Impact Of Model Structure Research Articles

Impact of model structure on radionuclide transport in urban stormwater

This paper compared simulation results from different SWMM model constructions after radiological contamination from a hypothetical dirty bomb. The model constructions included a utility-provided combined sewer model, a high-resolution mesh surface flow model, and a hybrid of the two. The models were run under three different rainfall configurations including a 3-h event, a 24-h event, and a 24-h event with 48 h of drainage to investigate the influence that model structure had on transport of 137Cs. Model structure was found to impact the contaminant loading of the stormwater model from the air modeling. It was also found to impact the amount of contaminant washoff (high resolution modeling produced more washoff), flow of contamination through outfalls (more contamination through outfalls connected to underground pipes), and 137Cs levels per pipe. The implications of these findings are that few stormwater models are considered ready to aid in response and recovery as originally designed.

Effect of load modeling on power system stability studies

Load models have been widely studied, mostly focusing on the model structure or on the parameter derivation. However, there are few studies investigating the effect of load modeling on the dynamic behavior of the overall power system. In this paper, the effect of load modeling on transient, frequency, and small-signal stability analysis is presented. For this purpose, a comprehensive review on relevant works is conducted. Simulations on a benchmark power system are performed to investigate the impact of load modeling on power system stability. Towards this objective, the exponential and the polynomial models are considered, assuming both voltage and frequency dependent implementations. Different load mixtures are examined by varying the load model parameters. The impact of model structure and model parameters on the overall system stability is systematically quantified by means of key performance indicators.

Probability prediction of pavement surface low temperature in winter based on bayesian structural time series and neural network

Reliable pavement surface temperature prediction models are important for decision making in winter road maintenance. Due to the complexity of thermal environment in winter, there is inevitably uncertainty in pavement temperature. Therefore, the prediction results of pavement surface temperature should not be a single value, but a probability distribution. This paper developed a prediction model for evaluating the probability distribution of pavement surface temperature in winter. The model consisted of two modules, namely, a Bayesian Structural Time Series module (BSTS) and a Bayesian Neural Network module (BNN). The BSTS module provided a relatively rough prediction of pavement surface temperature, while the BNN module provided refined prediction based on the results from the BSTS module. The model was validated with data collected from the field test. The impact of model structure, model parameters, and probability threshold on the predictions was analyzed. Results show that most measured pavement temperatures fell into the 95% confidence interval of the predicted values, indicating the prediction was reliable. For the BSTS module, the regression component plays the most important role in the prediction. The seasonal component also has obvious influence on the prediction results. While the influence of local linear trend component on the prediction result is not obvious. For the BNN module, although the slippery index and pavement surface condition do not directly contribute to the heat transfer process, these two factors have correlation with the pavement surface temperature. The value of probability threshold should be properly determined to balance the reliability and computational cost of the prediction model. The model presented in this paper is useful for transportation agencies in winter road maintenance

Land surface modeling over the Dry Chaco: the impact of model structures, and soil, vegetation and land cover parameters

Abstract. In this study, we tested the impact of a revised set of soil, vegetation and land cover parameters on the performance of three different state-of-the-art land surface models (LSMs) within the NASA Land Information System (LIS). The impact of this revision was tested over the South American Dry Chaco, an ecoregion characterized by deforestation and forest degradation since the 1980s. Most large-scale LSMs may lack the ability to correctly represent the ongoing deforestation processes in this region, because most LSMs use climatological vegetation indices and static land cover information. The default LIS parameters were revised with (i) improved soil parameters, (ii) satellite-based interannually varying vegetation indices (leaf area index and green vegetation fraction) instead of climatological vegetation indices, and (iii) yearly land cover information instead of static land cover. A relative comparison in terms of water budget components and “efficiency space” for various baseline and revised experiments showed that large regional and long-term differences in the simulated water budget partitioning relate to different LSM structures, whereas smaller local differences resulted from updated soil, vegetation and land cover parameters. Furthermore, the different LSM structures redistributed water differently in response to these parameter updates. A time-series comparison of the simulations to independent satellite-based estimates of evapotranspiration and brightness temperature (Tb) showed that no LSM setup significantly outperformed another for the entire region and that not all LSM simulations improved with updated parameter values. However, the revised soil parameters generally reduced the bias between simulated surface soil moisture and pixel-scale in situ observations and the bias between simulated Tb and regional Soil Moisture Ocean Salinity (SMOS) observations. Our results suggest that the different hydrological responses of various LSMs to vegetation changes may need further attention to gain benefits from vegetation data assimilation.

