Simple Model Research Articles

How representative are air transport functional complex networks? A quantitative validation.

Functional networks have emerged as powerful instruments to characterize the propagation of information in complex systems, with applications ranging from neuroscience to climate and air transport. In spite of their success, reliable methods for validating the resulting structures are still missing, forcing the community to resort to expert knowledge or simplified models of the system's dynamics. We here propose the use of a real-world problem, involving the reconstruction of the structure of flights in the US air transport system from the activity of individual airports, as a way to explore the limits of such an approach. While the true connectivity is known and is, therefore, possible to provide a quantitative benchmark, this problem presents challenges commonly found in other fields, including the presence of non-stationarities and observational noise, and the limitedness of available time series. We explore the impact of elements like the specific functional metric employed, the way of detrending the time series, or the size of the reconstructed system and discuss how the conclusions here drawn could have implications for similar analyses in neuroscience.

Virtual Hydrological Laboratories: Developing the Next Generation of Conceptual Models to Support Decision Making Under Change

AbstractAs hydrological systems are pushed outside the envelope of historical experience, the ability of current hydrological models to serve as a basis for credible prediction and decision making is increasingly challenged. Conceptual models are the most common type of surface water hydrological model used for decision support due to reasonable performance in the absence of change, ease of use and computational speed that facilitate scenario, sensitivity and uncertainty analysis. Hence, conceptual models in effect represent the current “shopfront” of hydrological science as seen by practitioners. However, these models have notable limitations in their ability to resolve internal catchment processes and subsequently capture hydrological change. New thinking is needed to confront the challenges faced by the current generation of conceptual models in dealing with a changing environment. We argue the next generation of conceptual models should combine the parsimony of conceptual models with our best available scientific understanding. We propose a strategy to develop such models using multiple hydrological lines of evidence. This strategy includes using appropriately selected physically resolved models as “Virtual Hydrological Laboratories” to test and refine the simpler models' ability to predict future hydrological changes. This approach moves beyond the sole focus on “predictive skill” measured using metrics of historical performance, facilitating the development of the next generation of conceptual models with hydrological fidelity (i.e., models that “get the right answers for the right reasons”). This quest is more than a scientific curiosity; it is expected by policy makers who need to know what to plan for.

Study on radon release behaviors of the tubular flow-through radon source with natural radium-bearing materials

To obtain a radon source that is easy to access natural radium-bearing raw material and does not require the approval of relevant departments, the tubular flow-through radon source using radium-bearing ore as the radon-generating material was designed. Based on the theories of radon diffusion-convection migration in porous media, the simplified calculation formula and numerical calculation models of radon activity concentration at the outlet of the tubular flow-through radon source were established, and the radon release rules of the tubular flow-through radon source were studied using a combination of simplified calculation formulas, numerical simulations, and experiments, respectively. The results show that (1) the radon activity concentration at the outlet of the tubular flow-through radon source was dominated by mass (or the volume) of the natural radium-bearing raw material, free radon production rate, and gas flow rate. (2) When the range of radon activity concentration at the outlet is 260–1308 Bq m-3, the relative deviation of the three methods is less than 2.62%. (3) The simplified calculation formula and numerical simulation method proposed in this paper can provide a reference for the design and control of the radon activity concentration of a tubular flow-through radon source.

Stability analysis of heavy machinery moving on weak subsoil. 3D FEM model vs. analytical models

In this paper, the authors present an extension of the scope of the previously conducted research to the full three-dimensional computer simulation (using the finite element method), which takes into account the interaction between: heavy caterpillar tracks system – working platform – weak subsoil. The article presents a computer model considering two caterpillars, resting on elastic-plastic sub-soil, with standard Mohr-Coulomb yield conditions, allowing for computer simulation of the behavior of the system up to achievement of ultimate limit state. The results of the above model are treated as the reference for a simplified Analytical Models of estimating the limit state, which might be used in design procedures. In turn, these Analytical Models are enhancements of previously presented one. The most important results concluding form the Analytical Model are simple interaction formulas, in the space of moments acting on the machine-subsoil system, limiting a domain of safety in given soil conditions.

