Mechanistic Model Research Articles

Multispectral and hyperspectral image fusion via multi-model deep prior regularization

ABSTRACT Due to the difficulty of collecting a sufficient number of photons within a narrow spectral window, hyperspectral images often have low spatial resolution. In order to solve this problem without increasing hardware costs, a large number of image super-resolution algorithms have been proposed and have shown promising results. However, the performance of these algorithms is inconsistent in different datasets. To fix this situation, we proposed a multi-model deep prior regularization method for multispectral and hyperspectral image fusion. This method leverages mechanistic knowledge to combine various super-resolution results into a single algorithm, significantly improving the quality of the hyperspectral images. Specifically, we first proposed a variable residual mixed attention network. This network takes a low-resolution hyperspectral image and a high-resolution multispectral image as inputs, producing deep prior results from three different modes by adjusting spatial and channel attention. Then, we reconstructed the prior results along the spectral dimension and used them as regularization terms in our mechanistic model. We transformed the problem of finding the optimal solution of the model into solving the Sylvester equation and applied a fast solver for this equation to create the high-resolution hyperspectral images. Additionally, we used a genetic algorithm to determine the best hyperparameters for different deep prior results. Through validation on three datasets, we have demonstrated that our method performs better than existing methods in both quantitative and qualitative comparisons.

Mechanistic Model of Fatigue in Ultrasonic Assisted Machining.

Anti-fatigue design in the machining process of aviation material requires advanced processes to enhance the surface integrity and a holistic model which can optimize the process aiming at maximum fatigue life. In the present study, the axial ultrasonic assisted milling process was utilized to machine the Inconel 718 while the process executes the thermomechanical cutting and peening action simultaneously. To optimize the process factors, a hybrid model using a combination of regression analysis and an analytical model was developed to correlate the machining factors, i.e., vibration amplitude, cutting velocity and feed rate to fatigue life. Herein, the former was used to map the process inputs to surface integrity aspects (SIAs), viz. roughness, hardness and residual stress; then, the SIA was mapped to fatigue life through a stress-based approach. The obtained results revealed that there is close agreement between the measured and predicted values of fatigue life where the prediction error is less than two times the dispersion. On the other hand, applying ultrasonic vibration at the highest amplitude together with the maximum feed rate and cutting velocity yield significant improvement in fatigue life, i.e., three times the same condition without ultrasonic vibration in light of the enhancement of compressive residual stress and work hardening of the surface layers.

Toward mechanistic models “augmented” by machine learning. Example of a drying simulation data set exploited by a Physics Informed Neural Network

Mechanistic modeling of drying is well-established since several decades. Based on fundamental balance equations and driven by relevant material parameters, it can predict the entire process, including configurations that were not observed before. Besides, thanks to their ability to tackle non-linear and dynamics problems, approaches based on Machine Learning (ML) based are capable of coping with complex situations even better than mechanistic modeling. The main drawback of mechanistic models is their complexity as operational tools, namely in providing the whole set of product characteristics, while the main drawback of ML tools is its restriction to the domain paved by the data set. This paper summarizes the physics of the mechanistic formulation and then presents the different possibilities of coupling the mechanistic and ML approaches to obtain an “augmented” mechanistic model. The idea is to merge the advantages of both worlds. Different strategies can be imagined: A full coupled method (hybrid model) A fully decoupled method A cascade coupling The second part of the paper gives examples of a fully decoupled approach the mechanistic model is used to populate a data set, which is further exploited by a neural network. The originality of the work is to compare a classical neural network with a Physics Informed Neural Network.

Composite dyadic models for spatio-temporal data.

Mechanistic statistical models are commonly used to study the flow of biological processes. For example, in landscape genetics, the aim is to infer spatial mechanisms that govern gene flow in populations. Existing statistical approaches in landscape genetics do not account for temporal dependence in the data and may be computationally prohibitive. We infer mechanisms with a Bayesian hierarchical dyadic model that scales well with large data sets and that accounts for spatial and temporal dependence. We construct a fully connected network comprising spatio-temporal data for the dyadic model and use normalized composite likelihoods to account for the dependence structure in space and time. We develop a dyadic model to account for physical mechanisms commonly found in physical-statistical models and apply our methods to ancient human DNA data to infer the mechanisms that affected human movement in Bronze Age Europe.

