limits. The influence on sulfate ranged from 2.0 to 1136.0mg/L, with several samples exceeding the limits. In contrast, with nitrate levels ranging from 1.4 to 5.0mg/L, the majority of the samples fall within the acceptable range. The overall intake of fluoride, chloride, sulfate, and nitrate is 0.00605, 138.911, 65.515, and 1.19, respectively, which is lower than the recommended daily consumption except for chloride. The groundwater sample contains fluoride and nitrate with HQ values less than one: 0.000064-0.0641 and 0.033654-0.120192. Humans in Riyadh, Saudi Arabia, do not pose a health risk when digesting or absorbing groundwater fluoride or nitrate.

Harnessing machine learning for accurate estimation of compressive strength of high-performance self-compacting concrete from non-destructive tests: A comparative study

This study investigates the compressive strength (CS-28d) of high-performance self-compacting concrete. To accurately estimate the CS-28d, the study employed regression techniques (linear regression (LR) and stepwise polynomial regression (SPR)) as well as machine learning techniques (M5Prime, Random Forest, Random Tree and REP Tree). The study was based on 600 experimental results were collected from literature, with the rebound number (R), and ultrasonic pulse velocity (Vp) used to predict the CS-28d. The results indicate that the M5Prime model performed the best, with the highest coefficient of determination (Pearson R of 95.06 %) and the lowest root mean square error (RMSE of 4.84516 %). Among the tested regression models, the SPR model demonstrated the best performance and an empirical equation was introduced to estimate the CS-28d of a high-performance self-compacting concrete with high accuracy. An in-depth exploration of the model's sensitivity to its input parameters was conducted to quantify their effect on the CS-28d outcome. The results showed that the rebound number (R) had the most significant impact, with a sensitivity analysis parameter of 95 %. This was followed by the ultrasonic pulse velocity (Vp) had a relatively insignificant effect of only 5 % on the CS-28d values.

Mass Timber Buildings: The associated risks of rainwater exposure during construction in the Portuguese climate

The increase in the number and complexity of wooden buildings creates a series of challenges, like moisture control and management during the construction phase. The risks associated with the direct exposure of timber elements to rain during construction depend on the severity of the climate and the specific details of each project and therefore must be analysed locally. In this context, the results of 16 weeks of moisture content monitorization of Cross Laminated Timber (CLT) floor-wall connection specimens exposed to the outdoors are presented and evaluated. The monitorization demonstrated that end-grain surfaces are the most sensitive to the action and movement of water in CLT and within a few days of exposure to rain reached alarming moisture content (MC>30 %), with water entrapment at the interface between elements, preventing drying in a reasonable time for construction workflow (MC>30 % after 39 days of drying). On the other hand, zones with faster drying exhibited severe physical damage such as delamination and cracks. Additionally, the study evaluates the feasibility of using satellite-obtained meteorological data to develop a degradation risk map (Scheffer Climate Index) for mainland Portugal. The high coefficient of determination (R2 ≥ 0.81) and model efficiency (EF ~ 0.80–0.90) demonstrate the reliability of the data. Finally, two degradation risk maps for mainland Portugal are presented.

Real-time in-situ coatings corrosion monitoring using machine learning-enhanced triboelectric nanogenerator

Current methods for monitoring coating corrosion are limited by their inability to provide real-time data and dependence on external power sources. This study presents a novel in-situ corrosion monitoring system using a solid-liquid triboelectric nanogenerator (TENG) that converts mechanical energy into electrical signals for self-powered sensing. TENG signals and electrochemical impedance spectra were measured on a dopamine-modified lignin-polydimethylsiloxane coating on steel in 1 M NaCl solution under no corrosion, indentation, pitting, and broken conditions, respectively. We extract time-frequency features from the TENG signals to predict the coating’s corrosion condition by applying a customised convolutional neural network (CNN). By extracting time-frequency features from the TENG signals and applying a custom CNN, a prediction accuracy of 99 % for corrosion classification was achieved. Furthermore, the CNN regression model predicted coating impedance values with a high coefficient of determination (R² = 0.98), demonstrating its effectiveness in tracking corrosion progression. The developed TENG also facilitates defect localisation via a matrix electrode beneath the coating. Our approach introduces a promising real-time technology for in-situ corrosion monitoring.

Determination of nutritional sufficiency ranges for pomelo (Citrus grandis Osbeck) grown on alluvial soils using DRIS.

