Field Data Sets Research Articles

Analysis of factors influencing energy consumption of electric vehicles: Statistical, predictive, and causal perspectives

Analyzing the influence of various factors on the energy consumption of Electric Vehicles (EVs) yields a positive impact on mitigating range anxiety. Nevertheless, prevailing research predominantly relies on statistical or predictive methodologies to assess the influence of factors on energy consumption, offering limited causal insights. This study addresses this gap by utilizing the double/debiased machine learning (DML) combined with bootstrap-of-little-bags (BIB) inference approach to make causal inferences concerning the influence of factors on energy consumption. Furthermore, the study introduces a comprehensive framework that integrates statistical, predictive, and causal perspectives to examine the consistency and discrepancies in the influence of twelve factors on energy consumption. Notably, it unveils variations in the influence of these factors on energy consumption concerning different trip categories. Field datasets are utilized for framework validation, covering 20,385 valid trips from eight cities and four vehicle models. The findings demonstrate that DML can yield interpretable causal inferences, and the framework allows these three perspectives to complement each other, unveiling insights that remain concealed when using a single method. Furthermore, even for the same factor, there are variations in the influence on energy consumption across different trip categories. The proposed framework opens avenues for a comprehensive understanding of influencing factors on energy consumption of EVs, crucial for eco-driving and energy consumption predictions.

Variable gauge length: Processing theory and applications to distributed acoustic sensing

AbstractMost distributed acoustic sensing systems will only process acoustic data with a fixed gauge length for the entire well depth. However, it has been shown that the gauge length is a critical parameter to improve the signal‐to‐noise ratio when used as a function of certain geophysical parameters, such as the apparent velocity. It can also be responsible for significant distortions if tuned incorrectly. In this paper, we first aim to reintroduce the concept of gauge length and derive a robust method to optimize its value based on the geophysical parameters whilst ensuring no distortion to the original signal. We then present a novel method of processing the distributed acoustic sensing data using the concept of a variable gauge length. We finish by showing applications of these techniques on synthetic vertical seismic profiling data and some of the results obtained on actual field datasets.

Intelligent Prestack Multi-trace Seismic Inversion Constrained by Probabilistic Geological Information

Seismic inversion uses observed data including well-logging and seismic data to infer rock parameters. However, it still suffers from non-unique solutions. The non-uniqueness increase when extrapolating away from well location. To mitigate this effect, we incorporate geological information into a deep learning (DL) framework in the form of generated probabilistic labels. We transform geological information into reservoir heterogeneity (RH) weights and geological pattern constraint (GPC) weights to generate probabilistic labels. The RH weights are used to capture the variability in accuracy during the spatial extrapolation of well data. The RH weights are used to balance the constraint weights between well and seismic data in the DL model. The GPC weights are used to characterize the changes of geological patterns from well locations to non-well locations. Based on variations in local prestack seismic waveforms, the DL model learns extrapolation patterns of well data with similar geological patterns. By combining the rock parameters derived from well logs with these two types of weights, we can generate a series of probability labels. Two types of geological information and well-logging data are deeply integrated into a supervised 2D-ResUNet framework. In this way, we develop a novel DL-based prestack multi-trace seismic inversion method. We validate our method on a 3D synthetic model and a 3D field dataset. The results show that our method exhibits great potential for characterizing the spatial variation of subsurface rocks and outperforms traditional deep learning methods.

Greenland Wind-Wave Bivariate Dynamics by Gaidai Natural Hazard Spatiotemporal Evaluation Approach

