Model Building Research Articles

Techno-Economic Analysis of a High-Rise Residential Building Adapted to Nearly Zero-Energy Building Standards

Zero-energy buildings have attracted great attention in China. Limited research about typical high-rise, zero-energy residential buildings in China was found. To figure out the potential of zero-energy buildings in northern China, a techno-economic analysis of a typical residential building adapted to the nearly zero energy building (NZEB) standards in the cold region of China was carried out in detail in this paper. Firstly, the feasibility of different building energy efficiency technologies was figured out in the passive design level. Secondly, the annual energy balance of the nearly zero-energy building model was investigated. Finally, detailed economic and environmental analyses were performed. The results show that the energy consumption of space heating and cooling of a typical high-rise, nearly zero-energy building could decrease to 11.1 kWh/(m2·a) in Beijing. The conclusions could provide a reference and design basis for the development of zero-energy residential buildings in northern China in the near future.

Multimodal Data Integration for Early Alzheimer’s Detection Using Random Forest and Support Vector Machines

Alzheimer's is a very challenging brain disease to recognize, diagnose, and treat correctly when it appears in its earliest forms. The primary contribution of this research study is about machine learning models, techniques, and approaches. In contrast, Random Forest and Support Vector Machine (SVM) are particularly suitable for identifying and staging Alzheimer's disease stages using multimodal data sources. In this paper, the aim was to develop well-performing predictive models to help diagnose Alzheimer's disease at an early stage by combining neuroimaging data (MRI/PET images), imaging-based biomarkers, both structural and functional measures from MRI(P) /PET image analysis along with subject-specific demographics like age using clinical features in a probabilistic fashion obtained from the Alzheimer's Disease Neuro-Imaging Initiative (ADNI) database. The methodology focuses on data pre-processing, feature selection, and model building using supervised learning algorithms. The accuracy of the Random Forest model is 78%, having a high performance in classifying some classes while showing different marks of performances across other courses. SVM reached an accuracy of 61%, or the model's performance is good in some classes and not reliable to identify samples from the others. The findings of this study underscore the capabilities and limits of these machine learning models in identifying Alzheimer’s disease and highlight the importance of feature engineering, data pre-processing, and model tuning to increase performance and correct class unevenness and misclassification.

Modeling Ligands into Maps Derived from Electron Cryomicroscopy.

Deciphering the protein-ligand interactions in a macromolecular complex is crucial for understanding the molecular mechanism, underlying biological processes, and drug development. In recent years, cryogenic sample electron microscopy (cryoEM) has emerged as a powerful technique to determine the structures of macromolecules and to investigate the mode of ligand binding at near-atomic resolution. Identifying and modeling non-protein molecules in cryoEM maps is often challenging due to anisotropic resolution across the molecule of interest and inherent noise in the data. In this article, the readers are introduced to various software and methods currently used for ligand identification, model building, and refinement of atomic coordinates using selected macromolecules. One of the simplest ways to identify the presence of a ligand, as illustrated with the enolase enzyme, is to subtract the two maps obtained with and without the ligand. The extra density of the ligand is likely to stand out in the difference map even at a higher threshold. There are instances, as shown in the case of metabotropic Glutamate receptor mGlu5, when such simple difference maps cannot be generated. The recently introduced method of deriving the Fo-Fc omit map can serve as a tool for validating and demonstrating the presence of the ligand. Finally, using the well-studied β-galactosidase as an example, the effect of resolution on modeling the ligands and solvent molecules in cryoEM maps is analyzed, and an outlook on how cryoEM can be used in drug discovery is presented.

A hybrid data assimilation method for reconstructing airflow path parameters of a multi-zone model

The indoor airborne pollutant has a great impact on indoor air quality. During the calculation process of pollutant distribution, some difficult-to-obtain parameters are critical in ensuring the calculation accuracy. In this study, we proposed a hybrid data assimilation method combined with Ensemble Kalman Filter (EnKF) and Gibbs sampling to reconstruct airflow path parameters (flow coefficient and flow index) in a multi-zone building model. Specially, we built a scaled multi-zone building model and used CO2 as the tracer gas to simulate the pollutant distribution in the building. The monitored sensor concentration data were used as prior information to reconstruct airflow path parameters of the scaled building model. And the performances of different data assimilation methods were compared. The results show that the EnKF method can reduce the calculation time by 46 % compared to the Gibbs sampling method, but the predicted accuracy decreases by 38 %. However, when we used the hybrid algorithm to reconstruct the airflow path parameters, the calculation time was reduced by 67 % compared to the EnKF method, and the predicted reconstruction accuracy is 7 % higher than the Gibbs sampling method.