Ore image classification based on small deep learning model: Evaluation and optimization of model depth, model structure and data size

The ore image classification technology based on deep learning is an effective way to improve the image sensor-based ore sorting classification capability. However, in practice, the image sensor-based ore sorting technique often has the problem of insufficient data, and has not systematically considered the impact of model structure and dataset size on the modeling efficiency and classification performance of deep learning. Therefore, this paper attempts to explore a more suitable small deep learning model for ore image classification by considering the model depth, model structure, and dataset size. Six Convolutional Neural Networks (CNNs) models are established with different depths based on Alex Net and VGG Net and the model structure is optimized by adding BN layer. Taking the gas-coal image dataset as case study, we systematically explore the influence of model depth, model structure, dataset size on the training process efficiency and classification accuracy. Meanwhile, the operational process of coal image classifiers is analyzed visually through the ways of Channel Visualization maps, Heatmaps, Grad-CAM map, and Guided Backpropagation maps.

Impact of model structure on the accuracy of hydrological modeling of a Canadian Prairie watershed

Study regionPrairie Pothole Region of Canada. Study focusThe Prairie region spans across approximately 870,000 km2 of the Great Plains region of Canada (80%) and the United States (20%). The presence of a large number of depressional wetlands (potholes) results in dynamic surface-water and stream connectivity during wet and dry year necessitating an improved understanding of watershed-scale interactions of the Prairie Potholes. The Soil Water Assessment Tool (SWAT) hydrological model with three structural variants is utilized to assess the degree of accuracy associated with increasing model complexity and its impact on the model calibration of the Upper Assiniboine River Basin at Kamsack. New hydrologic insight for the regionThe SWAT model was calibrated and verified with three different structural arrangements in 1) lumped pothole, 2) semi discretized pothole, 3) and fully discretized pothole representation. The fully discretized pothole version of the SWAT reflected streamflow best (KGE of 0.78) but with greater uncertainty, larger data and computational resource requirements. The fully discretized (modified) model, however, was able to capture the high flow and the fill-and-spill processes, which is a defining characteristic of the Prairie Pothole Region (PPR). Significant improvements to the predictive ability of SWAT in the case of the modified model was observed, thus allowing an enhanced understanding of the aggregate effect of potholes in this watershed.

Cost-utility of quadrivalent versus trivalent influenza vaccine in Brazil – comparison of outcomes from different static model types

BackgroundInfluenza burden in Brazil is considerable with 4.2–6.4 million cases in 2008 and influenza-like-illness responsible for 16.9% of hospitalizations. Cost-effectiveness of influenza vaccination may be assessed by different types of models, with limitations due to data availability, assumptions, and modelling approach. ObjectiveTo understand the impact of model complexity, the cost-utility of quadrivalent versus trivalent influenza vaccines in Brazil was estimated using three distinct models: a 1-year decision tree population model with three age groups (FLOU); a more detailed 1-year population model with five age groups (FLORA); and a more complex lifetime multi-cohort Markov model with nine age groups (FLORENCE). MethodsAnalysis 1 (impact of model structure) compared each model using the same data inputs (i.e., best available data for FLOU). Analysis 2 (impact of increasing granularity) compared each model populated with the best available data for that model. ResultsUsing the best data for each model, the discounted cost-utility ratio of quadrivalent versus trivalent influenza vaccine was R$20,428 with FLOU, R$22,768 with FLORA (versus R$20,428 in Analysis 1), and, R$19,257 with FLORENCE (versus R$22,490 in Analysis 1) using a lifetime horizon. Conceptual differences between FLORA and FLORENCE meant the same assumption regarding increased all-cause mortality in at-risk individuals had an opposite effect on the incremental cost-effectiveness ratio in Analysis 2 versus 1, and a proportionally higher number of vaccinated elderly in FLORENCE reduced this ratio in Analysis 2. DiscussionFLOU provided adequate cost-effectiveness estimates with data in broad age groups. FLORA increased insights (e.g., in healthy versus at-risk, paediatric, respiratory/non-respiratory complications). FLORENCE provided greater insights and precision (e.g., in elderly, costs and complications, lifetime cost-effectiveness). ConclusionAll three models predicted a cost per quality-adjusted life year gained for quadrivalent versus trivalent influenza vaccine in the range of R$19,257 (FLORENCE) to R$22,768 (FLORA) with the best available data in Brazil (Appendix A).

