Input Uncertainty Research Articles

Capturing the attenuating influence of a Three Gorges Dam on downstream sediment loads using artificial neural networks

ABSTRACT River sediment dynamics have been significantly altered by dam constructions, with the Three Gorges Dam (TGD) on the Yangtze River being a prominent example. This study aims to quantitatively assess the attenuating impact of TGD operations on downstream sediment discharge using artificial neural network (ANN) models. Seven ANN models were developed for stations along the middle and lower Yangtze River, trained on measured water and sediment data. The models effectively captured the relationships between upstream inputs and downstream outputs. Results revealed that variations in sediment discharge at each station were predominantly influenced by changes in sediment input from its immediate upstream section, rather than by changes in water discharge. This suggests that the TGD's impact on downstream sediment transport was primarily due to its sediment trapping efficiency, rather than alterations in the hydrological regime. The influence diminished gradually farther downstream, attributed to tributary contributions and sediment transport dynamics. ANNs, adept at handling input uncertainties, underscore the importance of boundary conditions; the integration of data-driven and physics-based models illuminates sediment dynamics in regulated rivers. This approach provides insight into dam management implications and sediment transport processes.

On application of the relative entropy concept in reliability assessment of some engineering cable structures

The main research problem studied in this work is an uncertain response and reliability assessment of the spatial cable structures due to the environmental stochasticity as well as material and geometrical imperfections. Some popular cable structures are analyzed for this purpose using the Stochastic Finite Element Method (SFEM) implemented with the use of three different techniques, namely the iterative generalized perturbation method, semi-analytical approach as well as the Monte-Carlo simulation. Uncertainty quantification delivered in this study is based on the series of FEM analyses of both static and dynamic structural problems. They enable the Least Squares Method determination of the structural polynomial responses linking extreme stresses and deformations with several uncorrelated uncertainty sources. Reliability assessment, fundamental in durability and Structural Health Monitoring, is completed using a comparison of the First Order Reliability Method (FORM) with probabilistic distance formulated by Bhattacharyya. Input uncertainties are assumed to be Gaussian according to the Maximum Entropy Principle. They have specific expected values following engineering design demands or the provisions of designing codes, whereas their standard deviations do not exceed the 10% level. The methods presented and the results obtained in this study may serve for further reliability analyses of large-scale civil engineering structures completed with both steel cables and also reinforced concrete plates like suspended bridges, for instance.

A multi-scale uncertainty quantification model encompassing minimum-size unit cells for effective properties of plain woven composites

The uncertainty quantification is crucial to the high-precision prediction of composites’ effective properties. However, the unclear input uncertainties of multiscale parameters, complex uncertainty propagations of inter-scale correlations, and unaffordable computational cost of massive simulations are three primary problems at present. In this work, an innovative model with high accuracy and feasible cost for the mechanical property prediction of plain woven composites is developed encompassing minimum-size unit cells and multi-scale uncertainty quantification. For the accuracy holding, an uncertainty analysis process consists of the traceability description, inter-scale propagation and quantification is established. The uncertainties of geometry are described by uniform distributions for fiber, fiber bundle and composite scales, respectively; that of constituent properties is described by normal distributions for fiber and matrix, and its propagations to bundle and composite scales are realized by Nataf transformation methods with the consideration of parameter correlations. For the cost control, minimum-size unit cells are formulated by exhaustive analysis of structural symmetries to reduce the computational cost without accuracy compromising for the single simulation, and 1/8 and 1/16 unit cells compared with traditional full-size ones are obtained. The evolution convergence for statistical uncertainties of effective properties is finally obtained with totally reduced computational cost of 89%.

Enhancing Outcomes in Opioid Use Disorder Treatment: An Economic Evaluation of Improving Medication Adherence for Buprenorphine Through Blister-Packaging.

