A Review of Modern Machine Learning Techniques in the Prediction of Remaining Useful Life of Lithium-Ion Batteries

Batteries, Год журнала: 2022, Номер 9(1), С. 13 - 13

Опубликована: Дек. 25, 2022

The intense increase in air pollution caused by vehicular emissions is one of the main causes changing weather patterns and deteriorating health conditions. Furthermore, renewable energy sources, such as solar, wind, biofuels, suffer from supply chain-related uncertainties. electric vehicles’ powered energy, stored a battery, offers an attractive option to overcome uncertainties certain extent. development implementation cutting-edge vehicles (EVs) with long driving ranges, safety, higher reliability have been identified critical decarbonizing transportation sector. Nonetheless, capacity time usage, environmental degradation factors, end-of-life repurposing pose significant challenges usage lithium-ion batteries. In this aspect, determining battery’s remaining usable life (RUL) establishes its efficacy. It also aids testing various EV upgrades identifying factors that will improve their efficiency. Several nonlinear complicated parameters are involved process. Machine learning (ML) methodologies proven be promising tool for optimizing modeling engineering domain (non-linearity complexity). contrast scalability temporal limits battery degeneration, ML techniques provide non-invasive solution excellent accuracy minimal processing. Based on recent research, study presents objective comprehensive evaluation these challenges. RUL estimations explained detail, including examples approach applicability. many thoroughly individually studied. Finally, application-focused overview offered, emphasizing advantages terms efficiency accuracy.

Язык: Английский

---

Predicting Redox Conditions in Groundwater at a National Scale Using Random Forest Classification

Environmental Science & Technology, Год журнала: 2024, Номер 58(11), С. 5079 - 5092

Опубликована: Март 7, 2024

Redox conditions in groundwater may markedly affect the fate and transport of nutrients, volatile organic compounds, trace metals, with significant implications for human health. While many local assessments redox have been made, spatial variability reaction rates makes determination at regional or national scales problematic. In this study, were predicted contiguous United States using random forest classification by relating measured water quality data from over 30,000 wells to natural anthropogenic factors. The model correctly oxic/suboxic 78 79% samples out-of-bag hold-out sets, respectively. Variables describing geology, hydrology, soil properties, hydrologic position among most important factors affecting likelihood oxic groundwater. Important variables tended relate aquifer recharge, travel time, prevalence electron donors, which are key drivers Partial dependence plots suggested that decreased sharply as streams approached gradually depth below table increased. probability increased base flow index values increased, likely due well-drained soils geologic materials high areas. topographic wetness (TWI) decreased. High occur areas a propensity standing overland flow, limit delivery dissolved oxygen recharge; higher TWI also tend discharge areas, contain long times. A second was developed predict elevated manganese (Mn) concentrations (i.e., ≥50 μg/L). Mn relied on same be used identify where Mn-reducing there is an risk domestic supplies concentrations. Model predictions produced study help regions country vulnerability stream groundwater-derived contaminants.

Язык: Английский

---

Estimating Lithium Concentrations in Groundwater Used as Drinking Water for the Conterminous United States

Environmental Science & Technology, Год журнала: 2024, Номер 58(2), С. 1255 - 1264

Опубликована: Янв. 2, 2024

Lithium (Li) concentrations in drinking-water supplies are not regulated the United States; however, Li is included 2022 U.S. Environmental Protection Agency list of unregulated contaminants for monitoring by public water systems. used pharmaceutically to treat bipolar disorder, and studies have linked its occurrence drinking human-health outcomes. An extreme gradient boosting model was developed estimate geogenic supply wells throughout conterminous States. The trained using measurements from ∼13,500 predictor variables related natural groundwater. predicts probability four concentration classifications, ≤4 μg/L, >4 ≤10 >10 ≤30 >30 μg/L. Model predictions were evaluated held out training with new data an accuracy 47–65%. Important include average annual precipitation, well depth, soil geochemistry. mapped at a spatial resolution 1 km2 represent depths associated public- private-supply wells. This hydrologists public-health researchers exposure compare national-scale better understanding dose–response low (<30 μg/L) Li.