Pomelo is an important tropical fruit with a high nutrient content and economic value in the Vietnamese Mekong Delta (VMD). The Diagnosis and Recommendation Integrated System (DRIS) helps determine the leaf nutrient status of various plants worldwide. However, the DRIS-based nutritional balance in pomelo leaves remains to be established. Therefore, in this study, we aimed to (i) construct the DRIS norms and indices for nutrients, including macronutrients (N, P, K, Ca, and Mg) and trace elements (Cu, Fe, Zn, and Mn) in pomelo leaves, and (ii) establish nutrient sufficiency value ranges for sustainable pomelo cultivation in the VMD. We collected 270 leaf samples at three stages of pomelo growth, i.e., flowering, fruit development, and postharvest, and calculated DRIS indices for various nutrients. The DRIS indices established for various nutrients in pomelo leaves were accurate and reliable, as indicated by the high coefficient of determination (R2 = 0.43-0.93, p < 0.05) between nutrient concentrations and their DRIS indices. We observed that pomelo leaves were deficient in N (IN = -6.82), P (IP = -24.0), and Fe (IFe = -0.40) at the flowering stage and most deficient in P (IP = -15.6), K (IK = -11.7), Fe (IFe = -0.50), and Mn (IMn = -2.31) at the fruit development stage. However, only N (IN = -2.64) and P (IP = -13.4) shortages were observed at the postharvest stage. Thus, in this study, we evaluated nutrient value ranges (deficient, balanced, and excess) in pomelo leaves at their different growth stages and established DRIS indices for various nutrients. The results contribute to our understanding of the nutritional status of pomelo leaves, which can help growers improve plant health for sustainable pomelo production.

Exploiting RSM and ANN modeling methods to optimize phosphate ions removal using LDH/alginate composite beads

In the present work, layered double hydroxide (MgAl) and alginate composite beads (LDH/alginate) were developed and used as low-cost and environmentally friendly adsorbent for phosphate ions removal. The successful incorporation of sodium alginate into the LDH structure was confirmed through SEM analysis. The dried beads exhibited a notably rough surface, which promotes molecular movement and potentially enhances pollutant adsorption. The pH of zero charge point (pHPZC) of the composite was found to be 7.41. This result implies that negative ions have a tendency to draw positive charges on the adsorbent surface via electrostatic interaction forces when the pH is lower than the pHPZC. On the other hand, a surface that has a pH higher than pHPZC mostly has a negative charge. The percentage removal and adsorption capacity were investigated as a function of contact time.Experiments were carried out by simultaneously varying three factors, in order to evaluate and compare the predictive capabilities of the response surface methodology (RSM) and the artificial neural network approach (ANN) for the adsorption process. Both methods demonstrated a strong ability to accurately predict the adsorption process. However, the response surface methodology exhibited a lower prediction error compared to the artificial neural network approach.The central composite design within response surface methodology (CCD-RSM) was employed to optimize the experimental conditions for the adsorption process. The model, which considers three factors: adsorbent dose (A), initial concentration (B), and contact time (C), proved to be significant. Among these, the quadratic term (B2) had the most substantial impact on the phosphate ion adsorption rate. The analysis yielded an R2 value of 0.94, indicating an excellent fit to the data. The findings suggest that increasing contact time and reducing the initial concentration improve phosphate ion removal efficiency, while the adsorbent dose has little to no effect. The Artificial Neural Network (ANN) model effectively predicted the adsorptive remediation of phosphate ions onto LDH/alginate composite beads, achieving a high coefficient of determination (R2 = 0.984) between the model outputs and the experimental data. The study highlights the significant potential of LDH/alginate composite beads as a natural adsorbent for the removal of phosphate ions from aqueous solutions. These results underscore the environmental friendliness and efficiency of the adsorbent used for phosphate adsorption.

Optimization and Validation of Extrusion Process Parameters for the Sensory Characteristics of Extruded Aerial Yam and Soybean Flour Blends