The opioid epidemic has severely impacted the US over the last 15 years. Buprenorphine is a partial opioid agonist indicated for the treatment of opioid use disorder (OUD) and is recognized as an effective treatment when taken as prescribed. However, adherence rates have been low in real-world settings. Blister-packaging has been shown to promote medication adherence across a variety of disease states, although it has never been studied in OUD. An economic analysis was conducted to assess the impact of increased adherence of blister-packaged buprenorphine on health care resource utilization (HCRU) and health care costs for 10,000 patients initiating therapy for OUD. The model analyzed a commercially insured population within the US over a one-year time horizon. Medication adherence was defined in the model as proportion of days covered (PDC) of at least 80%. Literature-based references were used to inform both the impact of blister-packaging on the number of patients who became adherent as well as the impact of medication adherence on HCRU and health care costs. Model input uncertainty was assessed in one-way sensitivity analyses. With the implementation of blister-packaging buprenorphine, adherence rates increased from 37.1% of patients in the pre-intervention period to 45.3%, resulting in an additional 818 patients becoming adherent post-intervention. The increase in adherence led to a reduction of medical costs of $12,138,757 (-$1,214 per-patient (PP)). Specifically, inpatient costs decreased by $7,127,073 (-$713 PP) while outpatient costs decreased by $5,013,319 (-$501 PP). Pharmacy costs increased by $3,432,705 ($343 PP). Despite the increase in pharmacy costs, total health care costs saw a reduction of $8,559,684 (-$856 PP). Blister-packaging buprenorphine for treatment of OUD has potential to improve medication adherence and health outcomes while reducing HCRU and health care costs. Future studies are necessary to assess the real-world application and impact of blister-packaging buprenorphine for OUD across various patient populations and health care settings.

Potential to Identify the Neutrino Mass Ordering with Reactor Antineutrinos in JUNO

Abstract The Jiangmen Underground Neutrino Observatory (JUNO) is a multi-purpose neutrino experiment under construction in South of China. This paper presents an updated estimate of JUNO's sensitivity to the neutrino mass ordering using the reactor antineutrinos emitted from eight nuclear reactor cores in the Taishan and Yangjiang nuclear power plants. This measurement is planned by studying the fine interference pattern caused by quasi-vacuum oscillations in the oscillated antineutrino spectrum at a baseline of 52.5~km and is completely independent of the CP violating phase and the neutrino mixing angle θ23. The sensitivity is obtained through a joint analysis of JUNO and TAO detectors utilizing the best available knowledge to date about the location and overburden of the JUNO experimental site, the local and global nuclear reactors, the JUNO and TAO detectors responses, the expected event rates and spectra of signal and backgrounds, and the systematic uncertainties of the analysis inputs. It is found that a 3σ median sensitivity to reject the wrong mass ordering hypothesis can be reached with an exposure of about 6.5 years × 26.6~GW thermal power.Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. Article funded by SCOAP3 and published under licence by Chinese Physical Society and the Institute of High Energy Physics of the Chinese Academy of Science and the Institute of Modern Physics of the Chinese Academy of Sciences and IOP Publishing Ltd.

DUAL LOOP VOLTAGE DROOP REGULATED CONTROLLER FOR DC MICROGRID USING HYBRID PSO AND GGO ALGORITHMS

Abstract DC microgrids are effectively employed in distribution network integration with renewable energy sources. The droop control technique is commonly utilized for DC microgrid regulation during load sharing. It minimizes the potential instability caused by disturbances such as input voltage changes, uncertainty parameters, and constant power load (CPL). Nevertheless, conventional droop control is inadequate in achieving precise current and satisfactory voltage regulation distribution for load sharing. The proposed approach comprises two separate loops for the DC bus voltage regulation and load current distribution, delivering a CVL and CPL to overcome the single optimization voltage controller technique. The primary loop uses the PSO algorithm to precisely manage the current load sharing among two parallel converters. Due to this the instability problems arising from disturbances, and it mitigated by the proposed topology. The multiobjective optimization technique provides notable resilience, rapid dynamic response, and robust stability over considerable variations in load. The secondary loop utilises the GGO algorithm to optimize the parameter to minimize the DC bus voltage regulation, voltage management, reasonable distribution of load power, and suitable reliability. The performance of the voltage regulation enhance & settling time for output voltage has been mitigated more than 31ms times through the proposed controller compared to conventional controller under the variation of CVL, CPL, and input voltage disturbance demonstrated in simulation results.