High Healthcare Costs in Childhood Inflammatory Bowel Disease: Development of a Prediction Model Using Linked Clinical and Health Administrative Data.

The incidence of pediatric-onset inflammatory bowel disease (IBD) and the costs of caring for individuals with IBD are both increasing. We calculated the direct healthcare costs of pediatric IBD in the first year after diagnosis and developed a model to predict children who would have high costs (top 25th percentile). Using data from the Canadian Children IBD Network inception cohort (≤16 years of age, diagnosed between 2013 and 2019) deterministically linked to health administrative data from Ontario, Canada, we estimated direct healthcare and medication costs accrued between 31 and 365 days after diagnosis. Candidate predictors included age at diagnosis, sex, rural/urban residence location, distance to pediatric center, neighborhood income quintile, IBD type, initial therapy, disease activity, diagnostic delay, health services utilization or surgery around diagnosis, regular primary care provider, and receipt of mental health care. Logistic regression with stepwise elimination was used for model building; 5-fold nested cross-validation optimized and improved model accuracy while limiting overfitting. The mean cost among 487 children with IBD was CA$15 168±15 305. Initial treatment (anti-tumor necrosis factor therapy, aminosalicylates, or systemic steroids), having a mental health care encounter, undergoing surgery, emergency department visit at diagnosis, sex, and age were predictors of increased costs, while having a regular primary care provider was a predictor of decreased costs. The C-statistic for our model was 0.71. The cost of caring for children with IBD in the first year after diagnosis is immense and can be predicted based on characteristics at diagnosis. Efforts that mitigate rising costs without compromising quality of care are needed.

The Complex Interrelationships of the Risk Factors Leading to Hamstring Injury and Implications for Injury Prevention: A Group Model Building Approach

There is a gap in the literature regarding the complex interrelationships among hamstring injury (HI) risk factors. System dynamics (SD) modeling is considered an appropriate approach for understanding the complex etiology of HI for effective injury prevention. This study adopted the SD method and developed a causal loop model (CLD) to elucidate the intricate relationships among HI risk factors. This is performed by combining literature evidence and insights from expert stakeholders through a group model building (GMB) approach. The GMB methodology facilitated the identification of sixty-five critical factors influencing the HI risk, revealing the dynamic interplay between factors. Stakeholder engagement underscored the importance of previous injury characteristics (level of influence of previous injury, severity of previous injury, quality and size of scar tissue) and the quality of rehabilitation. HI-CLD revealed that many factors had indirect effects on HI risk. The HI-causal loop model establishes a foundation for a future stock and flow quantitative SD model aiming to advance HI prevention strategies through an interdisciplinary collaborative effort. These findings underscore the complexity of HI prevention, necessitating a holistic approach that integrates the views of diverse professional expertise. Appropriate inter-professional collaboration and continuous athlete screening are important for effective injury prevention strategies.

Repeated Sieving for Prediction Model Building with High-Dimensional Data.