The aim of this study was to optimize and validate the extrusion process parameters (barrel temperature, screw speed, and feed moisture) for the sensory properties (texture, taste, appearance, and aroma) of extrudates made from blends of soybean and aerial yam flours. Five levels of barrel temperature (95, 100, 105, 110, and 115 °C), screw speed (85, 100, 115, 130, and 145 rpm), and feed moisture (31, 33, 35, 37, and 39%) were employed in 20 runs of the response surface methodology (RSM), which was based on the Box-Behnken design with three variables. A single-screw extruder at the laboratory scale was used to carry out the extrusion procedure. A high regression coefficient (R2 ≥ 0.9) indicates that the models are useful for navigating the design space. Numerical optimization results indicated that the optimal extrusion process parameters - barrel temperature of 114.12 °C, screw speed of 100.56 rpm, and feed moisture of 38.02% - produced extrudates with optimal sensory property scores of 5.34 for texture, 4.91 for taste, 6.97 for appearance, and 5.80 for aroma, with a desirability of 0.943. The correlation between the predicted and experimental values yielded a high coefficient of determination, indicating a good correlation. The “Fit and Diagnostic Case” statistics showed a low range of deviations between the predicted and observed values for the sensory characteristics. Therefore, the generated quadratic model accurately predicts the sensory characteristics of aerial yam-soybean flour blends and is thus validated.

Retrieval of Solar Shortwave Irradiance from All-Sky Camera Images

The present work proposes a new model based on a convolutional neural network (CNN) to retrieve solar shortwave (SW) irradiance via the estimation of the cloud modification factor (CMF) from daytime sky images captured by all-sky cameras; this model is named CNN-CMF. To this end, a total of 237,669 sky images paired with SW irradiance measurements obtained by using pyranometers were selected at the following three sites: Valladolid and Izaña, Spain, and Lindenberg, Germany. This dataset was randomly split into training and testing sets, with the latter excluded from the training model in order to validate it using the same locations. Subsequently, the test dataset was compared with the corresponding SW irradiance measurements obtained by the pyranometers in scatter density plots. The linear fit shows a high determination coefficient (R2) of 0.99. Statistical analyses based on the mean bias error (MBE) values and the standard deviation (SD) of the SW irradiance differences yield results close to −2% and 9%, respectively. The MBE indicates a slight underestimation of the CNN-CMF model compared to the measurement values. After its validation, model performance was evaluated at the Antarctic station of Marambio (Argentina), a location not used in the training process. A similar comparison between the model-predicted SW irradiance and pyranometer measurements yielded R2=0.95, with an MBE of around 2% and an SD of approximately 26%. Although the precision provided by the SD at the Marambio station is lower, the MBE shows that the model’s accuracy is similar to previous results but with a slight overestimation of the SW irradiance. Finally, the determination coefficient improved to 0.99, and the MBE and SD are about 3% and 11%, respectively, when the CNN-CMF model is used to estimate daily SW irradiation values.

Defining the segmental tension generated in a vertebral body tethering system for scoliosis.

Vertebral body tethering (VBT) uses a flexible tether affixed across the curve convexity with tension applied at each segment to treat scoliosis. Intraoperative tether tension may be achieved directly with a counter-tensioner or with an extension spring tube. The purpose of this study was to quantify the force generated with and without the extension spring tube using current FDA-approved VBT instrumentation, to understand the variation between surgeons using the same instrumentation, and to define the force range that is generated intra-operatively. Using a benchtop mechanical testing setup to simulate a spinal segment, we affixed the tether and applied tension using a tensioner and counter-tensioner alone (method T1) or by adding an extension spring tube (method T2). Eight orthopedic surgeons used T1 and T2 at six tensioner settings, and one surgeon completed three trials. A two-way ANOVA with a Tukey's HSD post hoc test (p < 0.05) compared the tensioner methods and testing levels. Inter- and intra-rater reliabilities were calculated using intraclass correlation coefficients (ICCs). Methods T1 and T2 exhibited linear tension-setting relationships, with high determination coefficients (R2 > 0.93). T2 consistently produced higher forces (increase of 62.1 N/setting), compared to T1 (increase of 50.6 N/setting, p < 0.05). Inter-rater reliability exhibited excellent agreement (ICC = 0.951 and 0.943 for T1 and T2, respectively), as did intra-rater reliability (ICC = 0.971).

Neuro-Musculoskeletal Modeling for Online Estimation of Continuous Wrist Movements from Motor Unit Activities.