Flammable futures—storylines of climatic impacts on wildfire events and palm oil plantations in Indonesia

Abstract Wildfire events are driven by complex interactions of climate and anthropogenic interventions. Predictions of future wildfire events, their extremity, and their impact on the environment and economy must consider the interactions between these drivers. Economic policy and land use decisions influence the susceptibility of an area to climate extremes, the probability of burning, and future decision making. To better understand how climate-driven drought events and adaptation efforts affect burned area, agricultural production losses, and land use decisions, we developed a storyline approach centered on Indonesia’s 2015 fire events, which saw significant (>5%) production losses of palm oil. We explored analogous events under three warming conditions and two storylines (multi-model ensemble mean climate change and high impact). We employed a model chain consisting of CMIP6 climate modeling to quantify climate change impacts, a wildfire climate impacts and adaptation model (FLAM) to predict burned areas, and the Global Biosphere Management Model (GLOBIOM) to predict resultant production losses and socio-economic consequences in the oil palm sector in Indonesia and, by extension, the EU. FLAM is a mechanistic fire model used to reproduce and project wildfires based on various criteria and input variables; GLOBIOM is a global economic land use model, which assesses land use competition and provides economic impacts based on scenario data. We found that total burned area and production loss can increase by up to 25% and lead to local price increases up to 70%, with only minor differences beyond 2.5 degrees of warming. Our results highlight the importance of considering the interactions of future warming, drought conditions, and extreme weather events when predicting their impacts on oil palm losses and burned area. This study sets the stage for further exploration on the impacts of land management policies on local and international environments and economies in the context of global warming.

Large-Scale Mechanistic Models of Brain Circuits with Biophysically and Morphologically Detailed Neurons.

Understanding the brain requires studying its multiscale interactions from molecules to networks. The increasing availability of large-scale datasets detailing brain circuit composition, connectivity, and activity is transforming neuroscience. However, integrating and interpreting this data remains challenging. Concurrently, advances in supercomputing and sophisticated modeling tools now enable the development of highly detailed, large-scale biophysical circuit models. These mechanistic multiscale models offer a method to systematically integrate experimental data, facilitating investigations into brain structure, function, and disease. This review, based on a Society for Neuroscience 2024 MiniSymposium, aims to disseminate recent advances in large-scale mechanistic modeling to the broader community. It highlights (1) examples of current models for various brain regions developed through experimental data integration; (2) their predictive capabilities regarding cellular and circuit mechanisms underlying experimental recordings (e.g., membrane voltage, spikes, local-field potential, electroencephalography/magnetoencephalography) and brain function; and (3) their use in simulating biomarkers for brain diseases like epilepsy, depression, schizophrenia, and Parkinson's, aiding in understanding their biophysical underpinnings and developing novel treatments. The review showcases state-of-the-art models covering hippocampus, somatosensory, visual, motor, auditory cortical, and thalamic circuits across species. These models predict neural activity at multiple scales and provide insights into the biophysical mechanisms underlying sensation, motor behavior, brain signals, neural coding, disease, pharmacological interventions, and neural stimulation. Collaboration with experimental neuroscientists and clinicians is essential for the development and validation of these models, particularly as datasets grow. Hence, this review aims to foster interest in detailed brain circuit models, leading to cross-disciplinary collaborations that accelerate brain research.

Model-based experiments as epistemic evidence in paleoecology

Where ordinary experiments are impossible and observational data scarce and indirect—particularly in paleoecosystems—computational experiments are often our only means to learn about reality. There are good arguments to count such model-based predictions as evidence, testing hypotheses and updating our beliefs about the world. However, the epistemic weight of computational experiments depends on an adequate model representation of the target system, transparency about predictive uncertainty, and the avoidance of confirmation bias. I argue that mechanistic models are particularly suited for paleoecological predictions but that iterative uncertainty analyses should guide their development. Using a Bayesian framework I propose preregistration and blinded analysis as tools to strengthen the epistemic value of computational experiments. Here, a preregistration marks the boundary between exploratory model development, which establishes credence in the model, and predictive model application, which tests hypotheses. As good modeling practice I suggest clarifying epistemic goals at the outset of a project and accordingly choose methods to maximize the epistemic weight of the computational experiment.