Background: The prediction of patients' outcomes is a key component in personalized medicine. Oftentimes, a prediction model is developed using a large number of candidate predictors, called high-dimensional data, including genomic data, lab tests, electronic health records, etc. Variable selection, also called dimension reduction, is a critical step in developing a prediction model using high-dimensional data. Methods: In this paper, we compare the variable selection and prediction performance of popular machine learning (ML) methods with our proposed method. LASSO is a popular ML method that selects variables by imposing an L1-norm penalty to the likelihood. By this approach, LASSO selects features based on the size of regression estimates, rather than their statistical significance. As a result, LASSO can miss significant features while it is known to over-select features. Elastic net (EN), another popular ML method, tends to select even more features than LASSO since it uses a combination of L1- and L2-norm penalties that is less strict than an L1-norm penalty. Insignificant features included in a fitted prediction model act like white noises, so that the fitted model will lose prediction accuracy. Furthermore, for the future use of a fitted prediction model, we have to collect the data of all the features included in the model, which will cost a lot and possibly lower the accuracy of the data if the number of features is too many. Therefore, we propose an ML method, called repeated sieving, extending the standard regression methods with stepwise variable selection. By selecting features based on their statistical significance, it resolves the over-selection issue with high-dimensional data. Results: Through extensive numerical studies and real data examples, our results show that the repeated sieving method selects far fewer features than LASSO and EN, but has higher prediction accuracy than the existing ML methods. Conclusions: We conclude that our repeated sieving method performs well in both variable selection and prediction, and it saves the cost of future investigation on the selected factors.

Definition of a PVT coupled water source heat pump system through optimization of individual components

Exploiting renewable resources considering their discontinuity compared to building needs, necessitates energy system designs that simultaneously maximize generation and storage performance in different seasons. This article presents the optimization of RESHeat European project, concerning a high-performance system which combines water-source heat pump, cooled photovoltaic panels and thermal storage. The research regards the Italian demo site, near Rome, to define a solution for Mediterranean regions which alongside the requirement for heating have notable need for summer cooling. The targets are system's efficiency, with a minimum sCOP of 5, and a minimum 70 % coverage from renewable sources focusing ambient temperature management.A dynamic model of plant and building was used for parametric analysis, which covered separately heating and cooling modes, involving significant parameters: operating temperatures of the heat pump, number of strings comprising the photovoltaic array, volume of the buffer tank connecting PVTs and WSHP and setpoint temperature of a cooler added for tank temperature control in cooling. 184 alternative configurations were evaluated by Multi-Criteria Decision Making (MCDM) method based on energy, economic and logistic criteria, and weighted by five indices. The optimal system includes 15 PVT panels strings for 25 kWp, a 3 m³ storage tank and cooling setpoint temperature of 25 °C.

DEVELOPMENT OF A SUSTAINABLE BUSINESS MODEL FOR DENIM

Globally, over a billion dollars are spent on denim jeans each year, and that number is still growing. As a result of this expansion, the production of waste in the systems is also increasing. According to the Environmental Protection Agency (EPA), roughly 5% of all landfill space is textile waste, including denim. One of the least ecological industries in the world is the denim industry. This calls for more intensive interventions in the design of sustainable development, as it calls into question the key criteria of sustainability. This topic has been covered by various models of industrial sustainability, although these models focus on solving global problems, assuming small improvements and growth. Consequently, a sustainable business model approach is necessary to drastically rethink how firms should operate to support system-level sustainability. This paper will provide comprehensive background information on the denim business and its operations before offering some notable case studies of well-known companies that use environmentally friendly practices. The result of this document will be a comprehensive analysis and building of a sustainable business model for the denim sector.

A Study on the Impact of Window Partition Walls on the Spread of Fire on Building Facades

This paper investigates the impact of window partition walls on the spread of fire on building facades under the impact of environmental wind through Fire Dynamics Simulator simulation experiments. A four-story building model was constructed using a Fire Dynamics Simulator incorporating six different wind speed conditions and six different partition wall widths. The fire-blocking performance of window partition walls of varying widths was systematically compared and analyzed, and the data indicated: (1) Under calm wind conditions, the installation of window partition walls is observed to facilitate the vertical spread of facade fires. Moreover, as the width of these partition walls increases, this facilitative effect becomes increasingly prominent; (2) Under wind speeds of 0 to 5 m/s, the temperature on the leeward side is lower when window partition walls are present than when they are absent. This indicates that window partition walls inhibit the horizontal spread of building facade fires, and wider window partition walls have better horizontal fire resistance performance.