The current work presents a case study for the state-of-the-art multimodal risk assessment approach, which is especially appropriate for environmental wind-wave dynamic systems that are either directly physically observed or numerically modeled. High dimensionality of the wind-wave environmental system and cross-correlations between its primary dimensions or components make it quite challenging for existing reliability methods. The primary goal of this investigation has been the application of a novel multivariate hazard assessment methodology to a combined windspeed and correlated wave-height unfiltered/raw dataset, which was recorded in 2024 by in situ NOAA buoy located southeast offshore of Greenland. Existing hazard/risk assessment methods are mostly limited to univariate or at most bivariate dynamic systems. It is well known that the interaction of windspeeds and corresponding wave heights results in a multimodal, nonstationary, and nonlinear dynamic environmental system with cross-correlated components. Alleged global warming may represent additional factor/covariate, affecting ocean windspeeds and related wave heights dynamics. Accurate hazard/risk assessment of in situ environmental systems is necessary for naval, marine, and offshore structures that operate within particular offshore/ocean zones of interest, susceptible to nonstationary ocean weather conditions. Benchmarking of the novel spatiotemporal multivariate reliability approach, which may efficiently extract relevant information from the underlying in situ field dataset, has been the primary objective of the current work. The proposed multimodal hazard/risk evaluation methodology presented in this study may assist designers and engineers to effectively assess in situ environmental and structural risks for multimodal, nonstationary, nonlinear ocean-driven wind-wave-related environmental/structural systems. The key result of the presented case study lies within the demonstration of the methodological superiority, compared to a popular bivariate copula reliability approach.

Advancing nitrogen nutrition index estimation in summer maize using continuous wavelet transform

Rapid and non-destructive diagnosis of plant nitrogen (N) status is crucial to optimize N management during the growth of summer maize. This study aimed to evaluate the effectiveness of continuous wavelet analysis (CWA) in estimating the nitrogen nutrition index (NNI), to determine the most suitable wavelet analysis method, and to identify the most sensitive wavelet features across the visible to near-infrared spectrum (325–1,025 nm) for accurate NNI estimation. Field experiments were conducted across two sites (Kaifeng and Weishi) during the 2022 and 2023 growing seasons using four summer maize cultivars (XD20, ZD958, DH661, and DH605) under varying N application rates (0, 80, 160, 240, and 320 kg N ha-1). Canopy reflectance spectra and plant samples were collected from the V6 to V12 growth stages. The wavelet features for each spectral band were calculated across different scales using the CWA method, and their relationships with NNI, plant dry matter (PDM), and plant N concentration (PNC) were analyzed using four regression models. The results showed that NNI varied from 0.61 to 1.19 across different N treatments during the V6 to V12 stages, and the Mexican Hat wavelet was identified as the most suitable mother wavelet, achieving an R² value of 0.73 for NNI assessment. The wavelet features derived from the Mexican Hat wavelet were effective in estimating NNI, PDM, and PNC under varying N treatments, with the most sensitive wavelet features identified as 745 nm at scale 7 for NNI, 819 nm at scale 5 for PDM, and 581 nm at scale 6 for PNC using linear regression models. The direct and indirect methods for NNI estimation were compared using independent field data sets. Both methods proved valid to predict NNI in summer maize, with relative root mean square errors of 10.8% for the direct method and 13.4% for the indirect method. The wavelet feature at 745 nm, scale 7, from the direct method (NNI = 0.14 WF (745 nm, 7) + 0.3) was found to be simpler and more accurate for NNI calculation. These findings provide new insights into the application of the CWA method for precise spectral estimation of plant N status in summer maize.

Multiple Data Imputation Methods Advance Risk Analysis and Treatability of Co-occurring Inorganic Chemicals in Groundwater.