A predictive fuzzy logic and rule-based control approach for practical real-time operation of urban stormwater storage system

Predictive real-time control (RTC) strategies are usually more effective than reactive strategies for the intelligent management of urban stormwater storage systems. However, it remains a challenge to ensure the practicality of RTC strategies that use accessible, non-idealized predictive information while improving their efficiency for successive rainfall events instead of specific phases. This study developed a predictive fuzzy logic and rule-based control (PFL-RBC) approach to address the continuous control of individual storage systems. This approach incorporates total rainfall depth forecast information with an intra-storm fuzzy logic system to optimize peak flow control and several rule-based strategies for pre-storm water detention, reuse, and release control. Computational experiments were conducted using a storage tank case study to test the proposed approach under various rainfall conditions and storage sizes. The results showed that PFL-RBC outperformed static rule-based control in infrequent design storms and realistic continuous rainfall events, reducing flood peaks and volumes by 55 %∼87 % and 7 %∼20 %, respectively, and significantly increasing water detention time and reuse volume. Meanwhile, PFL-RBC required less predictive information to achieve a 6 %∼15 % advantage in peak flow control compared to optimized model predictive control. More importantly, PFL-RBC was reliable in the face of input uncertainty, with <25 % performance loss for water quantity control when the realistic forecast error ranged from -50 % to +50 %. These findings suggest that the proposed approach has great potential to enhance the efficiency and practicality of stormwater storage operations.

Robust optimization of geometrical properties of flow diverter stents for treating cerebral aneurysm: A proof-of-concept study

This study presents a novel approach to optimize the design of flow diverter (FD) stents for cerebral aneurysm (CA) treatment. By addressing sources of uncertainty in cardiovascular simulations, including geometrical and physical properties and boundary conditions, we aim to assess the applicability of robust optimization techniques to the FD design, establishing a foundation for acquiring robust FDs that are capable of operating consistently under various real-world scenarios. Blood flow in a simplified 2-dimensional CA and FD model was simulated through computational fluid dynamics (CFD). A design space exploration method, incorporating Latin hypercube sampling and Kriging surrogate models, was employed to obtain the optimal solution. The objective was to maximize the reduction in velocity and vorticity within the CA sac. This study used non-intrusive polynomial chaos expansion (PCE) to quantify and propagate the input uncertainties through the computational model and compute the statistical moments of velocity and vorticity reductions. To assess the effect of uncertain sources on objective functions, a sensitivity analysis method based on Sobol indices was applied. Robust optimization involved simultaneously optimizing the mean and standard deviation of velocity reduction. Additionally, we accounted for patients’ specific conditions and repeated the robust optimization. The results indicate that blood Hematocrit and inlet velocity are the most impactful uncertain sources in FD optimization. Moreover, the obtained Pareto front shows that in robust designs, FD struts are concentrated in the distal region of the CA neck, while optimal designs have more struts in the proximal region. This study proposes an FD that compromises robustness and optimality with a velocity reduction of 72.31 % and a standard deviation of 0.00343.

Simulating streamflow in a transboundary river catchment: The implications of hybrid rainfall data

The lack of observed data for the hydrological modelling of catchments across borders has hindered the management of transboundary water resources. This study investigated the implications of different degrees of hybridization of observed rainfall data using ERA5-Land reanalysis precipitation data in simulating streamflow in the Ruvubu River catchment across the Burundi-Tanzania border. The hydrological model of the Ruvubu River catchment was set up and iteratively updated using 0%–100% hybrid rainfall data, parameterized, simulated, and then evaluated against observed streamflow data at the catchment outlet. The findings show that the performance of the hydrological model decreased as the degree of rainfall data hybridization increased. However, model parameters satisfactorily compensated for input uncertainty when the model had hybrid rainfall data not exceeding 15% hybridization. Subsequently, errors in the simulated flow were also minimal. This implies that the simulated flow from the model when it has hybrid rainfall data not exceeding 15% hybridization can represent the simulated flow when it has 0% hybrid rainfall data. These findings show the need to develop thresholds of rainfall data hybridization for different global precipitation datasets in data-scarce transboundary river catchments.