Quality-preserving multilevel mesh generation for building models

ABSTRACT Three-dimensional (3D) high-fidelity building models are crucial for digital twin cities. Multi-scale modeling is crucial for accurately capturing the complexity of urban environments; however, existing methods often lack the flexibility required to satisfy the varied demands of diverse applications. In this paper, we propose a framework to continuously simplify 3D building models, it dynamically adjusts mesh accuracy to meet diverse needs, enabling seamless transitions from high-poly to medium and low-poly meshes. Our method contains three main stages: First, in the mesh repair stage, we introduce a novel mesh repair framework based on an advanced Marching Cubes 33 algorithm to ensure a watertight and clean topology mesh. Then, for generating medium-poly meshes, we employ a feature-preserving mesh simplification algorithm that combines curvature and area-weighted quadratic error metrics to achieve a high-fidelity simplified mesh. For generating low-poly meshes, we use the concept of carving visual hull to generate low-poly mesh with high visual similarity. Our method has been evaluated on building subsets of the Free3D and Thingi10 K datasets, as well as on various building models generated from different observation scales, and its superiority has been validated through extensive comparisons with state-of-the-art techniques.

Pollutant Dispersion Around a Single Tall Building

An experiment was carried out using a scale model of a tall building, with the goal of investigating the role of individual buildings in the dispersion of air pollution. Pollutant dispersion around an isolated building with a height-to-length aspect ratio of 1.4 is investigated using simultaneous particle image velocimetry and planar laser induced fluorescence. Dye is released from a ground-level point source five building heights upstream of the tall building. It was found that in this case the scalar plume was dispersed laterally strongly by the building, but only slightly vertically. It is hypothesized that this is due to 94% of the plume impinging below the stagnation point on the front of the building and being drawn into the horseshoe vortex. We expect this fraction would be lower in a case in which the building is in an array of smaller buildings, and that this would lead to more vertical dispersion.

A hybrid transfer learning to continual learning strategy for improving cross-building energy prediction in data increment scenario

For a data-driven building energy prediction model, the insufficient data available for new buildings and existing buildings with limited data pose a challenge to achieving the expected prediction performance. Transfer learning (TL) and continual learning (CL) are two widely applied effective methods for addressing the shortage of building data. The actual building projects typically exist in a scenario of data increment, wherein their energy consumption data gradually accumulates with usage. Previous studies have indicated that the TL strategy can transfer the learned data distribution patterns from source buildings to target buildings of the same type. Meanwhile, the CL strategy can extract the dynamic characteristics from continuously accumulated data. Both strategies can be employed to enhance the performance of energy consumption prediction models for buildings with insufficient data. However, existing studies on data scarcity in energy prediction primarily focus on buildings with fixed data quantities, and have not provided sufficient analysis for newly constructed buildings in a data increment scenario. Therefore, this study proposes a TL-to-CL strategy to enhance the energy prediction performance of newly constructed buildings in a data increment scenario. When the available data is limited, TL should be utilized. But as the data volume increases to be sufficient, TL should be transformed to CL at the reasonable time. Additionally, this study provides a reference for determining the transform time for the TL-to-CL strategy. The validation was conducted using two years of energy consumption data from 36 buildings in the Genome Project 2 dataset. The results demonstrate that the proposed TL-to-CL strategy (transform time = 4-week) obtain the overall better prediction performance than pure TL, pure CL and combined TL-CL without transformation between the two) strategies. The prediction improvement ratio (PIR) can be as high as 0.9 compared with the traditional LSTM.

Progressive Collapse and Post Impact Damage Assessment in a Regular Beam-Slab Building & Flat-Slab Building on the Intermediate Floor