Effect of HPO42− and brushite on gypsum reactivity and implications for utilization of phosphogypsum in plaster production

The implementation of a sustainable development strategy holds the potential to enhance productivity, profitability, and overall efficiency by leveraging by-products from other industries. In plaster manufacturing the substitution of natural gypsum with synthetic gypsum is worth exploring. Phosphogypsum (PG) represents a significant fraction of synthetic gypsum production, arising as a by-product of phosphoric acid manufacturing. However, the presence of impurities in PG, particularly phosphorus, poses challenges to its use as a plaster material. To identify the phases present in PG, quantify their occurrences and understand their effects on dehydration and hydration processes, our study aimed to examine simplified models. These models are specifically designed to provide insights into the underlying mechanisms governing the hydration reactivity of hemihydrate. Pure gypsum and brushite were synthesized, and mechanical mixing was carried out to explore how brushite affects the characteristics of gypsum. Additionally, a solid solution of Ca(SO4)1-x(HPO4)x·2H2O, where 0 < x < 1, was prepared to characterize and quantify the syncrystallized HPO42− as well as the effects of this impurity on typical gypsum use cases. The synthesized materials underwent physical and chemical characterization using SEM, XRD, IR spectroscopy, DSC, pH and conductivity measurements, and ion chromatography. The results demonstrated that the HPO42− ions can substitute for sulfate ions in gypsum to form solid solutions. The maximum quantity of HPO42− ions in the gypsum lattice is approximately 10%, leading to the crystallization of gypsum into sand rose shape rather than needle crystals. As the concentration of HPO42− ions increases, a new phase known as ardealite (Ca(SO4)1-x(HPO4)x·2H2O) with 0.42 < x < 0.54 emerges, followed by brushite (CaHPO4·2H2O). During the calcination of brushite at 160 °C, a mixture of brushite and monetite formed that exhibits different properties compared to its natural counterpart. In the presence of syncrystallized HPO42− ions, the transformation from anhydrite III to anhydrite II occurred at higher temperatures than for pure gypsum. The reactivity of calcined samples indicate that HPO42− ions from the dehydration products of brushite caused a delay in hydration, with the maximum delay observed at a calcination temperature of 160 °C. At this temperature, monetite begins to crystallize, and it is characterized by its inert nature, exhibiting no reactivity with water. This explains the observed decrease in setting times as the calcination temperature rises. Moreover, syncrystallized HPO42− ions induced a significant delay in the hydration process. Our research revealed that adjusting the calcination temperature can mitigate the retarding effect associated with this impurity, suggesting an industrial tendency to use the lowest possible temperature during the calcination process of PG containing syncrystallized phosphate impurity. In addition, we have observed that adjusting the pH of the mixing water to lower values can significantly accelerate setting times.

First experimental time-of-flight-based proton radiography using low gain avalanche diodes

Objective. Ion computed tomography (iCT) is an imaging modality for the direct determination of the relative stopping power (RSP) distribution within a patient’s body. Usually, this is done by estimating the path and energy loss of ions traversing the scanned volume utilising a tracking system and a separate residual energy detector. This study, on the other hand, introduces the first experimental study of a novel iCT approach based on time-of-flight (TOF) measurements, the so-called Sandwich TOF-iCT concept, which in contrast to any other iCT systems, does not require a residual energy detector for the RSP determination. Approach. A small Sandwich TOF-iCT demonstrator was built based on low gain avalanche diodes (LGADs), which are 4D-tracking detectors that allow to simultaneously measure the particle position and time-of-arrival with a precision better than 100 μm and 100 ps, respectively. Using this demonstrator, the material and energy-dependent TOF was measured for several homogeneous PMMA slabs in order to calibrate the acquired TOF against the corresponding water equivalent thickness (WET). With this calibration, two proton radiographs (pRads) of a small aluminium stair phantom were recorded at MedAustron using 83 MeV and 100.4 MeV protons. Main results. Due to the simplified WET calibration models used in this very first experimental study of this novel approach, the difference between the measured and theoretical WET ranged between 37.09% and 51.12%. Nevertheless, the first TOF-based pRad was successfully recorded showing that LGADs are suitable detector candidates for Sandwich TOF-iCT. Significance. While the system parameters and WET estimation algorithms require further optimization, this work was an important first step to realize Sandwich TOF-iCT. Due to its compact and cost-efficient design, Sandwich TOF-iCT has the potential to make iCT more feasible and attractive for clinical application, which, eventually, could enhance the treatment planning quality.