Accurately assessing and managing risks associated with inorganic pollutants in groundwater is imperative. Historic water quality databases are often sparse due to rationale or financial budgets for sample collection and analysis, posing challenges in evaluating exposure or water treatment effectiveness. We utilized and compared two advanced multiple data imputation techniques, AMELIA and MICE algorithms, to fill gaps in sparse groundwater quality data sets. AMELIA outperformed MICE in handling missing values, as MICE tended to overestimate certain values, resulting in more outliers. Field data sets revealed that 75% to 80% of samples exhibited no co-occurring regulated pollutants surpassing MCL values, whereas imputed values showed only 15% to 55% of the samples posed no health risks. Imputed data unveiled a significant increase, ranging from 2 to 5 times, in the number of sampling locations predicted to potentially exceed health-based limits and identified samples where 2 to 6 co-occurring chemicals may occur and surpass health-based levels. Linking imputed data to sampling locations can pinpoint potential hotspots of elevated chemical levels and guide optimal resource allocation for additional field sampling and chemical analysis. With this approach, further analysis of complete data sets allows state agencies authorized to conduct groundwater monitoring, often with limited financial resources, to prioritize sampling locations and chemicals to be tested. Given existing data and time constraints, it is crucial to identify the most strategic use of the available resources to address data gaps effectively. This work establishes a framework to enhance the beneficial impact of funding groundwater data collection by reducing uncertainty in prioritizing future sampling locations and chemical analyses.

Large-Kernel Central Block Masked Convolution and Channel Attention-Based Reconstruction Network for Anomaly Detection of High-Resolution Hyperspectral Imagery

In recent years, the rapid advancement of drone technology has led to an increasing use of drones equipped with hyperspectral sensors for ground imaging. Hyperspectral data captured via drones offer significantly higher spatial resolution, but this also introduces more complex background details and larger target scales in high-resolution hyperspectral imagery (HRHSI), posing substantial challenges for hyperspectral anomaly detection (HAD). Mainstream reconstruction-based deep learning methods predominantly emphasize spatial local information in hyperspectral images (HSIs), relying on small spatial neighborhoods for reconstruction. As a result, large anomalous targets and background details are often well reconstructed, leading to poor anomaly detection performance, as these targets are not sufficiently distinguished from the background. To address these limitations, we propose a novel HAD network for HRHSI based on large-kernel central block masked convolution and channel attention, termed LKCMCA. Specifically, we first employ the pixel-shuffle technique to reduce the size of anomalous targets without losing image information. Next, we design a large-kernel central block masked convolution to make the network pay more attention to the surrounding background information, enabling better fusion of the information between adjacent bands. This, coupled with an efficient channel attention mechanism, allows the network to capture deeper spectral features, enhancing the reconstruction of the background while suppressing anomalous targets. Furthermore, we introduce an adaptive loss function by down-weighting anomalous pixels based on the mean absolute error. This loss function is specifically designed to suppress the reconstruction of potentially anomalous pixels during network training, allowing our model to be considered an excellent background reconstruction network. By leveraging reconstruction error, the model effectively highlights anomalous targets. Meanwhile, we produced four benchmark datasets specifically for HAD tasks using existing HRHSI data, addressing the current shortage of HRHSI datasets in the HAD field. Extensive experiments demonstrate that our LKCMCA method achieves superior detection performance, outperforming ten state-of-the-art HAD methods on all datasets.

Research on Flow Field Prediction in a Multi-Swirl Combustor Using Artificial Neural Network

In aero-engine combustion research, the pursuit of cost-effective and rapid methods for acquiring precise flow fields across various operating conditions remains a significant challenge. This study offers novel insights into the rapid modeling of complex multi-swirling flows, introducing flow-field-based analytical methods to evaluate flow topologies, spray dispersion, ignition dynamics, and flame propagation patterns. A data-driven model is proposed to predict the swirling velocity field inside a multi-swirl combustor, using spatial coordinates and air pressure drops as input features. Particle Image Velocimetry (PIV) experiments under different air pressure drops are performed to generate the necessary flow field dataset. A fully connected deep neural network is designed and optimized with a focus on prediction accuracy, training efficiency, and mitigation of over-fitting. The predicted flow characteristics, including swirling jets, shear layers, recirculation zones, and velocity profiles, align closely with the PIV experimental results. This demonstrates the model’s capability to effectively capture the intricate multi-swirling flow structures and the complex relationships between input parameters and the resulting flow field. Furthermore, the trained model shows excellent generalization capability, accurately predicting flow fields under previously unseen operating conditions. Finally, combustion-relevant characteristics, such as ignition and flame propagation, are successfully extracted and analyzed from the predicted flow fields using the proposed deep learning framework.