Decoding movement intentions from motor unit (MU) activities remains an ongoing challenge, which restricts our comprehension of the intricate transition mechanism from microscopic neural drive to macroscopic movements. This study presents an innovative neuro-musculoskeletal (NMS) model driven by MU activities for online estimation of continuous wrist movements. The proposed model employs a physiological and comprehensive utilization of MU firings and waveforms, thus facilitating the localization of MUs to muscle-tendon units (MTU) as well as the computation of MU-specific neural excitation. Subsequently, the MU-specific neural excitation was integrated to form the MTU-specific neural excitation, which were then inputted into a musculoskeletal model to accomplish the joint angle estimation. To assess the effectiveness of this model, high-density surface electromyography and angular data were collected from the forearms of eight subjects during their performance of wrist flexion-extension task. Two pieces of 8 × 8 electrode arrays and a motion capture system were employed for data acquisition. Following offline model calibration with a global optimization algorithm, online angle estimation results demonstrated a significant superiority of the proposed model over the state-of-the-art NMS models (p < 0.05), yielding the lowest normalized root mean square error (0.10 ± 0.02) and the highest determination coefficient (0.87 ± 0.06). This study provides a novel idea for the decoding of joint movements from MU activities. The research findings hold the potential to advance the development of NMS models towards the control of multiple degrees of freedom, with promising applications in the fields of motor control, biomechanics, and neuro-rehabilitation engineering.

Relationships between body length, body mass and otolith dimensions in three Sprat species (Teleostei: Ehiravidae) from the Caspian Sea

AbstractReliable and detailed knowledge regarding the relationship between otolith size and the size of fish is important for fisheries management and for ecological studies on predicting fish size and on predator–prey interactions. Therefore, these relationships are estimated for three sprat (genus Clupeonella) species, including C. caspia, C. engrauliformis and C. grimmi from the Caspian Sea. The relative size and mass of the otoliths corrected for body size and mass were also estimated for the three sprat species. After biometry of collected specimens, the Sagittal otoliths were extracted from the cranium of collected specimens and photographed. Photos were used to estimate length and width of otoliths. The results revealed significant relationships between otolith length and width vs. fish length and otolith mass vs. body mass in all three species (r2 &gt; 0.5). The higher coefficient of determination (r2) for relationships between total length (fish) – Otolith length and width was concluded in comparison with relationships between fish mass and otolith mass in the studied sprat species. Otoliths of C. grimmi showed the highest relative length and were significantly (P &lt; 0.05) heavier than sagittal otoliths of two other sprat species. There is no previous report on size–mass relationships between fish and otolith measurements among the three studied sprat species. These estimated equations can be used in back-calculation studies, especially for these three sprats as the dominant prey for piscivorous predators such as Caspian seals and sturgeons, in their habitats.

Unveiling the Genetic Architecture of Semen Traits in Thai Native Roosters: A Comprehensive Analysis Using Random Regression and Spline Function Models.

Improving reproductive traits, particularly semen quality and quantity, is crucial for optimizing poultry production and addressing the current limitations in native chicken reproduction. The aim of this study was to develop a genetic model to estimate genetic parameters guiding the selection of individual Thai native roosters. Using data collected from 3475 records of 242 Thai native grandparent roosters aged 1-4 years, we evaluated semen traits (mass movement, semen volume, and sperm concentration) over 54 weeks. A random regression test-day model incorporating five covariance functions, including a linear spline function with four, five, six, and eight knots (SP4, SP5, SP6, and SP8) and second-order Legendre polynomial function (LG2), was used to estimate genetic parameters. The results showed that the SP8 model consistently outperformed the other models across all traits, with the lowest mean square error, highest coefficient of determination, and superior predictive ability. Heritability estimates for mass movement, semen volume, and sperm concentration ranged from 0.10 to 0.25, 0.22 to 0.25, and 0.11 to 0.24, respectively, indicating moderate genetic influence on these traits. Genetic correlations between semen volume and sperm concentration were highest in the SP8 model, highlighting a strong genetic association between these traits. The SP8 model also revealed a high genetic correlation between mass movement and semen volume, supporting the potential for selecting mass movement as a predictor of semen volume. In conclusion, this study highlights the effectiveness of random regression models with linear spline functions to evaluate the genetic parameters of semen traits in native Thai roosters. The SP8 model is a robust tool for breeders to enhance the reproductive performance of native Thai chickens, contributing to sustainable poultry production systems.

Using convolutional neural network combined with multi-scale channel attention module to predict soil properties from visible and near-infrared spectral data