Are the Italian microzonation level 2 abacuses applicable in the Friuli Venezia Giulia (Italy) plain? Comparison between the national abacuses and the numerically simulated amplification factors and between the derived elastic response spectra

Seismic shaking of an area is strongly affected by the local geology. The so-called local site effects must be considered for the estimation of seismic effects on structures and urban planning. Thus, the seismic microzonation is the process aimed at identifying and mapping the subsoil local response in a given area, usually at urban/municipality scale and in terms of ground shaking parameters and susceptibility to ground instabilities. In Italy, for areas that can be schematised as a 1D subsoil model (e.g. alluvial plain), a simplified approach is proposed to quantify the seismic amplification (amplification factor, AF). This approach consists of tables of correspondences, called seismic abacuses, available for the whole national area as well as for some regional territories, and derived for simplified subsoil models. In this work, the results of the comparison between the AF values retrieved from national abacuses applied in the Friuli Venezia Giulia (Italy) plain municipalities and those from 1D numerical simulations are presented. In general, the abacuses underestimate the local seismic site effects a part for sites with a shallow bedrock. No correlations/trends were identified between the AF derived from abacuses and those from numerical simulations. Moreover, considering the elastic acceleration response spectra, it emerges that in the 49.5% of the FVG analysed sites the abacuses approach, even though it underestimates the real seismic response, is a more suitable approximation compared to the soil class simplified approach proposed by the Italian regulation. Finally, what emerges is that the limit of 30 m, as indicated in the Italian regulation, to consider a deep or shallow bedrock seams underestimated, and the AFs are not correlated with the seismic bedrock depth when it is higher than 100 m.

Adsorption of an antiseptic in a functionalized fixed-bed: Analysis of breakthrough scenarios and validation of simplistic models defending a novel proposition

BackgroundChlorhexidine gluconate (CHG) and cetrimide, which are widely used in various pharmaceutical compositions, are considered potentially hazardous compounds. This combination was largely used during and after Covid 19 pandemic for sanitization. Removal of these two compounds from pharmaceutical waste-water with commercial and functionalized activated carbon in a packed bed column is reported. MethodsEffects of changes in bed height, flow rate, and initial concentration on the performance of the packed bed are analyzed using Yoon-Nelson, BDST and Thomas models for commercial scale-up operation. The effects of primary design parameters like bed depth and operating parameters like inflow rate and inlet concentration of influent wastewater are studied on the extent of removal of cetrimide and chlorhexidine gluconate. Granular activated carbon (GAC) is functionalized using HF and NH4OH. The extent of enhanced adsorption using the functionalized GAC is demonstrated using breakthrough curves. Significant findingsK. H. Chu's iconic proposition is validated. Breakthrough time (BT) increases with bed heights and it is less in the case of cetrimide as compared to chlorhexidine gluconate. This shows that cetrimide wins in the competition and occupies the pores much faster than CHG.Mostly, BT-CHG (GAC) < BT-CHG (FAC-HF) < BT-CHG (FAC-NH3) and BT-cetrimide (GAC) < BT-cetrimide (FAC-NH3) < BT-cetrimide (FAC-HF) for a particular bed height.BT−CHG(FAC−HF)BT−cetrimide(FAC−HF)<BT−CHG(FAC−NH3)BT−cetrimide(FAC−NH3)<BT−CHG(GAC)BT−cetrimide(GAC)

Relativistic energy density functional from momentum space to coordinate space within a coherent density fluctuation model