Язык: Английский

---

Improving crop-specific groundwater use estimation in the Mississippi Alluvial Plain: Implications for integrated remote sensing and machine learning approaches in data-scarce regions

Journal of Hydrology Regional Studies, Год журнала: 2024, Номер 52, С. 101674 - 101674

Опубликована: Фев. 1, 2024

The Mississippi Alluvial Plain (MAP) in the United States (US). Understanding local-scale groundwater use, a critical component of water budget, is necessary for implementing sustainable management practices. MAP one most productive agricultural regions US and extracts more than 11 km3/year irrigation activities. Consequently, groundwater-level declines region pose substantial challenge to sustainability, hence, we need reliable pumping monitoring solutions manage this resource appropriately. We incorporate remote sensing datasets machine learning improve an existing lookup table-based model use previously developed by U.S. Geological Survey (USGS). Here, employ Distributed Random Forests, ensemble algorithm predict annual monthly (2014–2020) throughout at 1-km resolution, using data from flowmeters Delta. Our compares favorably with USGS model, higher R2 (0.51 compared 0.42 previous model), lower root mean square error (RMSE) absolute (MAE)— 0.14 m 0.09 m, respectively our 0.15 0.1 model. Therefore, work advances ability scarce or limited in-situ withdrawal availability.

Язык: Английский

---

Evaluation of the lithium resource in the Smackover Formation brines of southern Arkansas using machine learning

Science Advances, Год журнала: 2024, Номер 10(39)

Опубликована: Сен. 27, 2024

Global demand for lithium, the primary component of lithium-ion batteries, greatly exceeds known supplies, and this imbalance is expected to increase as world transitions away from fossil fuel energy sources. High concentrations lithium in brines have been observed Smackover Formation southern Arkansas (>400 milligrams per liter). We used published newly collected brine concentration data train a random forest machine-learning model using geologic, geochemical, temperature explanatory variables create map predicted across Arkansas. Using these maps with reservoir parameters geologic information, we calculated that there are 5.1 19 million tons Arkansas, which represents 35 136% current US resource estimate. Based on calculations, 2022, 5000 dissolved were brought surface within waste streams oil, gas, bromine industries.

Язык: Английский

---

Application of machine learning in delineating groundwater contamination at present times and in climate change scenarios

Current Opinion in Environmental Science & Health, Год журнала: 2024, Номер 39, С. 100554 - 100554

Опубликована: Май 5, 2024

Язык: Английский

---

Prediction of wild pistachio ecological niche using machine learning models

Ecological Informatics, Год журнала: 2022, Номер 72, С. 101907 - 101907

Опубликована: Ноя. 16, 2022

Язык: Английский

---

Riverine Particulate Carbon, Nitrogen, and Phosphorus Are Decoupled From Land Cover at the Continental Scale

Global Biogeochemical Cycles, Год журнала: 2025, Номер 39(3)

Опубликована: Март 1, 2025

Abstract While inland freshwater networks cover less than 4% of the Earth's terrestrial surface, these ecosystems play a disproportionately large role in global cycles [C]arbon, [N]itrogen, and [P]hosphorus, making streams rivers critical regulators nutrient balance at regional continental scales. Foundational studies have established relative importance hydrologic regime, land cover, instream removal processes for controlling transport processing C, N, P river networks. However, particulate can make up proportion total material during high flows. To constrain patterns biogeochemistry riverine particulates, we characterized modeled dissolved concentration variability scale using open‐access data from 27 National Ecological Observatory Network (NEON) sites across United States. We analyzed Boosted Regression Trees (BRTs) to statistically identify if characteristics could predict quantity quality stream particulates. The BRT models revealed that does not strongly dynamics NEON but indicate might be more important catchment alone. In addition, our study demonstrates consistent particulates forms, highlighting their likely significance biogeochemical along continuum.

Язык: Английский

---

Heavy metals prediction system in groundwater using online sensor and machine learning for water management: the case of typical industrial park

Environmental Pollution, Год журнала: 2025, Номер 374, С. 126270 - 126270

Опубликована: Апрель 22, 2025

Язык: Английский

---

Prioritizing water availability study settings to address geogenic contaminants and related societal factors

Environmental Monitoring and Assessment, Год журнала: 2024, Номер 196(3)

Опубликована: Фев. 24, 2024

Abstract Water availability for human and ecological uses depends on both water quantity quality. The U.S. Geological Survey (USGS) is developing strategies prioritizing regional-scale watershed basin-scale studies of across the nation. Previous USGS ranking processes incorporated primarily factors but are now considering additional quality factors. This study presents a based potential impacts geogenic constituents consideration societal related to High-concentration constituents, including trace elements radionuclides, among most prevalent contaminants limiting in USA globally. Geogenic commonly occur groundwater because subsurface water–rock interactions, their distributions controlled by complex geochemical processes. constituent mobility can also be affected activities (e.g., mining, energy production, irrigation, pumping). Societal relations drinking sources information often overlooked when evaluating research priorities. Sociodemographic characteristics, data gaps resulting from historical data-collection disparities, infrastructure condition/age examples consider regarding environmental justice. paper approaches areas contiguous suite conventional physical variables with without Simultaneous could provide decision makers more diverse, interdisciplinary tools increase equity reduce bias focused future studies.

Язык: Английский

---