Progressive collapse, where a localized member failure causes widespread structural collapse, has become a critical concern nowadays, due to its potential to cause significant financial losses and loss of human life. Triggers include natural disasters like earthquakes and floods, as well as accidents, attacks and explosions. Reinforced concrete flat slab structures, which are eminent for their architectural flexibility and have larger spans, are particularly susceptible to disproportionate collapse, due to the lack of floor beams, which can redistribute loads after a column failure, unlike moment frame buildings. This research examines how multi-story reinforced concrete flat slab buildings behave, under prescribed gravity load combinations, compared to conventional framed buildings. The effects of removing columns at specified locations from an intermediate floor of the multistorey building are also examined. However, this investigation covers both the column removal approaches to check a possibility of disproportionate collapse which are; static removal and dynamic instantaneous removal. Furthermore, the research also assesses the efficacy of perimeter beams, in minimising the risk of gradual collapse in flat slab structures, by scrutinising their ability to reduce joint displacement, chord rotation, and demand capacity ratio. ETABS v18 was used to analyze all the 18 models. The findings revealed that buildings are more prone to progressive collapse when corner columns are removed, as opposed to edge and interior columns, due to higher Demand-Capacity Ratios (DCR) and joint displacement. In comparison to dynamic analysis, the static evaluation exhibited greater DCR values and vertical joint displacement. Furthermore, since they have a more efficient load redistribution mechanism, traditional framed structures performed better than flat slab models. The simulations additionally indicated that, adding edge perimeter beams, substantially lowered the possibility of progressive collapse in flat slab structures. Moreover, the tested flat slab building models, with and without perimeter beams showed no indications of progressive collapse, when specified columns were removed from the intermediate floors, since the DCR values of the crucial columns stayed within the permissible range of 2.0. In conclusion, structures built in compliance with IS 1893:2016 code and designed to withstand seismic loads demonstrate stronger resistance against significant damage, brought about by column failures.

A fast modeling method of pulsed eddy current with double-coil based on model-driven BP neural network

ABSTRACT The analytical model serves as an important guiding role in practical testing. Currently, there is scarce research on the building of the analytical model for pulsed eddy current (PEC) with double-coil excited by a voltage source. This paper derives a frequency-domain model for PEC from a double-coil’s equivalent circuit. To tackle low modeling efficiency due to complex impedance changes and mutual inductance changes, a BP neural network (BPNN) maps lift-off, thickness, and frequency to impedance changes and mutual inductance changes. Training samples for the BPNN come from analytical models of impedance change and mutual inductance change. The study finds that a BPNN with four hidden layers and Bayesian regularization significantly improves prediction accuracy. Finally, the correctness of the BPNN method is verified. The PEC signals obtained by the BPNN method, the analytical model, and the Comsol finite element method are compared under different thicknesses and lift-offs, and it is found that the consistency of the PEC signals of these three methods is very well. And, the amplitudes of the residual signals between the BPNN method and the analytical model of the PEC signals are relatively small, and the relative errors at the peaks are no more than 0.4%.

Hybrid modeling of T-shaped partial least squares regression and transfer learning for formulation and manufacturing process development of new drug products

T-shaped partial least squares regression (T-PLSR) is a valuable machine learning technique for the formulation and manufacturing process development of new drug products. An accurate T-PLSR model requires experimental data with multiple formulations and process conditions. However, it is usually challenging to collect comprehensive experimental data using large-scale manufacturing equipment because of the cost, time, and large consumption of raw materials. This study proposes a hybrid modeling of T-PLSR and transfer learning (TL) to enhance the prediction performance of a T-PLSR model for large-scale manufacturing data by exploiting a large amount of small-scale manufacturing data for model building. The proposed method of T-PLSR+TL was applied to a practical case study focusing on scaling up the tableting process from an experienced compaction simulator to a less-experienced rotary tablet press. The T-PLSR+TL models achieved significantly better prediction performance for tablet quality attributes of new drug products than T-PLSR models without using large-scale manufacturing data with new drug products. The results demonstrated that T-PLSR+TL is more capable of addressing new drug products than T-PLSR by using small-scale manufacturing data to cover a scarcity of large-scale manufacturing data. Furthermore, T-PLSR+TL holds the potential to streamline formulation and manufacturing process development activities for new drug products using an extensive database.