In this theoretical study, we have derived a simplified analytical expression for the binding energy per nucleon as a function of density and isospin asymmetry within the relativistic mean-field model. We have generated a new parameterization for the density-dependent DD-ME2 parameter set using the Relativistic-Hartree-Bogoliubov approach. Moreover, this work attempts to revisit the prior polynomial fitting in Kumar A. et al. Phys. Rev. C, 103 (2021) 024305 for the non-linear NL3 force parameter to provide a simplified set of equations for the energy density functional which is used for calculating the surface properties of finite nuclei. The current study improves the existing fitting procedure by effectively proposing a simpler model that provides comparably precise results while lowering the computational expense. To study the surface properties of finite nuclei with these parameterizations, we have adopted the coherent density fluctuation model, which effectively translates the quantities of nuclear matter from momentum space to coordinate space at local density. The isospin properties, such as symmetry energy and its surface and volume components, slope parameter, finite nuclear incompressibility, and surface incompressibility for even-even nuclei, are calculated for different mass regions. Moreover, we have studied the effect of density, weight function, and choice of relativistic force parameters on the surface properties. The significance of this work will help to determine the properties of nuclei along the nuclear landscape and can facilitate an improved understanding of the island of stability, heavy-ion collision, and nucleosynthesis, among others.

Metaheuristic optimized time–frequency features for enhancing Alzheimer’s disease identification

Background:Alzheimer’s disease (AD) is a chronic disorder characterized by progressive cognitive dysfunctions and memory loss. Electroencephalography (EEG) is a non-invasive tool to detect AD using machine learning models and signal-derived features. Feature selection enhances AD detection performance, reduces training and testing time, and builds simpler models. Metaheuristic algorithms (MHAs) have successfully selected optimum features in several classification tasks and can be valuable in enhancing AD detection. Method:In this work, an AD detection model is proposed based on a systematic investigation of adaptive signal decomposition techniques and MHAs. The four techniques comprising empirical mode decomposition, variational mode decomposition, discrete wavelet transform, and Low-Complexity Orthogonal Wavelet Filter Banks (LCOWFBs) decompose EEG signals into subbands. Thirty-four features based on signal, temporal, spectral, and entropy analysis are extracted for each subband. The salient features are selected using seven different MHAs: particle swarm algorithm, salp swarm algorithm, reptile search (RSA), Grey Wolf, dwarf mongoose optimization, snake optimizer, and Fick’s law algorithm (FLA). Results:The two publicly available AD EEG datasets validated the performance of the proposed framework. LCOWFBs achieved the highest binary classification accuracy of 99.72% using 15 salient features selected by RSA and multi-class accuracy of 88.92% with seven salient features selected by FLA. The present method investigates brain regions that are affected by AD. Conclusion:The results indicate that the combination of LCOWFBs and RSA achieved maximum AD detection performance for both datasets with shorter computational time than earlier reported methods in the literature. The present model can be extended to identify chronic disorders, including hypertension, sleep disorders, and Parkinson’s disease.

Influence of modelling hypotheses on strength assessment of CFRP stepped repairs

Composite stepped repairs can achieve high strength recovery without the addition of bolts or fasteners to the structure. They are therefore a major issue in the field of aerospace composite structure damage repair. However, there is no standardized method to design this type of repairs. Many analytical, semi-analytical and finite element models were proposed throughout the years to predict the strength of stepped repairs, using various hypotheses and simplifications. The aim of this paper is to investigate the influence of modelling hypotheses on stress distribution and strength prediction of composite stepped repairs. Five simplified stepped joint models using macro-element (ME) modelling and finite element method (FE) are compared to a full 3D FE model of a stepped repaired panel. The influence of step length and adhesive fracture toughness was investigated to determine the field of validity of each model. Among FE models, it was shown that modelling the equivalent joint under 2D generalized plain strain gives a very close strength prediction to the 3D stepped repair model while saving computation time. Simplified macro-element models under bar or beam hypotheses are fairly close to the results of FE modelling, but the deviation between those and FE is sensitive to step length and adhesive fracture toughness.