Combining a Data Driven and Mechanistic Model to Predict Capacity and Potential Curve‐Degradation

AbstractThis work compares a state of the art data‐driven model to predict the state of health (SoH) in lithium ion batteries with a new prediction model based on the mechanistic framework. The mechanistic approach attributes the degradation to individual components such as loss of available capacity on each electrode as well as loss of cyclable lithium. By combining the mechanistic framework with data‐driven models for the component losses based on a design of experiment, we achieve a cycle aging model that can predict capacity degradation as well as degradation‐induced changes to the discharge potential curve. Using this cycle aging model alongside with a semi‐empirical calendar aging model, we present a holistic aging model that we validate on independent validation tests containing time‐variant load profiles. While the purely data‐driven model is better at predicting the SoH, the mechanistic model clearly has it advantages in a deeper understanding that can potentially enhance the current methods of tracking and updating the characteristic open‐circuit voltage curve over lifetime.

Dilational rheological properties of extended anionic surfactants at decane-water interface

The dynamic interfacial dilational properties of extended surfactants octadecyl-(polypropylene oxide)m-(polyethylene oxide)n-carboxylic sodium (C18POmEOnC) with different numbers of polypropylene oxide (PO) and polyethylene oxide (EO) groups at the decane–water interface were investigated by means of oscillating drop method. The influences of interface ageing time, dilational frequency, bulk concentration and interfacial pressure on the dilational properties were investigated. The experimental results demonstrate that, at low concentrations, the adsorption film is predominantly influenced by diffusion exchange, while intermolecular interactions at the interface play a more significant role at high concentrations. Beyond the critical micelle concentration (CMC), the primary relaxation process in the adsorption film is driven by the formation of micellar structures. Additionally, the weak hydrophobic PO groups exhibit compressibility and extensibility, facilitating the formation of spring-like helical structures at the decane–water interface. As a result, the dilational modulus increases and the phase angle decreases in low concentration with the increase of PO number. With increasing concentration, the weak hydrophilic EO chains gradually transition from a partially flat state to an upright orientation toward the aqueous phase, forming sublayer structures through polar interactions. At this stage, the increase in EO group chain length has a more pronounced impact on the elastic modulus compared to PO groups. These experimental findings suggest the existence of a mechanistic model for the extended surfactant adsorption films.

Building Digital Twins for Cardiovascular Health: From Principles to Clinical Impact.

The past several decades have seen rapid advances in diagnosis and treatment of cardiovascular diseases and stroke, enabled by technological breakthroughs in imaging, genomics, and physiological monitoring, coupled with therapeutic interventions. We now face the challenge of how to (1) rapidly process large, complex multimodal and multiscale medical measurements; (2) map all available data streams to the trajectories of disease states over the patient's lifetime; and (3) apply this information for optimal clinical interventions and outcomes. Here we review new advances that may address these challenges using digital twin technology to fulfill the promise of personalized cardiovascular medical practice. Rooted in engineering mechanics and manufacturing, the digital twin is a virtual representation engineered to model and simulate its physical counterpart. Recent breakthroughs in scientific computation, artificial intelligence, and sensor technology have enabled rapid bidirectional interactions between the virtual-physical counterparts with measurements of the physical twin that inform and improve its virtual twin, which in turn provide updated virtual projections of disease trajectories and anticipated clinical outcomes. Verification, validation, and uncertainty quantification builds confidence and trust by clinicians and patients in the digital twin and establishes boundaries for the use of simulations in cardiovascular medicine. Mechanistic physiological models form the fundamental building blocks of the personalized digital twin that continuously forecast optimal management of cardiovascular health using individualized data streams. We present exemplars from the existing body of literature pertaining to mechanistic model development for cardiovascular dynamics and summarize existing technical challenges and opportunities pertaining to the foundation of a digital twin.