Predictive Modelling of Mineral Prospectivity Using Satellite Remote Sensing and Machine Learning Algorithms

In today’s world of falling returns on fixed exploration budgets, complex targets, and ever-increasing volumes of multi-parameter datasets, the effective management and integration of existing data are essential to any mineral exploration operation. Machine learning (ML) algorithms like Convolutional Neural Networks (CNN), Random Forest (RF), and Support Vector Machine (SVM) are powerful data-driven methods that are not implemented very often with remote sensing-derived hydrothermal alternation information and limited field datasets for mapping mineral prospectivity. The application of machine learning algorithms with satellite remote sensing data and limited field data, they have not been compared and evaluated together thoroughly in this field. A data science approach was applied to create nine predictor maps, incorporating limited field data and satellite remote sensing information. A confusion matrix, statistical measures, and a Receiver Operating Characteristic (ROC) curve were used to evaluate the prediction models efficacy on both the training and test datasets. The results suggested that the RF model exhibited the highest predictive accuracy, consistency and interpretability among the three ML models evaluated in this study. RF model also achieved the highest predictive efficiency in capturing known copper (Cu) deposits within a small prospective area. In comparison to the SVM and CNN models, the RF model outperformed them in terms of predictive accuracy and interpretability. These results imply that the RF model is the most suitable for Cu potential mapping in the Pakistan’s North Waziristan region. Consequently, all the models including the RF model were used to generate a prospectivity map, which contained low to very-high potential zones, to support further exploration in the region. The newly discovered deposit inside the predicted prospective areas demonstrates the robustness and efficacy of the prospectivity modelling approach as proposed in this research for generating exploration targets.

Advancing mechanistic understanding of cohesive sediment transport: Integrating flume experiments, field measurements, and modelling approaches in a gravel-bed river

Gravel-bed rivers draining mountainous forested headwater regions are critically important for drinking water supply and ecological integrity. These rivers, however, have been increasingly impacted by intensifying anthropogenic and natural (especially climate change exacerbated) landscape disturbances that commonly increase hillslope/channel connectivity and the delivery of cohesive sediment (<63 μm) and associated pollutants. Despite the known deleterious threats of excess cohesive sediments, there is still limited understanding of their transport and intra-gravel storage due to the complexities of such processes. Accordingly, the objectives of this study were to: i) calibrate and validate a semi-empirical cohesive sediment transport model (RIVFLOC) using the observations from flume experiments; ii) estimate the intra-gravel storage capacity for cohesive sediment with the calibrated model based on the field dataset (collected in two field campaigns between 2019 and 2021), and; iii) investigate mechanisms of cohesive sediment transport dynamics in this gravel-bed river, identifying knowledge gaps and areas for future research. Our results showed that despite the increased floc settling velocity, deposition was hindered by turbulent flow fields. The model predicted that ~60 % of upstream cohesive sediment would ingress within the 10 km study reach due to the flow interaction with the gravel-bed. Despite the agreement between flume and field observations on ingress rates and preferential ingress of coarser (~100 μm) flocs, notable differences were observed between modelled and field datasets, highlighting unknowns regarding cohesive sediment exfiltration without framework mobilization. This study uniquely integrates field measurements, flume experiments, and modelling strategies to evaluate the transport and fate of cohesive sediment in a gravel-bed river. Accordingly, our findings advance current knowledge on the mechanistic understanding of cohesive sediment transport and highlight future research directions.