White dwarf cooling in f(R,T) gravity

In recent times, astounding observations of both over- and under-luminous type Ia supernovae have emerged. These peculiar observations hint not only at surpassing the Chandrasekhar limit but may also suggest potential modifications in the physical attributes of their progenitors such as their cooling rate. This, in turn, can influence their temporal assessments and provide a compelling explanation for these intriguing observations. In this spirit, we investigate here the cooling process of white dwarfs in [Formula: see text] gravity with the simplest model [Formula: see text], where [Formula: see text] is the model parameter. Our modeling suggests that the cooling timescale of white dwarfs exhibits an inverse relationship with the model parameter [Formula: see text]. This unveils that in the realm of [Formula: see text] gravity, the energy release rate for white dwarfs increases as [Formula: see text] increases. Furthermore, we also report that the luminosity of the white dwarfs also depends on [Formula: see text] and an upswing in [Formula: see text] leads to an amplification in the luminosity. As a result, utilizing white dwarf luminosity could possibly define bounds on [Formula: see text] gravity models.

Influence of sedimentary architecture on static connectivity and geothermal doublet performance (Mezőberény, SE Hungary)

Although the Pannonian Basin in Hungary boasts significant geothermal potential, most geothermal wells in the region suffer from reinjection issues. The aim of this study is to examine the impact of the sedimentary architecture of a paralic lacustrine succession on the static connectivity of a geothermal aquifer. This problem was examined for a case-study geothermal doublet drilled in Mezőberény, southeast Hungary, which experienced injectivity problems shortly after commencing operations in 2012, ultimately leading to shut-in of the wells. A characterization of the architecture of the deltaic aquifer has been undertaken through a novel workflow that combines sedimentological and sequence stratigraphic analysis of the well-doublet and stochastic 3D modelling techniques to assess the likely static connectivity between the production and injection wells. Planform dimension of aquifer-forming mouth-bar elements and fluvial channel bodies were estimated from geological analogues, selected based on the thickness values of such sedimentary bodies interpreted on the gamma-ray logs of the well doublet and nearby hydrocarbon wells. In total five mouth-bar and two fluvial channel input geometry groups were selected, and fifty object-based realizations were constructed to analyse the static connectivity of different geological scenarios. The internal lithology of mouth-bar and fluvial-channel elements were modelled by sequential indicator simulations (ten realizations). Results of connected volume analysis indicate that the sandstone content of the aquifer is located within vertically segmented fluvial-channel and mouth-bar elements separated by mud-prone prodelta and lower delta-front deposits. This architectural configuration, in combination with low net-sandstone content suggests limited connectivity between the wells. Across the set of stochastic realizations, the volume of sandstone connecting the injection and production wells is ∼4% of the total aquifer volume, on average. Results of this study shows that simplified reservoir models can lead to over-estimation of connectivity of geothermal doublets. The proposed workflow combines sedimentological analysis and 3D modelling supported by geological analogues to predict sandstone reservoir connectivity of deltaic geothermal reservoirs. The described workflow can be carried out in target areas using existing boreholes to assist borehole planning.

Intersecções entre o Eficientismo Processual Penal e o Neoliberalismo

This article aims to address the meta-legal relations between the management of the speed of the criminal process and the social constraints arising from neoliberal reason. To this end, it is based on the premise established by Foucault and deepened by Dardot and Laval that neoliberalism has become socially rooted to the point of transmuting from a simple economic model to a rationality that spreads to the most diverse social fields. The work is guided by the question: are there harmful effects beyond the loss of quality in the motivation of decisions with the adoption of efficiency logic by the criminal judge? The objective is to verify whether the management principles of this reason may be harmful to the proper functioning of the criminal process. This is an exploratory, interdisciplinary research, based exclusively on bibliographical references. The conclusion is that there are no conditions for intensifying the pace of the criminal process without a loss of quality that is not only technical, but also democratic, since acceleration without other measures ends up catalyzing authoritarian clots that have never been completely expunged from the Brazilian system of criminal procedure.

Benchmarking machine learning-based real-time respiratory signal predictors in 4D SBRT.