A Framework for Multi-Physics Modeling, Design Optimization and Uncertainty Quantification of Fast-Spectrum Liquid-Fueled Molten-Salt Reactors

The analysis of liquid-fueled molten-salt reactors (LFMSRs) during steady state, operational transients and accident scenarios requires addressing unique reactor multi-physics challenges with coupling between thermal hydraulics, neutronics, inventory control and species distribution phenomena. This work utilizes the General Nuclear Field Operation and Manipulation (GeN-Foam) code to perform coupled thermal-hydraulics and neutronics calculations of an LFMSR design. A framework is proposed as part of this study to perform modeling, design optimization, and uncertainty quantification. The framework aims to establish a protocol for the studies and analyses of LFMSR which can later be expanded to other advanced reactor concepts too. The Design Analysis Kit for Optimization and Terascale Applications (DAKOTA) statistical analysis tool was successfully coupled with GeN-Foam to perform uncertainty quantification studies. The uncertainties were propagated through the input design parameters, and the output uncertainties were characterized using statistical analysis and Spearman rank correlation coefficients. Three analyses are performed (namely, scalar, functional, and three-dimensional analyses) to understand the impact of input uncertainty propagation on temperature and velocity predictions. Preliminary three-dimensional reactor analysis showed that the thermal expansion coefficient, heat transfer coefficient, and specific heat of the fuel salt are the crucial input parameters that influence the temperature and velocity predictions inside the LFMSR system.

A convex modelling based reliability analysis of rock structures with limited data of inputs modelled via alternative uncertainty models: Application for rock slopes

Reliability methods are used to deal with the variability in rock properties. These methods require probabilistic characteristics of properties, to be estimated via large quantities of objective data. This data is difficult to obtain due to different difficulties in rock testing, and/or subjective nature of certain properties. This invokes epistemic uncertainties in properties as indicated by their information levels, impeding the application of probabilistic reliability methods. This study proposes a non-probabilistic reliability methodology for problems with insufficient information of inputs. The methodology is further generalized to mixed uncertainty problems. A super-ellipsoid convex model with minimized volume is used to model objective inputs with limited data. The subjective inputs are modelled as fuzzy/interval models. The super-ellipsoid model is mapped to a unit multi-dimensional sphere, with non-probabilistic reliability index (βNP) defined as the shortest distance from its origin to the limit surface. A double-loop optimization algorithm is developed to estimate fuzzy/intervals of βNP. The method is demonstrated for a rock slope, with inputs modelled via mixed uncertainty models. The method efficiently estimates the converged fuzzy/intervals of βNP by propagating the impreciseness of inputs. The effect of variations in uncertainty and convex models of inputs on βNP is further investigated.

Parsimonious streamflow forecasting system based on a dynamical systems approach

Operational hydrologic forecasting often relies on the use of physically-based computer models at the catchment scale. However, our understanding of streamflow generation is still evolving and incomplete, and oftentimes linearization assumptions must be made to make physically-based hydrologic and hydraulic models tractable. In addition to uncertainty in the physical runoff-generation mechanisms, many physically-based approaches do not formally assimilate observed streamflow data. Streamflow is their output, deterministically simulated from observed inputs (such as precipitation, temperature, etc.). Divergences between observed and simulated streamflow may be due to input uncertainty, calibration issues, or problems with the inherent physical process representation. Thus, data-driven approaches that are responsive to observed streamflow by including assimilation as an essential feature are promising in operational hydrology. In this study, we use a novel forcings-weighted k-Nearest Neighbor (kNN) nonlinear timeseries forecasting method developed in the context of the modeling of nonlinear dynamical systems. The rainfall-runoff transform is assumed to be a low-dimensional system of (unknown) nonlinear differential equations; the deterministic state space is reconstructed based on the streamflow timeseries; and a forecast is generated based on how “similar” the current hydrograph is to historical events encoded within the reconstructed state space, weighted by the similarity of forcings. We tested the model in a long-term test period with no real data assimilation, and in forecast mode with daily data assimilation. When used for forecasting, we also illustrate two different uncertainty quantification approaches: forecast postprocessing and internal-to-the-model ensemble generation. The method is conceptually and computationally simple, but offers promising results. We tested the model in three different scenarios. First, investigating a long-term test period with no real data assimilation in a rainfall-dominated research catchment at the Coweeta Hydrologic Lab in North Carolina, USA. The simulation exhibited an NSE of 0.73. Next, we included a conceptual snowmelt model to drive the long term simulation and studied boreal-temperate transition zone catchments in northern Minnesota, USA. The study area included a small (<0.1 mi2) reference research catchment at the Marcell Experimental Forest, and a large river basin (>19 k mi2) with incomplete forcings information. The kNN method yielded NSEs from 0.28 to 0.50. And finally, we conducted a pseudo-operational forecasting test for a large river basin in northern Minnesota, USA with a focus on performance forecasting a flood-of-record on the Rainy River. The test involved reforecasting, data assimilation, and implementation of ensemble and postprocessing bias correction uncertainty-quantification schemes. The technique demonstrated good performance capturing a flood-of-record on a large catchment (−5.7 to −3.3 % crest error with 1-day of lead time). For the entire test period in forecast mode, the 1-day lead time NSE was 0.99, going down to 0.79 for the 7-day lead time, showing its general utility as an operational forecasting tool.