GMDIC: a digital image correlation measurement method based on global matching for large deformation displacement fields

The digital image correlation method is a non-contact optical measurement method, which has the advantages of full-field measurement, simple operation, and high measurement accuracy. The traditional DIC method can accurately measure displacement and strain fields, but there are still many limitations. (i) In the measurement of large displacement deformations, the calculation accuracy of the displacement field and strain field needs to be improved due to the unreasonable setting of parameters such as subset size and step size. (ii) It is difficult to avoid under-matching or over-matching when reconstructing smooth displacement or strain fields. (iii) When processing large-scale image data, the computational complexity will be very high, resulting in slow processing speeds. In recent years, deep-learning-based DIC has shown promising capabilities in addressing the aforementioned issues. We propose a new, to the best of our knowledge, DIC method based on deep learning, which is designed for measuring displacement fields of speckle images in complex large deformations. The network combines the multi-head attention Swin-Transformer and the high-efficient channel attention module ECA and adds positional information to the features to enhance feature representation capabilities. To train the model, we constructed a displacement field dataset that conformed to the real situation and contained various types of speckle images and complex deformations. The measurement results indicate that our model achieves consistent displacement prediction accuracy with traditional DIC methods in practical experiments. Moreover, our model outperforms traditional DIC methods in cases of large displacement scenarios.

Supervised Deep Learning for Detecting and Locating Passive Seismic Events Recorded with DAS: A Case Study.

Exploring shallow mineral resources requires acquiring denser seismic data, for which Distributed Acoustic Sensing is an effective enabler and relevant to mining operations monitoring. Passive seismic can be of interest in characterizing the subsurface; however, dealing with large amounts of data pushes against the limits of existing computational systems and algorithms, especially for continuous monitoring. Hence, more than ever, novel data analysis methods are needed. In this article, we investigate using synthetic seismic data, paired with real noise recordings, as part of a supervised deep-learning neural network methodology to detect and locate induced seismic sources and explore their potential use to reconstruct subsurface properties. Challenges of this methodology were identified and addressed in the context of induced seismicity applications. Data acquisition and modelling were discussed, preparation workflows were implemented, and the method was demonstrated on synthetic data and tested on relevant seismic monitoring field dataset from the Otway CO2 injection site. Conducted tests confirmed the effects of time shifts, signal-to-noise ratios, and geometry mismatches on the performance of trained models. Those promising results showed the method's applicability and paved the way for potential application to more field data, such as seismic while drilling.

Deep learning networks-based tomato disease and pest detection: a first review of research studies using real field datasets.

Recent advances in deep neural networks in terms of convolutional neural networks (CNNs) have enabled researchers to significantly improve the accuracy and speed of object recognition systems and their application to plant disease and pest detection and diagnosis. This paper presents the first comprehensive review and analysis of deep learning approaches for disease and pest detection in tomato plants, using self-collected field-based and benchmarking datasets extracted from real agricultural scenarios. The review shows that only a few studies available in the literature used data from real agricultural fields such as the PlantDoc dataset. The paper also reveals overoptimistic results of the huge number of studies in the literature that used the PlantVillage dataset collected under (controlled) laboratory conditions. This finding is consistent with the characteristics of the dataset, which consists of leaf images with a uniform background. The uniformity of the background images facilitates object detection and classification, resulting in higher performance-metric values for the models. However, such models are not very useful in agricultural practice, and it remains desirable to establish large datasets of plant diseases under real conditions. With some of the self-generated datasets from real agricultural fields reviewed in this paper, high performance values above 90% can be achieved by applying different (improved) CNN architectures such as Faster R-CNN and YOLO.

Robust full waveform inversion with deep Hessian deblurring

SUMMARY Full waveform inversion (FWI) is a technique widely used in geophysics to obtain high-resolution subsurface velocity models from waveform seismic data. Due to its large computation cost, most flavours of FWI rely only on the computation of the gradient of the loss function to estimate the update direction, therefore ignoring the contribution of the Hessian. Depending on the level of computational resources one can afford, an approximate of the inverse of the Hessian can be calculated and used to speed up the convergence of FWI towards the global (or a plausible local) minimum. In this work, we propose to use an approximate Hessian computed from a linearization of the wave equation as commonly done in least-squares migration. More precisely, we rely on the link between a migrated image and a doubly migrated image (i.e. an image obtained by demigration–migration of the migrated image) to estimate the inverse of the Hessian. However, instead of using non-stationary compact filters to link the two images and approximate the Hessian, we propose to use a deep neural network to directly learn the mapping between the FWI gradient (output) and its Hessian (blurred) counterpart (input). By doing so, the network learns to act as an approximate inverse Hessian: as such, when the trained network is applied to the FWI gradient, an enhanced update direction is obtained, which is shown to be beneficial for the convergence of FWI. The weights of the trained (deblurring) network are then transferred to the next FWI iteration to expedite convergence. We demonstrate the effectiveness of the proposed approach on two synthetic data sets and a field data set.