Online welding status monitoring method of T-joint double-sided double arc welding based on multi-source information fusion

In modern manufacturing, double-sided double arc (DSDA) welding brings advantages in properties and efficiency for T-shaped joints. However, the double-side forming makes it difficult to detect the imperfections and ensure weld joint quality, which few studies have covered. Most existing methods deal with single-arc welding with limited sensing sources. To fill the gap, an online welding status monitoring method for DSDA welding is proposed based on the fusion of welding current, arc voltage, arc sound, and weld pool images. In pre-processing, the automatic weld pool region of interest (ROI) detection method based on the lightweight YOLO-L model and the waveform denoising algorithm are designed. In feature engineering, waveform signals are analyzed in both the time and frequency domain. A weld pool feature extractor based on a convolutional neural network (CNN) is proposed with good effectiveness and interpretability. The output feature combination is refined by Fisher-based selection and evaluation. In model building, an ensemble learning model based on three high-fit basic learners is proposed, with an accuracy of 98.538 %. The proposed model shows significant advantages over the single basic classifier and other ensemble methods. Experiments verify high precision and robustness, laying a foundation for accurate real-time monitoring of DSDA welding production.

Ensuring information security and stability: The application of blockchain in naming system

In recent years, network information security and stability problems have emerged one after another, and people's personal privacy and property security are being threatened. A large part of these problems are caused by two major characteristics of traditional DNSinsecurity and instability. Therefore, this article aims to explore a new internet naming system the Blockchain Naming Systemto address the current issues of internet information security and stabilityespecially the problems caused by the traditional DNS system. This article adopts scientific research methods such as literature survey, comparative research, experimental exploration, analogy research, and model building. By analyzing the current situation, studying the feasibility of the blockchain naming system, and comparing its advantages and disadvantages with the traditional DNS system, it is not difficult to conclude that this new blockchain naming system has huge potential. It can not only ensure stability and security but also contribute to other fields.

The role of the foundation flexibility on the seismic response of a modern tall building: Vertically incident plane waves

The soil-structure interaction (SSI) is an integral part of the seismic response of structures. The knowledge about its effects is based largely on numerical simulations involving simplified models. In this paper, we examine the consequence of several assumptions in SSI models of buildings, with emphasis on the rigid foundation assumption. To this end, we analyze several variants of a linear three-dimensional finite element model of the structure and soil corresponding to a uniquely instrumented 50-story skyscraper, with a basement on a pile foundation, located in southwest China. The models are excited by triaxial motion due to a vertically incident plane wave. The most realistic model (with basement and piles, both with bond contact with the soil), predicted an 8 % reduction of the apparent frequency and a 30 %–50 % increase in the apparent damping ratio of the 1st mode of vibration, relative to the model with fixed basement. The model with a rigid foundation (basement) embedded in the soil underestimated the reduction in apparent frequency by a factor of two. The SSI effects were the most pronounced for the model with flexible basement and no piles (16 %–19 % reduction in apparent frequency and 73 %–86 % increase in apparent damping ratio). The basement-structure interaction reduced the apparent frequency by 5 %–8 % depending on the direction. It is concluded that, for such a building, a detailed model of the structure and its foundation is needed to predict reliably the SSI effects.

Fragility and vulnerability models for poorly-detailed infilled buildings accounting for building-to-building and masonry infill variability

While in poorly-detailed masonry-infilled reinforced concrete (RC) buildings the significance of masonry infills is acknowledged, less attention has been paid to the variability surrounding their mechanical properties and its potential impact on fragility and, consequently, vulnerability models to be used in seismic risk estimation. This study addresses this gap by proposing updated fragility and vulnerability models of existing poorly-detailed masonry infilled and pilotis RC frames that consider the variability in building layout and infill properties. A large building portfolio, representative of those designed according to Italian codes between 1970 and 1980, is selected as case study, incorporating building-to-building variability through various geometrical configurations. Updated modelling uncertainty parameters, specifically derived for existing infilled RC frames, are used, including the variability surrounding the properties of different infill typologies and the occurrence of RC frame shear failure. The portfolio was analysed through multiple-stripe analyses, followed by the development of upgraded fragility models for buildings with different heights and geometrical configurations. Finally, the corresponding vulnerability models were derived using damage-to-loss models calibrated with post-earthquake observation damage data from the 2009 L'Aquila earthquake. The updated fragility and vulnerability curves can be used for more refined seismic risk estimates of large building portfolios, that account for building-to-building and infill-to-infill variability, revised modelling uncertainty estimates and a more robust intensity measure (AvgSa).