Adopting a cell-based method to identify mineralized geological structures in support of mineral prospectivity mapping at the Kuh–Lakht epithermal gold deposit, Central Iran

Mineral prospectivity mapping (MPM) is a new tool in mineral exploration with the objective of gathering and integrating multi-source evidence to provide decision-making criteria for exploration targeting. Ideally, the MPM approach accounts for the uncertainties associated with different predictive modeling techniques. The cell-based association (CBA) is a computational workflow that addresses the prediction-related bias of the MPM protocols. In this contribution, the CBA technique was employed to prioritize grid cells according to the importance of mineralized geological features and to identify geological settings favorable for ore deposition. The Kuh-Lakht epithermal gold deposit in Central Iran was adopted as a case study to examine the relevance of the proposed workflow and define target areas for brownfield exploration. To assess the significance of the proposed technique in dealing with evidence layers rather than training locations (i.e., mineralized/non-mineralized sites), two geochemical evidence layers, single-element and multi-element signatures, were used to train the CBA technique. An appraisal analysis through prediction-area plots and success-rate curves indicated that multi-element geochemical signatures are more relevant for delineating mineralized geological structures and vectoring for concealed gold mineralization. Notably, the input of single-element geochemical data into the CBA workflow enabled the prediction of additional exploration targets, whereas multi-element geochemical data failed to achieve the same results. The results indicated that the CBA technique can efficiently cope with the uncertainty of insufficient data input and permit the delineation of exploration targets. Additionally, our results suggest that the incorporation of a ranking system of critical geological factors that facilitate the gold mineralization into conventional MPM protocols can improve the robustness of mineral exploration targeting.

The Effect of Model Input Uncertainty on the Simulation of Typical Pollutant Transport in the Coastal Waters of China

Route planning to evade potential pollution holds critical importance for aquaculture vessels. This study establishes a fish-feed pollutant drift model based on the Lagrangian particle tracking algorithm and designs four sets of sensitivity experiments in the East China Sea. The research investigates the impact of model input uncertainties on the drift trajectory, centroid position, and sweeping area of the fish-feed pollutants. Numerical results indicate that the uncertainty in the background flow field significantly affects the uncertainty in the centroid position and sweeping area in the numerical simulations. Specifically, when a 35% random error is added to the background flow field, the centroid shift distance reaches its maximum, and the sweeping area also attains its largest value. The uncertainty in the background wind field affects the centroid position of particles but to a much lesser extent compared to the background flow field. When considering only the uncertainty of the background wind field, the sweeping area does not significantly differ from the control experiment as the uncertainty of the background wind field increases. The initial release position has little effect on the drift direction of the fish-feed pollutants but does affect the drift distance; it has minimal impact on the trajectory but significantly affects the final position of the pollutant centroid. By analyzing the model uncertainties, this study reveals the key factors influencing the drift of fish-feed pollutants. This information is crucial for aquaculture vessels in planning routes, considering environmental factors, and reducing potential pollution risks.