Development of 1 ×1 km gridded emission inventory for air quality assessment and mitigation strategies from open biomass burning in Karnataka, India

In this work, we have developed a high-resolution (1 ×1 km grid) emission inventory of greenhouse gases and air pollutants from open biomass burning (OBB) in Karnataka for 2000, 2005, 2010, 2015, 2020, and 2022. Secondary and field datasets were used to validate the emission inventories. The GIS-based statistical model and activity data were used to estimate greenhouse and air pollutant gas emissions from OBB (Forest, grassland, and crop fires). The uncertainty analysis of the emissions were done using the Monte Carlo Simulation (MCS) model and Low Emission Analysis Platform (LEAP) tool was used for projecting the emissions to 2050 based on the annual average growth rate. According to our results, the total OBB burned areas (%) in Karnataka during the aforementioned years were 24.43, 15.69, 15.95, 11.41, 26.73, and 5.78. The results showed that total annual average OBB emissions in Karnataka for 2000–2022 in Gg were SO2 (6.67), NOx (9.48), NH3(9.80), CO (670.12), OC (59.78), BC (5.09), CO2 (11071 .11), CH4(33.68), PM10(84.29), PM2.5(81.08), NMVOC (74.13), and NO2 (4.80) respectively. The districts of Kalburgi, Raichur, Bagalkot, Uttara Kannada, and Shivamogga have the highest emissions from OBB in Karnataka from 2000 to 2022. The emission uncertainty results showed ±26.92, ±28.17, and ±30.27 % for the crop, grassland, and forest fires, respectively. Under the business-as-usual (BAU) scenarios, the CH4 emissions from forest, grassland, and crop fires are projected at 8823, 2.64, and 5262 metric tonnes by 2030 and 58046, 10, and 30700 metric tonnes by 2050, respectively. The developed high-resolution inventories are the most up-to-date surface emission datasets with hotspots in Karnataka from OBB emission. National Clean Air Programme (NCAP) and India’s Net-Zero Emission target by 2070, our high-resolution emission inventories and mitigation strategies may be helpful for air quality monitoring, health benefits analysis, and policy-making studies in Karnataka, India.

One‐dimensional deep learning inversion of marine controlled‐source electromagnetic data

AbstractThis paper explores the application of machine learning techniques, specifically deep learning, to the inverse problem of marine controlled‐source electromagnetic data. A novel approach is proposed that combines the convolutional neural network and recurrent neural network architectures to reconstruct layered electrical resistivity variation beneath the seafloor from marine controlled‐source electromagnetic data. The approach leverages the strengths of both convolutional neural network and recurrent neural network, where convolutional neural network is used for recognizing and classifying features in the data, and recurrent neural network is used to capture the contextual information in the sequential data. We have built a large synthetic dataset based on one‐dimensional forward modelling of a large number of resistivity models with different levels of electromagnetic structural complexity. The combined learning of convolutional neural network and recurrent neural network is used to construct the mapping relationship between the marine controlled‐source electromagnetic data and the resistivity model. The trained network is then used to predict the distribution of resistivity in the model by feeding it with marine controlled‐source electromagnetic responses. The accuracy of the proposed approach is examined using several synthetic scenarios and applied to a field dataset. We explore the sensitivity of deep learning inversion to different electromagnetic responses produced by resistive targets distributed at different depths and with varying levels of noise. Results from both numerical simulations and field data processing consistently demonstrate that deep learning inversions reliably reconstruct the subsurface resistivity structures. Moreover, the proposed method significantly improves the efficiency of electromagnetic inversion and offers significant performance advantages over traditional electromagnetic inversion methods.

