An official website of the United States government
Abstract Applications of process‐based models (PBM) for predictions are confounded by multiple uncertainties and computational burdens, resulting in appreciable errors. A novel modeling framework combining a high‐fidelity PBM with surrogate and machine learning (ML) models is developed to tackle these challenges and applied for streamflow prediction. A surrogate model permits high computational efficiency of a PBM solution at a minimum loss of its accuracy. A novel probabilistic ML model partitions the PBM‐surrogate prediction errors into reducible and irreducible types, quantifying their distributions that arise due to both explicitly perceived uncertainties (such as parametric) or those that are entirely hidden to the modeler (not included or unexpected). Using this approach, we demonstrate a substantial improvement of streamflow predictive accuracy for a case study urbanized watershed. Such a framework provides an efficient solution combining the strengths of high‐fidelity and physics‐agnostic models for a wide range of prediction problems in geosciences.
more »
« less
Predicting biomass comminution: Physical experiment, population balance model, and deep learning
An extended population balance model (PBM) and a deep learning-based enhanced deep neural operator (DNO+) model are introduced for predicting particle size distribution (PSD) of comminuted biomass through a large knife mill. Experimental tests using corn stalks with varied moisture contents, mill blade speeds, and discharge screen sizes are conducted to support model development. A novel mechanism in the extended PBM allows for including additional input parameters such as moisture content, which is not possible in the original PBM. The DNO+ model can include influencing factors of different data types such as moisture content and discharge screen size, which significantly extends the engineering applicability of the standard DNO model that only admits feed PSD and outcome PSD. Test results show that both models are remarkably accurate in the calibration or training parameter space and can be used as surrogate models to provide effective guidance for biomass preprocessing design.
more »
« less
- PAR ID:
- 10623989
- Publisher / Repository:
- Elsevier
- Date Published:
- Journal Name:
- Powder Technology
- Volume:
- 441
- ISSN:
- 0032-5910
- Page Range / eLocation ID:
- 119830
- Subject(s) / Keyword(s):
- Deep neural operator Machine learning Corn stover Preprocessing Knife mill
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Plastic-bonded granular materials (PBM) are widely used in industrial sectors, including building construction, abrasive applications, and defense applications such as plastic-bonded explosives. The mechanical behavior of PBM is highly nonlinear, irreversible, rate dependent, and temperature sensitive governed by various micromechanical attributions such as grain crushing and binder damage. This paper presents a thermodynamically consistent, microstructure-informed constitutive model to capture these characteristic behaviors of PBM. Key features of the model include a breakage internal variable to upscale the grain-scale information to the continuum level and to predict grain size evolution under mechanical loading. In addition, a damage internal state variable is introduced to account for the damage, deterioration, and debonding of the binder matrix upon loading. Temperature is taken as a fundamental external state variable to handle non-isothermal loading paths. The proposed model is able to capture with good accuracy several important aspects of the mechanical properties of PBM, such as pressure-dependent elasticity, pressure-dependent yield strength, brittle-to-ductile transition, temperature dependency, and rate dependency in the post-yielding regime. The model is validated against multiple published datasets obtained from confined and unconfined compression tests, covering various PBM compositions, confining pressures, temperatures, and strain rates.more » « less
-
Abstract The size distribution of suspended particles influences several processes in aquatic ecosystems, including light propagation, trophic interactions, and biogeochemical cycling. The shape of the particle size distribution (PSD) is commonly modeled as a single‐slope power law in oceanographic studies, which can be used to further estimate the relative contributions of different particle size classes to particle number, area, and volume concentration. We use a data set of 168 high size‐resolution PSD measurements in Arctic oceanic waters to examine variability in the shape of the PSD over the particle diameter range 0.8 to 120 μm. An average value of −3.6 ± 0.33 was obtained for the slope of a power law fitted over this size range, consistent with other studies. Our analysis indicates, however, that this model has significant limitations in adequately parameterizing the complexity of the PSD, and thus performs poorly in predicting the relative contributions of different size intervals such as those based on picoplankton, nanoplankton, and microplankton size classes. Similarly, median particle size was also generally a poor indicator of these size class contributions. Our results suggest that alternative percentile diameters derived from the cumulative distribution functions of particle number, cross‐sectional area, and volume concentration may provide better metrics to capture the overall shape of the PSD and to quantify the contributions of different particle size classes.more » « less
-
Abstract Much of our conceptual understanding of midlatitude atmospheric motion comes from two-layer quasigeostrophic (QG) models. Traditionally, these QG models do not include moisture, which accounts for an estimated 30%–60% of the available energy of the atmosphere. The atmospheric moisture content is expected to increase under global warming, and therefore, a theory for how moisture modifies atmospheric dynamics is crucial. We use a two-layer moist QG model with convective adjustment as a basis for analyzing how latent heat release and large-scale moisture gradients impact the scalings of a midlatitude system at the synoptic scale. In this model, the degree of saturation can be tuned independently of other moist parameters by enforcing a high rate of evaporation from the surface. This allows for study of the effects of latent heat release at saturation, without the intrinsic nonlinearity of precipitation. At saturation, this system is equivalent to the dry QG model under a rescaling of both length and time. This predicts that the most unstable mode shifts to smaller scales, the growth rates increase, and the inverse cascade extends to larger scales. We verify these results numerically and use them to verify a framework for the complete energetics of a moist system. We examine the spectral features of the energy transfer terms. This analysis shows that precipitation generates energy at small scales, while dry dynamics drive a significant broadening to larger scales. Cascades of energy are still observed in all terms, albeit without a clearly defined inertial range. Significance Statement The effect of moist processes, especially the impact of latent heating associated with condensation, on the size and strength of midlatitude storms is not well understood. Such insight is particularly needed in the context of global warming, as we expect moisture to play a more important role in a warmer world. In this study, we provide intuition into how including condensation can result in midlatitude storms that grow faster and have features on both larger and smaller scales than their dry counterparts. We provide a framework for quantifying these changes and verify it for the special case where it is raining everywhere. These findings can be extended to the more realistic situation where it is only raining locally.more » « less
-
The Soil Moisture Active Passive (SMAP) mission measures important soil moisture data globally. SMAP's products might not always perform better than land surface models (LSM) when evaluated against in situ measurements. However, we hypothesize that SMAP presents added value for long-term soil moisture estimation in a data fusion setting as evaluated by in situ data. Here, with the help of a time series deep learning (DL) method, we created a seamlessly extended SMAP data set to test this hypothesis and, importantly, gauge whether such benefits extend to years beyond SMAP's limited lifespan. We first show that the DL model, called long short-term memory (LSTM), can extrapolate SMAP for several years and the results are similar to the training period. We obtained prolongation results with low-performance degradation where SMAP itself matches well with in situ data. Interannual trends of root-zone soil moisture are surprisingly well captured by LSTM. In some cases, LSTM's performance is limited by SMAP, whose main issue appears to be its shallow sensing depth. Despite this limitation, a simple average between LSTM and an LSM Noah frequently outperforms Noah alone. Moreover, Noah combined with LSTM is more skillful than when it is combined with another LSM. Over sparsely instrumented sites, the Noah-LSTM combination shows a stronger edge. Our results verified the value of LSTM-extended SMAP data. Moreover, DL is completely data driven and does not require structural assumptions. As such, it has its unique potential for long-term projections and may be applied synergistically with other model-data integration techniques.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1016/j.powtec.2024.119830