Mass property estimation on TSE(3) via unscented Kalman filter using RCS thrusters

Estimation of spacecraft mass properties in the presence of noise is a nontrivial problem that, when properly addressed, can substantially reduce the crew time devoted to cargo management, control effort expended to manage fuel mass, or both. The estimation of such properties is treated here using a dual unscented Kalman filter (UKF) where the dynamics of the rigid-body spacecraft are formulated on the special Euclidean group SE(3) and its tangent bundle TSE(3) to avoid singularity and nonuniqueness issues found in attitude parameterization sets. The dual UKF on TSE(3) is developed, and a specific methodology is proposed for the estimation of the constant mass properties. An excitation approach is employed in the mass property estimation scheme. Arbitrary sensor suite placements are considered and implemented in the measurement model to capture real-world spacecraft behavior. Furthermore, a reaction control system with duty cycle constraints and noisy thrust values is implemented to account for uncertainty in the excitation inputs. With these features, the methods contained herein are found to be effective at estimating mass properties.

L1 Adaptive Fault-Tolerant Control for Nonlinear Systems Subject to Input Constraint and Multiple Faults

This paper investigates an L1 adaptive fault-tolerant control scheme for nonlinear systems with input constraint, external disturbances, and multiple faults, which include actuator faults and sensor faults. Faults and input constraint are important factors that affect the stability and performance of a control system. Actuators and sensors are the most vulnerable components, with the former receiving more attention in comparison. In this paper, sensor faults are first transformed into pseudo-actuator faults through the augmented matrix approach, which facilitates their handling together with actuator faults. Saturation constraints on the control signal are not conducive to the design of the controller. The conversion of an input-saturated function to a time-varying linear system is completed based on function approximation and Lagrange’s mean value theorem. Moreover, a nonlinear system with unknown input gain and uncertainties is constructed using these methods. Next, an L1 adaptive fault-tolerant controller is designed to cope with uncertainties, including system uncertainties, external disturbances, faults, and approximation errors. In the L1 adaptive controller, the online estimation of the time-varying parameters allows for updating of the system state, while the combination of the two is transmitted to the control law such that it can compensate for the effects of the uncertainties. The stability and performance boundaries are further derived using the Lyapunov theory and the L1 reference system. Finally, simulations are carried out to demonstrate the effectiveness of the proposed controller.

Robust optimization of a novel ultraviolet (UV) photoreactor for water disinfection: A neural network approach

To optimize the ultraviolet (UV) water disinfection process, it is crucial to determine the ideal geometric dimensions of a corresponding model that enhance performance while minimizing the impact of uncertain photoreactor inputs. As water treatment directly affects people's lives, it is crucial to eliminate the risks associated with the non-ideal performance of disinfection photoreactors. Input uncertainties greatly affect photoreactor performance, making it essential to develop a robust optimization algorithm in advance to mitigate these effects and minimize the physical and financial resources required for constructing the photoreactors. In the suggested algorithm, a two-objective genetic algorithm is integrated with a non-intrusive polynomial chaos expansion (PCE) technique. Additionally, the Sobol sampling method is employed to select the necessary samples for understanding the system's behavior. An artificial neural network surrogate model is trained using sufficient data points derived from computational fluid dynamics (CFD) simulations. A novel type of UV photoreactors working based on exterior reflectors is chosen to optimize the process with three uncertain input parameters, including UV lamp power, UV transmittance of water, and diffusive fraction of the reflective surface. In addition, four geometrical design variables are considered to find the optimal configuration of the photoreactor. The standard deviation (SD) and the reciprocal of log reduction value (LRV) are set as the objective functions, calculated using PCE. The optimal design provides a LRV of 3.95 with SD of 0.2. The coefficient of variation (CoV) of the model significantly declines up to 7%, indicating the decreased sensitivity of the photoreactor to the input uncertainties. Additionally, it is discovered that the robust model exhibits minimal sensitivity to changes in reflectivity in various flow rates, and its output variability aligns with the SD obtained through robust optimization.