Introducing a new model for solid-state batteries: Parameter estimation and sensitivity analysis on diffusion, concentration, and electrochemical kinetics

Despite their potential, solid-state batteries (SSBs) are challenging to optimize due to the complexity of their materials and the early stage of their modeling, especially compared to liquid-electrolyte technologies. Many existing models rely on empirical equations that cannot represent internal mass-transport and kinetic phenomena. Even studies based on mechanistic models often focus on theoretical explanations of experimental phenomena or generating hard-to-obtain experimental data. In this work, a simple yet versatile mechanistic model – able to simulate any battery composed of a metallic anode, solid electrolyte and intercalation cathode – is proposed and used in a parameter estimation routine to identify the material properties of a Li/LiPON/LiCoO2 battery. After validation, a parametric study is made to address how each material property impacts concentration profiles, discharge curves, overpotentials, and core key-performance indexes (KPIs) related to energy efficiency and battery capacity. Findings reveal that cathode diffusion is critical to enhancing Extraction Efficiency, in some cases reaching 98.9% of the theoretical capacity without enhancing any other parameter. In the case of Energy Efficiency being a priority over capacity, electrolyte diffusion and concentration showed a pivotal role, in the range of parameters studied, to increase Voltaic Efficiency up to 97.8%. Despite not directly affecting the Extraction Efficiency, it is discussed how high overpotentials caused by Electrolyte Concentration and Kinetic Constants still can harm battery capacity, for charge–discharge cycles limited by boundaries in electrical potential. Finally, studies simultaneously varying two or more parameters show that, despite extreme property enhancements (up to ca 500x higher than the standard case) indeed lead to exceptional results (up to 96.4% for Voltaic Efficiency and 98.9% for Extraction Efficiency), to use such materials at its full potential will result in other zones of the device usually neglected – such as the overpotentials from the metal anode or even the current collector – may become the new stumbling block, underscoring the significance of considering the battery as a holistic system that will inherently exhibit bottlenecks. This also implies, however, in the perpetual potential for enhancement across its components.

Microbial-Guided prediction of methane and sulfide production in Sewers: Integrating mechanistic models with Machine learning

Accurate modeling of methane (CH4) and sulfide (H2S) production in sewer systems was constrained by insufficient consideration of microbial processes under dynamic environmental conditions. This study introduces a microbial-guided machine learning (ML) framework (Micro-ML), which integrates microbial process representations from mechanistic models (microbial information) with ML models. Results indicate that Micro-ML model enhanced predictions of CH4 and H2S production, where microbial information provides more information for model optimization. The feature importance of microbial information performed comparable weightings for 58.12 % and 55.16 %, respectively, but their relative significance in influencing Micro-ML model performance varies considerably. The application of Micro-ML performed great potential in reducing CH4 and H2S production (decreased ∼ 80 % and 90 %). The integrated model not only improves the accuracy of CH4 and H2S predictions but also offers a valuable tool for effective management strategies for sewer systems.

Analysis of resting status reveals distinct elevational variation in metabolisms of lizards.

Animals spend a considerable proportion of their life span at rest. However, resting status has often been overlooked when investigating how species respond to environmental conditions. This may induce a large bias in understanding the local adaptation of species across environmental gradients and their vulnerability to potential environmental change. Here, we conducted an empirical study on montane agamid lizards, combined with mechanistic modeling, to compare elevational variations in body temperature and metabolisms (cumulative digestion and maintenance cost) between resting and active status. Our study on three populations of an agamid lizard along an elevational gradient revealed a trend of decreasing body temperature toward higher elevations, the main contributor of which was resting status of the lizards. Using population-specific reaction norms, we predicted greater elevational variation in hourly and cumulative digestion for resting lizards than for active lizards. Climate-change impacts, estimated as the change in cumulative digestion, also show greater elevational variation when resting status is factored into the analysis. Further, our global analysis of 98 agamid species revealed that in about half of their combined distributional range, the contribution of resting status in determining the elevational variation in cumulative digestion and maintenance cost of lizards was greater than the contribution made by a lizard's active status. Our study highlights the importance of considering resting status when investigating how species respond to environmental conditions, especially for those distributed over tropical and subtropical mountain areas.