Stereotactic body radiotherapy of thoracic and abdominal tumors has to account for respiratory intrafractional tumor motion. Commonly, an external breathing signal is continuously acquired that serves as a surrogate of the tumor motion and forms the basis of strategies like breathing-guided imaging and gated dose delivery. However, due to inherent system latencies, there exists a temporal lag between the acquired respiratory signal and the system response. Respiratory signal prediction models aim to compensate for the time delays and to improve imaging and dosedelivery. The present study explores and compares six state-of-the-art machine and deep learning-based prediction models, focusing on real-time and real-world applicability. All models and data are provided as open source and data to ensure reproducibility of the results and fosterreuse. The study was based on 2502 breathing signals ( h) acquired during clinical routine, split into independent training (50%), validation (20%), and test sets (30%). Input signal values were sampled from noisy signals, and the target signal values were selected from corresponding denoised signals. A standard linear prediction model (Linear), two state-of-the-art models in general univariate signal prediction (Dlinear, Xgboost), and three deep learning models (Lstm, Trans-Enc, Trans-TSF) were chosen. The prediction performance was evaluated for three different prediction horizons (480, 680, and 920ms). Moreover, the robustness of the different models when applied to atypical, that is, out-of-distribution (OOD) signals, was analyzed. The Lstm model achieved the lowest normalized root mean square error for all prediction horizons. The prediction errors only slightly increased for longer horizons. However, a substantial spread of the error values across the test signals was observed. Compared to typical, that is, in-distribution test signals, the prediction accuracy of all models decreased when applied to OOD signals. The more complex deep learning models Lstm and Trans-Enc showed the least performance loss, while the performance of simpler models like Linear dropped the most. Except for Trans-Enc, inference times for the different models allowed for real-time application. The application of the Lstm model achieved the lowest prediction errors. Simpler prediction filters suffer from limited signal history access, resulting in a drop in performance for OODsignals.

A rapid method for turbomachinery aerodynamic design and optimization using active subspace

Preliminary design is a crucial step in the turbomachinery aerodynamic design process, which involves the initial conceptualization and serves as the foundation for subsequent detailed design and analysis. During preliminary design, designers typically use a combination of empirical correlations, simplified flow models, and computational tools to evaluate the impact of different design parameters on the machine’s performance. This iterative process heavily relies on the designer’s experience and knowledge, and errors made during preliminary design may not be rectifiable in later design phases. To address this challenge, this paper proposes a novel optimization strategy for turbomachinery aerodynamic preliminary design using active subspace. The method utilizes active subspace techniques to reduce the dimensionality of the design space, making efficient and accurate exploration of the space possible. In addition, the open-access &lt;italic&gt;Multall&lt;/italic&gt; turbomachinery design suite is employed to evaluate the aero-thermal performance of different designs. The method was then utilized to optimize the aerodynamic design of the blade profile for a ducted fan. Results show that the optimization strategy can identify the most significant directions in the design space, providing designers with a clear path to achieving an optimal design. Overall, the proposed method can be of good value to the turbomachinery design community attribute to its reduced dependence on the designer’s experience and elevated efficiency of the design process.

Approaches to simplify industrial energy models for operational optimisation

Minimising energy consumption in today's industrial sector is a crucial objective for achieving established climate goals. One effective strategy to enhance efficiency is optimising energy system operations within industries. The initial step in establishing operational optimisation involves developing a comprehensive model of the energy system. This model necessitates a specific structure to meet optimisation requirements. However, creating a model from scratch incurs substantial effort. While numerous companies possess energy models, they often lack the requisite structure for optimisation. Consequently, simplifying existing models can significantly reduce the effort needed to implement operational optimisation.This paper investigates the simplification of intricate industrial energy system models for optimisation purposes. The subsequent sections analyse two distinct approaches. One approach involves linearisation, while the other utilises neural networks. To facilitate a comparative analysis of these approaches, a reference model is developed. The assessment of these methodologies includes an investigation into optimisation robustness, computation time, accuracy concerning the reference model, and the effort required for developing and maintaining the simplified models. It proved that both approaches are suitable for operational optimisation. Linearisation exhibits superior computational efficiency compared to the neural network approach. The linearisation modelling approach together with the optimisation only required a few milliseconds for the calculation. The neural network approach needed 3 h for the calculation of the optimum with the genetic algorithm. The simulation of the neural network itself only required a few milliseconds. Hence, an improvement of the genetic algorithm is needed. However, the accuracy of linearisation falls short of that achieved by neural networks. The linearisation achieves a mean average percentage error from only 13%. In comparison the neural network's mean average error is 2.3%. Therefore, the linearisation must be improved. The impact using a piecewise linearisation on the results will be analysed in further research.