Rural children know cavity-nesting birds of the Atlantic Forest but may underappreciate their critical habitat

Abstract Cavity-nesting birds are a diverse and charismatic community, with a common need for tree cavities that makes them vulnerable to land management by humans. However, little research has formally integrated human social aspects into management recommendations for the conservation of cavity-nesting birds. In agroecosystems, people's management decisions modify and define the habitat availability for native cavity-nesting species. These behaviors during adulthood are related to people's worldviews and are shaped, in part, by childhood experiences. On-going forest loss may reduce opportunities for children to interact with and learn from cavity-nesting birds and their habitats. We used a social-ecological framework to assess rural children's knowledge and representations of native cavity-nesting birds and their habitats in agroecosystems of the threatened Atlantic Forest of Argentina. We employed “freelists” and “draw-and-explain” strategies with 235 children from 19 rural schools, and then compared results with a 4-yr dataset of trees (n = 328) and tree-cavity nests (n = 164) in the same study area. Children listed a high diversity (93 taxa) of native cavity-nesting birds, especially parrots (Psittacidae), toucans (Ramphastidae), and woodpeckers (Picidae), which they mostly recognized as cavity-nesters. However, children drew agricultural landscapes with few of the habitat features that these birds require (e.g., tree cavities, native forest). Exotic trees were overrepresented in drawings (40% of mentions) compared to our field dataset of nests (10%) and trees on farms (15%). Although children mentioned and depicted a high diversity of native cavity-nesting birds, our results may reveal a problematic extinction of experience regarding how these birds interact with their habitat. To strengthen children's contextualized knowledge and promote their long-term commitment to the conservation of cavity-nesting species, we recommend fostering meaningful experiences for children to interact with native cavity-nesting birds and recognize their habitat needs. A version of this article translated into Spanish is available in the Supplementary Material 1.

Intercropping Legumes Improves Long Term Productivity and Soil Carbon and Nitrogen Stocks in Sub‐Saharan Africa

AbstractFood, feed, and fiber production needs to increase to support demands of the growing population in Sub‐Saharan Africa (SSA), while soil fertility continues to decline. Intercropping, the cultivation of two or more crop species on the same field, can provide yield benefits and is suggested to positively affect soil organic carbon (C) and nitrogen (N) stocks. This study uses the biogeochemical model system LandscapeDNDC with the objective to (a) represent maize‐legume intercropping systems in different bioregions in SSA by simultaneously simulating both crops and their interactions and (b) assess long‐term (20 years) impacts of intercropping under varying mineral fertilizer inputs (0–150 kg N ha−1 yr−1) on productivity as well as soil organic C and N stocks. We test LandscapeDNDC on 82 field data sets (site‐year‐treatment combinations) from 18 sites to represent yields and soil C/N dynamics of maize‐legume intercropping systems. Using the model for long‐term scenario simulations showed that intercropping allows to sustain productivity and to improve or maintain SOC stock in low or zero fertilizer systems if all residues are returned to the soil. In contrast, for sole‐cropped maize systems, a decline in SOC stocks was simulated unless a minimum of 35 kg N ha−1 yr−1 of fertilizer was applied at full residue return. We conclude that intercropping using legumes alongside sufficient residue return allows for stabilizing long‐term yields while avoiding SOC losses even with low fertilizer N inputs. Overall, our study confirms the potential of intercropping as a sustainable agricultural practice that could significantly contribute to food security in SSA.