Uncertainty and variability in computational and mathematical models of cardiac physiology.

Key points Mathematical and computational models of cardiac physiology have been an integral component of cardiac electrophysiology since its inception, and are collectively known as the Cardiac Physiome.We identify and classify the numerous sources of variability and uncertainty in model formulation, parameters and other inputs that arise from both natural variation in experimental data and lack of knowledge.The impact of uncertainty on the outputs of Cardiac Physiome models is not well understood, and this limits their utility as clinical tools.We argue that incorporating variability and uncertainty should be a high priority for the future of the Cardiac Physiome.We suggest investigating the adoption of approaches developed in other areas of science and engineering while recognising unique challenges for the Cardiac Physiome; it is likely that novel methods will be necessary that require engagement with the mathematics and statistics community. The Cardiac Physiome effort is one of the most mature and successful applications of mathematical and computational modelling for describing and advancing the understanding of physiology. After five decades of development, physiological cardiac models are poised to realise the promise of translational research via clinical applications such as drug development and patient‐specific approaches as well as ablation, cardiac resynchronisation and contractility modulation therapies. For models to be included as a vital component of the decision process in safety‐critical applications, rigorous assessment of model credibility will be required. This White Paper describes one aspect of this process by identifying and classifying sources of variability and uncertainty in models as well as their implications for the application and development of cardiac models. We stress the need to understand and quantify the sources of variability and uncertainty in model inputs, and the impact of model structure and complexity and their consequences for predictive model outputs. We propose that the future of the Cardiac Physiome should include a probabilistic approach to quantify the relationship of variability and uncertainty of model inputs and outputs.

Impact of model structure on the performance of dynamic data reconciliation

Dynamic data reconciliation (DDR) is a technique used to estimate the true values of process variables when plant measurements are corrupted by measurement noise. DDR integrates information from both measurements and process models such that the reconciled values become more reliable and better represent the current state of the process. Process models play a key role in the performance of DDR. Empirical or black-box models are identified and used in the DDR when phenomenological models are unavailable, impractical to obtain, or whose solutions require excessive computation time for real-time applications. Black-box models usually have higher degree of uncertainty and use a wide variety of structures. This article examines the impact of model structure on the performance of DDR, and more importantly, on the performance of controllers when the DDR is embedded inside feedback control loops. Simulation results of a binary distillation column demonstrated that the model structure can have a major impact on the performance of DDR. The DDR using simple linear models can successfully attenuate the noise propagation; however, further significant improvement of DDR performance can be achieved if more advanced models, such as nonlinear models, are used.

Estimating Exchange Rate Exposures: Issues in Model Structure

We show that both return measurement horizon and model specification have noticeable impacts on estimates of exposure from equity prices for U.S. firms. While increases in the return horizon leads to increases in the precision of the estimates, this effect is less significant than the impact of model structure. We demonstrate that the inclusion and of a market return variable and its particular construction has a dramatic influence on the sign and size of the exposures due to a strong relation between firm size and exposure for U.S. firms. We propose using CRSP cap-based portfolios as the control for market factors and show that this produces exposures with stronger relation to foreign cash flows and correlations with firm size.