A framework for data regression of heat transfer data using machine learning

Machine Learning (ML) algorithms are emerging in various industries as a powerful complement/alternative to traditional data regression methods. A major reason is that, unlike deterministic models, they can be used even in the absence of detailed phenomenological knowledge. Not surprisingly, the use of ML algorithms is being explored also in heat transfer applications. It is of particular interest in systems dealing with complex geometries and underlying phenomena (e.g. fluid phase change, multi-phase flow, heavy fouling build-up). However, heat transfer systems present specific challenges that need addressing, such as the scarcity of high-quality data, the inconsistencies across published data sources, the complex (and often correlated) influence of inputs, the split of data between training and testing sets, and the limited extrapolation capabilities to unseen conditions. In an attempt to help overcome some of these challenges and, more importantly, to provide a systematic approach, this article reviews and analyses past efforts in the application of ML algorithms to heat transfer applications, and proposes a regression framework for their deployment to estimate key quantities (e.g. heat transfer coefficient), to be used for improved design and operation of heat exchangers. The framework consists of six steps: i) data pre-treatment, ii) feature selection, iii) data splitting philosophy, iv) training and testing, v) tuning of hyperparameters, and vi) performance assessment with specific indicators, to support the choice of accurate and robust models. A relevant case study involving the estimation of the condensation heat transfer coefficient in microfin tubes is used to illustrate the proposed framework. Two data-driven algorithms, Deep Neural Networks and Random Forest, are tested and compared in terms of their estimation and extrapolation capabilities. The results show that ML algorithms are generally more accurate in predicting the heat transfer coefficient than a well-known semi-empirical correlation proposed in past studies, where the mean absolute error of the most suitable ML model is 535 [Wm2K-1], compared to the error using the correlation of 1061 [Wm2K-1]. In terms of extrapolation, the selected ML model has a mean absolute error of 1819 [Wm2K-1], while for the correlation is 1111 [Wm2K-1], indicating a disadvantage of the use of semi-empirical models, although the comparison was not entirely suitable, given that the correlation was used as is and no training was done. In addition, feature selection enables simpler models that depend only on features that are potentially most related to the target variable. Special attention is needed however, as overfitting and limited extrapolation capabilities are common difficulties that are encountered when deploying these models.

Simplifying Complex Observation Models in Continuous POMDP Planning with Probabilistic Guarantees and Practice

Solving partially observable Markov decision processes (POMDPs) with high dimensional and continuous observations, such as camera images, is required for many real life robotics and planning problems. Recent researches suggested machine learned probabilistic models as observation models, but their use is currently too computationally expensive for online deployment. We deal with the question of what would be the implication of using simplified observation models for planning, while retaining formal guarantees on the quality of the solution. Our main contribution is a novel probabilistic bound based on a statistical total variation distance of the simplified model. We show that it bounds the theoretical POMDP value w.r.t. original model, from the empirical planned value with the simplified model, by generalizing recent results of particle-belief MDP concentration bounds. Our calculations can be separated into offline and online parts, and we arrive at formal guarantees without having to access the costly model at all during planning, which is also a novel result. Finally, we demonstrate in simulation how to integrate the bound into the routine of an existing continuous online POMDP solver.

A Pre-convolved Representation for Plug-and-Play Neural Illumination Fields

Recent advances in implicit neural representation have demonstrated the ability to recover detailed geometry and material from multi-view images. However, the use of simplified lighting models such as environment maps to represent non-distant illumination, or using a network to fit indirect light modeling without a solid basis, can lead to an undesirable decomposition between lighting and material. To address this, we propose a fully differentiable framework named Neural Illumination Fields (NeIF) that uses radiance fields as a lighting model to handle complex lighting in a physically based way. Together with integral lobe encoding for roughness-adaptive specular lobe and leveraging the pre-convolved background for accurate decomposition, the proposed method represents a significant step towards integrating physically based rendering into the NeRF representation. The experiments demonstrate the superior performance of novel-view rendering compared to previous works, and the capability to re-render objects under arbitrary NeRF-style environments opens up exciting possibilities for bridging the gap between virtual and real-world scenes.