Computational Modelling of the Impact of Evaporation on In-Vitro Dermal Absorption

PurposeVolatiles are common in personal care products and dermatological drugs. Determining the impact of evaporation of volatiles on skin permeation is crucial to evaluate and understand their delivery, bioavailability, efficacy and safety. We aim to develop an in-silico model to simulate the impact of evaporation on the dermal absorption of volatiles.MethodThe evaporation of volatile permeants was modelled using vapour pressure as the main factor. This model considers evaporation as a passive diffusion process driven by the concentration gradient between the air-vehicle interface and the ambient environment. The evaporation model was then integrated with a previously published physiologically based pharmacokinetic (PBPK) model of skin permeation and compared with published in vitro permeation test data from the Cosmetics Europe ADME Task Force.ResultsThe evaporation-PBPK model shows improved predictions when evaporation is considered. In particular, good agreement has been obtained for the distributions in the evaporative loss, and the overall percutaneous absorption. The model is further compared with published in-silico models from the Cosmetics Europe ADME Task Force where favourable results are achieved.ConclusionThe evaporation of volatile permeants under finite dose in vitro permeation test conditions has been successfully predicted using a mechanistic model with the intrinsic volatility parameter vapour pressure. Integrating evaporation in PBPK modelling significantly improved the prediction of dermal delivery.

Time-varying stimuli that prolong IKK activation promote nuclear remodeling and mechanistic switching of NF-κB dynamics.

Temporal properties of molecules within signaling networks, such as sub-cellular changes in protein abundance, encode information that mediate cellular responses to stimuli. How dynamic signals relay and process information is a critical gap in understanding cellular behaviors. In this work, we investigate transmission of information about changing extracellular cytokine concentrations from receptor-level supramolecular assemblies of IκB kinases (IKK) downstream to the nuclear factor κB (NF-κB) transcription factor (TF). In a custom robot-controlled microfluidic cell culture, we simultaneously measure input-output (I/O) encoding of IKK-NF-κB in dual fluorescent-reporter cells. When compared with single cytokine pulses, dose-conserving pulse trains prolong IKK assemblies and lead to disproportionately enhanced retention of nuclear NF-κB. Using particle swarm optimization, we demonstrate that a mechanistic model does not recapitulate this emergent property. By contrast, invoking mechanisms for NF-κB-dependent chromatin remodeling to the model recapitulates experiments, showing how temporal dosing that prolongs IKK assemblies facilitates switching to permissive chromatin that sequesters nuclear NF-κB. Remarkably, using simulations to resolve single-cell receptor data accurately predicts same-cell NF-κB time courses for more than 80% of our single cell trajectories. Our data and simulations therefore suggest that cell-to-cell heterogeneity in cytokine responses are predominantly due to mechanisms at the level receptor-associated protein complexes.

Estimating Fishing Exploitation Rates to Simulate Global Catches and Biomass Changes of Pelagic and Demersal Fish

AbstractRobust projections of future trends in global fish biomass, production and catches are needed for informed fisheries policy in a changing climate. Trust in future projections, however, relies on establishing that models can accurately simulate past relationships between exploitation rates and ecosystem states. In addition, historical simulations are important to describe how the oceans have changed due to fishing. Here we use fisheries catch, catch‐only assessment models and effort data to estimate regional fishing exploitation levels, defined as the fishing mortality relative to fishing mortality at maximum sustainable yield (F/FMSY). These estimates are given for large pelagic, forage and demersal fish types across all large marine ecosystems and the high seas between 1961 and 2004; and with a ‘ramp‐up’ between 1841 and 1960. We find that global exploitation rates for large pelagic and demersal fish consistently exceed those for forage fish and peak in the late 1980s. We use the rates to globally simulate historical fishing patterns in a mechanistic fish community model. The modeled catch aligns with the reconstructed catch, both for total catch and catch distribution by functional type. Simulations show a clear deviation from an unfished model state, with a 25% reduction in biomass in large pelagic and demersal fish in shelf regions in recent years and a 50% increase in forage fish, primarily due to reduced predation. The simulations can set a baseline for assessing the effect of climate change relative to fishing. The results highlight the influential role of fishing as a primary driver of global fish community dynamics.