Physics-Informed Deep-Learning Models Improve Forecast Scalability, Reliability

_ This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper URTeC 4042557, “Physics-Informed Deep-Learning Models for Improving Shale and Tight Forecast Scalability and Reliability,” by Kainan Wang, SPE, Lichi Deng, and Yuzhe Cai, SPE, Chevron, et al. The paper has not been peer reviewed. _ In the complete paper, the authors present a workflow that combines probabilistic modeling and deep-learning models trained on an ensemble of physics models to improve scalability and reliability for shale and tight reservoir forecasting. Their approach is applied to synthetic cases and many wells in the Permian Basin. Through hindsight studies, these models have been demonstrated to generate realistic and diverse production curves, capture the physics of unconventional flow, quantify well-production-outlook uncertainty, and help interpretation of subsurface uncertainty. Introduction In the realm of unconventional asset development, scalable forecasting is a key component in forecast reliability. In recent years, data-driven machine-learning models and workflows have emerged as potent tools for predicting well performance, particularly in scenarios where wells share similar reservoir properties, completion designs, and operational conditions. A novel approach known as physics-informed machine learning (PIML) has gained prominence. These models leverage the strengths of machine learning while honoring the constraints imposed by physical laws. By incorporating observational, inductive, or learning bias, they bridge the gap between empirical data and fundamental physics. In this paper, the authors showcase the application of a PIML system tailored specifically for unconventional reservoirs. Their methodology involves the selection of an appropriate physics modeling framework, coupled with the development of multiple machine-learning models that augment the limited field data set. Data Acquisition and Generation Bottomhole pressure (BHP) data are key to reliable forecasting because they reflect constraints to well production. Downhole pressure gauges can provide measurements of BHP accurately, but these measurements are not always available. Numerical models can be used to correlate surface measurements to BHP with some accuracy, taking into consideration the well configuration and fluid properties, yet it is too time-consuming to develop such models for many wells in relation to the speed needed for unconventional reservoir development. An in-house machine-learning model was developed for BHP calculation to further augment the data set of gauge measurements and numerical modeling. Through a workflow that chooses the wells with higher-fidelity data as a training set, and careful calibration and iterations between incrementally adding wells of high-fidelity data to reduce uncertainty, the trained machine- learning model is able to predict BHP accurately for many more wells. Another type of engineering data important for reliable well-performance analysis and forecasting is reservoir-fluid description. In this study, the authors leveraged an internally developed pressure/volume/temperature (PVT) machine-learning model trained on proprietary PVT data, which was also shown to provide reliable predictions to key PVT properties such as saturation pressure, formation volume factor, and viscosity from limited inputs at the reservoir and separator level. These PVT models not only enhance scalability of any downstream forecasting models but also play a crucial role in physics-informed uncertainty analysis for these models.

Flow field reconstruction of compressor blade cascade based on deep learning methods

In the design process of compressors, calculating the S1 flow surface involves solving the Navier-Stokes equations, which results in slow convergence and makes determining cascade characteristics time-consuming. However, deep learning offers significant advantages in flow field reconstruction by not only automatically extracting complex flow features and reducing prediction time but also providing high accuracy in reconstruction. This paper implements the rapid reconstruction of the compressor S1 flow surface cascade flow field using two deep learning models: U-Net and 1D-CNN. Using a double-circular arc airfoil as an example, we selected four key design parameters that define the geometry and position of the airfoil, ultimately designing 5,292 sets of cascade geometries. By performing batch meshing and CFD simulations, we built a cascade flow field dataset. The U-Net neural network uses design parameters as input and outputs the aerodynamic distribution of the cascade flow field. After training, it can directly predict the flow field based on the design parameters. Since the U-Net model cannot directly obtain the aerodynamic parameter distribution and flow field aerodynamic coefficients on the airfoil surface, a 1D-CNN model is used as a complementary approach. The 1D-CNN model takes the design parameters as input and outputs the aerodynamic parameter distribution on the airfoil surface and the flow field aerodynamic coefficients. The prediction results show that the U-Net model achieves an average relative error of less than 1% in cascade flow field reconstruction, while the 1D-CNN model achieves an average relative error of less than 1% in predicting the pressure recovery coefficient and less than 2% in predicting the total pressure loss coefficient. This study presents a method for the rapid reconstruction of compressor blade cascade flow fields, which helps improve design efficiency and shorten the design cycle.