Improving Flood Forecasting Skill by Combining Ensemble Precipitation Forecasts and Multiple Hydrological Models in a Mountainous Basin

Ensemble precipitation forecasts (EPFs) derived from single numerical weather predictions (NWPs) often miss extreme events, and individual hydrological models (HMs) often fail to accurately capture all types of flows, including flood peaks. To address these shortcomings, this study introduced four “EPF + HM” schemes for ensemble flood forecasting (EFF) by combining two EPFs and two HMs. A generator-based post-processing (GPP) method was applied to correct biases and under-dispersion within the raw EPF data. The effectiveness of these schemes in delivering high-quality flood forecasts was assessed using both deterministic and probabilistic metrics. The results indicate that, once post-processed by GPP, all proposed schemes show improvements in both deterministic and probabilistic performances, with skillful flood forecasts for 1–7 lead days. The deterioration in forecast performance with extended lead times is also lessened. Notably, the results indicate that uncertainty within hydrological models has a more pronounced impact on capturing flood peaks than uncertainty in precipitation inputs. This study recommends combining individual EPF with multiple hydrological models for reliable flood forecasting. In conclusion, effective flood forecasting necessitates employing post-processing techniques to correct EPFs and accounting for the uncertainty inherent in hydrological models, rather than relying solely on the uncertainty of the input data.

Novel gradient-enhanced Bayesian neural networks for uncertainty propagation

Uncertainty propagation (UP) is crucial for assessing the impact of input uncertainty on structural responses, holding significant importance in engineering applications. However, achieving accurate and efficient UP remains challenging, especially for highly nonlinear structures. Bayesian neural networks (BNN) have gained attention for addressing UP issues, yet current BNN models only utilize input samples and corresponding structural responses for training. However, incorporating gradients of structural responses with respect to input samples provides valuable information. This study proposes a novel approach called gradient-enhanced Bayesian neural networks (GEBNN) to tackle UP tasks. In the GEBNN, a modified evidence lower bound (MELBO) loss is developed to consider both structural responses and gradient information simultaneously. This includes disparities between actual and predicted responses, as well as disparities between actual and predicted derivatives. Additionally, a gradient screening strategy based on the marginal probability density functions (PDFs) of input samples is established to identify significant derivative data for GEBNN training. Once the GEBNN is configured, it is utilized to replace the computationally intensive finite element model to efficiently provide UP results. Various applications, including nonlinear numerical examples, and mechanical, civil, and aeronautical structures, are presented to demonstrate the effectiveness of the GEBNN. The results show that the GEBNN significantly enhances the computational accuracy of the BNN in solving UP tasks.

Sensitivity analysis of the Cercignani - Lampis accommodation coefficients in prototype rarefied gas flow and heat transfer problems via the Monte Carlo method

In rarefied gas dynamics, the Cercignani-Lampis (CL) scattering kernel, containing two accommodation coefficients (ACs), namely the tangential momentum and normal energy ones, is widely employed to characterize gas-surface interaction, particularly in non-isothermal setups, where both momentum and energy may simultaneously be exchanged. Here, a formal and detailed sensitivity analysis of the effect of the CL ACs on the main output quantities of several prototype problems, namely the cylindrical Poiseuille, thermal creep and thermomolecular pressure difference (TPD) flows, as well as the plane Couette flow and heat transfer (Fourier flow), is performed. In each problem, some uncertainties are randomly introduced in the ACs (input parameters) and via a Monte Carlo propagation analysis, the deduced uncertainty of the corresponding main output quantity is computed. The output uncertainties are compared to each other to determine the flow configuration and the gas rarefaction range, where a high sensitivity of the output quantities with respect to the CL ACs is observed. The flow setups and rarefaction regimes with high sensitivities are the most suitable ones for the estimations of the ACs, since larger modeling and experimental errors may be acceptable. In the Poiseuille and Couette flows, the uncertainties of the flow rate and shear stress respectively are several times larger than the input uncertainty in the tangential momentum AC and much smaller than the uncertainty in the normal energy AC in a wide range of gas rarefaction. In the thermal creep flow, the uncertainty of the flow rate depends on the input ones of both ACs, but, in general, it remains smaller than the input uncertainties. A similar behavior with the thermal creep flow is obtained in the TPD flow. On the contrary, in the Fourier flow, the uncertainty of the heat flux may be about the same or even larger than the input ones of both ACs in a wide range of gas rarefaction. It is deduced that in order to characterize the gas-surface interaction via the CL ACs by matching computations with measurements, it is more suitable to combine the Poiseuille (or Couette) and Fourier configurations, rather than, as it is commonly done, the Poiseuille and thermal creep ones. For example, in order to estimate the normal energy AC within an accuracy of 10 %, experimental uncertainties should be less than 4 % in the thermal creep or TPD flows, while may be about 10 % in the Fourier flow.