Phenology forecasting models for detection and management of invasive annual grasses

AbstractNon‐native annual grasses can dramatically alter fire frequency and reduce forage quality and biodiversity in the ecosystems they invade. Effective management techniques are needed to reduce these undesirable invasive species and maintain ecosystem services. Well‐timed management strategies, such as grazing, that are applied when invasive grasses are active prior to native plants can control invasive species spread and reduce their impact; however, anticipating the timing of key phenological stages that are susceptible to management over vast landscapes is difficult, as the phenology of these species can vary greatly over time and space. To address this challenge, we created range‐wide phenology forecasts for two problematic invasive annual grasses: cheatgrass (Bromus tectorum), and red brome (Bromus rubens). We tested a suite of 18 mechanistic phenology models using observations from monitoring experiments, volunteer science, herbarium records, timelapse camera imagery, and downscaled gridded climate data to identify the models that best predicted the dates of flowering and senescence of the two invasive grass species. We found that the timing of flowering and senescence of cheatgrass and red brome were best predicted by photothermal time models that had been adjusted for topography using gridded continuous heat‐insolation load index values. Phenology forecasts based on these models can help managers make decisions about when to schedule management actions such as grazing to reduce undesirable invasive grasses and promote forage production, quality, and biodiversity in grasslands; to predict the timing of greatest fire risk after annual grasses dry out; and to select remote sensing imagery to accurately map invasive grasses across topographic and latitudinal gradients. These phenology models also have the potential to be operationalized for within‐season or within‐year decision support.

Modeling modulated tool path turning using coupled smoothed particle hydrodynamics and finite element method

This paper describes a full-scale, three-dimensional coupled smoothed particle hydrodynamics (SPH) and finite element model for modulated tool path (MTP) turning. The chip breaking mechanism due to modulated motion of the tool is demonstrated by the developed machining model. In contrast, the simulation of conventional turning with the same machining conditions predicts long continuous chips. The cutting force predicted by the simulation is validated with a mechanistic force model based on the instantaneous chip thickness. This work expands the capabilities of machining simulations to predict complex machining phenomena such as MTP turning through a full-scale realistic simulation. The encouraging simulation results show the potential to study more complex phenomena, such as evaluating the parameters of tool path modulation, simulating ultrasonic machining, and studying machining stability.

Water and Thermal Management in PEM Fuel Cells Using Feasible Humidity Plots and Model Predictive Controllers

Water and thermal management are critical for the performance, efficiency and longevity of PEM fuel cells (PEMFCs). Effective water and thermal management require the design of control systems that can maintain the water balance and temperature at stable and optimal levels. In this paper, we consider a stack of PEM fuel cells integrated with a water recovery and cooling system. A mechanistic dynamic model is developed to be able to predict the water content and temperature in response to the fuel cell inputs. Water management uses a cascade arrangement of a supervisory Model Predictive Controller (MPC) and local anode and cathode PID humidity controllers to balance the membrane water content. Thermal management consists of a separate MPC controller to regulate the fuel stack temperature. One novelty of this work lies in identifying and utilizing the feasible region for the relative humidities of the anode and cathode when controlling the membrane water content. We introduce the feasible humidity plots (FHP) which define the feasible values for the anode and cathode relative humidities for a given fuel cell design and its operating conditions. This useful information helps to assign the set-point values to the local PID humidity controllers of the water management system. It is shown by simulations that the water and thermal management MPC controllers work in tandem and successfully track the desired set-point changes in humidity and temperature while rejecting external disturbances such as load changes. In addition, the control system is robust against modeling errors and possible model-plant mismatch introduced by fuel cell aging.

Dynamic entrainment: A deep learning and data-driven process approach for synchronization in the Hodgkin-Huxley model.

Resonance and synchronized rhythm are significant phenomena observed in dynamical systems in nature, particularly in biological contexts. These phenomena can either enhance or disrupt system functioning. Numerous examples illustrate the necessity for organs within the human body to maintain their rhythmic patterns for proper operation. For instance, in the brain, synchronized or desynchronized electrical activities can contribute to neurodegenerative conditions like Huntington's disease. In this paper, we utilize the well-established Hodgkin-Huxley (HH) model, which describes the propagation of action potentials in neurons through conductance-based mechanisms. Employing a "data-driven" approach alongside the outputs of the HH model, we introduce an innovative technique termed "dynamic entrainment." This technique leverages deep learning methodologies to dynamically sustain the system within its entrainment regime. Our findings show that the results of the dynamic entrainment technique match with the outputs of the mechanistic (HH) model.