With the increasing demand for precision agriculture and sustainable land management, it is crucial to quickly and accurately predict soil properties. However, there are still challenges in predicting large-scale soil properties. This study proposes a new convolutional neural network architecture (MCA-CNN) combined with visible and near-infrared spectroscopy for predicting soil properties. The architecture introduces a multi-scale channel attention mechanism, aiming to more effectively capture complex features in the spectrum. The performance of the MCA-CNN model was tested on a large-scale dataset and compared with partial least squares (PLS), extreme gradient boosting tree (XGBoost), and conventional convolutional neural network (CNN) models. The predictive performance of different models in organic carbon (OC), nitrogen (N), pH, Clay, Salt, and Sand was analyzed. The results showed that the MCA-CNN model had the best predictive performance on all soil properties, achieving the lowest root mean square error (RMSE) and the highest coefficient of determination (R2), the R2 of OC, N, pH, Clay, Silt, and Sand are 0.95, 0.93, 0.95, 0.80, 0.58, 0.64, respectively. The RMSE of OC, N, pH, Clay, Silt, and Sand are 16.60 g kg−1, 0.96 g kg−1, 0.31, 4.89 %, 8.27 %, and 11.61 %, respectively. Further research has found that soil types (mineral soil and organic soil) affect the predictive performance of OC. It is necessary to distinguish between mineral soil and organic soil when predicting soil OC content. In addition, the effectiveness of the model transfer strategy was explored for situations with small soil sample sizes. By fine-tuning the pre-trained model, the performance of the model can be significantly improved, which provides a feasible solution for predicting soil properties in situations of data scarcity. In summary, this study not only confirms the potential application of deep learning in the field of soil science, but also provides new technological avenues for future soil management and agricultural practices.

A comprehensive multivariate approach for GxE interaction analysis in early maturing rice varieties.

The genotype evaluation process requires analysis of GxE interactions to ascertain the responsiveness of a genotype to various environments, including the development of early maturing rice. However, the concept of interaction is relatively specific to grain yield. In contrast, grain yield is highly polygenic, so assessment should be carried out with multivariate approaches. Therefore, multivariate assessment in evaluating GxE interactions should be developed, especially for early maturing rice genotypes. The study aimed to develop a comprehensive multivariate approach to improve the comprehensiveness and responsiveness of GxE interaction analysis. The study was conducted in Bone and Soppeng districts, South Sulawesi, Indonesia, in two seasons. The study used a randomized complete block design, where replications were nested across two seasons and locations. Two check varieties and five early maturing varieties were replicated three times in each environment. Based on this study, a new approach to GxE interaction analysis based on multiple regression index analysis, BLUP analysis, factor analysis, and path analysis was considered adequate, especially for evaluating early maturing rice. This approach combined days to harvest, biological yield, and grain yield in multiple linear regression with weighting based on the combination of all analyses. The effectiveness of the GxE interaction assessment was reflected by high coefficient of determination (R2) and gradient (b) values above 0.8 and 0.9, respectively. Inpari 13 (R2=0.9; b=1.05), Cakrabuana (R2=0.98; b=0.99), and Padjajaran (R2=0.95; b=1.07) also have good grain yield with days to harvesting consideration, namely 7.83ton ha-1, 98.12 days; 7.37ton ha-1, 95.52 days; and 7.29ton ha-1, 97.23 days, respectively. Therefore, this index approach can be recommended in GxE interaction analysis to evaluate early maturing rice genotypes. Furthermore, Inpari 13, Cakrabuana, and Padjajaran are recommended as adaptive early maturing varieties.

A dynamic solvent chamber propagation estimation framework using RNN for warm solvent injection in heterogeneous reservoirs

Warm solvent injection (WSI), injecting low-temperature solvent into formations to reduce the viscosity of heavy oil, is a clean technology for heavy oil production through reducing greenhouse gas emissions and water usage. The success of WSI operation depends on the uniform development and propagation of solvent chambers in reservoirs. However, reservoir heterogeneity stemming from shale barriers plays a detrimental role in the conformance of solvent chamber development and oil production rate. In this work, we developed a novel recurrent neural network (RNN)-based framework with the capability of efficiently tracking and estimating the solvent chamber positions in heterogeneous reservoirs based on only production time-series data. The developed estimation model utilizes the “sequence-to-sequence" mapping methodology to correlate observed production time-series sequence and solvent chamber edge sequence via a long short-term memory (LSTM) algorithm. The trained RNN models exhibit high accuracy, evidenced by the predicted dynamic solvent chamber locations match the corresponding true locations from numerical simulation, with a high coefficient of determination (R2) and a low mean squared error. Specifically, the achieved R2 values exceed 0.98 on both the training and testing data. The developed RNN-based workflow was tested via several cases from both regularly- and irregularly-shaped shale barriers, and the results were promising. The predicted solvent chambers showed strong agreement with those obtained from numerical simulations. The major benefits of this workflow include reducing computational time and saving overall monitoring and tracking costs for conventional techniques. The present work would provide a good demonstration of the capability of practical integration of machine learning methods in solving engineering problems.