Search results for: prediction

Authors: Husham Bayazed

Abstract:

Immune responses following surgical trauma play a pivotal role in predicting postoperative outcomes from healing and recovery to postoperative complications. Postoperative complications, including infections and protracted recovery, occur in a significant number of about 300 million surgeries performed annually worldwide. Complications cause personal suffering along with a significant economic burden on the healthcare system in any community. The accurate prediction of postoperative complications and patient-targeted interventions for their prevention remain major clinical provocations. Recent Findings: Recent studies are focusing on immune dysregulation mechanisms that occur in response to surgical trauma as a key determinant of postoperative complications. Antecedent studies mainly were plunging into the detection of inflammatory plasma markers, which facilitate in providing important clues regarding their pathogenesis. However, recent Single-cell technologies, such as mass cytometry or single-cell RNA sequencing, have markedly enhanced our ability to understand the immunological basis of postoperative immunological trauma complications and to identify their prognostic biological signatures. Summary: The advent of proteomic technologies has significantly advanced our ability to predict the risk of postoperative complications. Multiomic modeling of patients' immune states holds promise for the discovery of preoperative predictive biomarkers and providing patients and surgeons with information to improve surgical outcomes. However, more studies are required to accurately predict the risk of postoperative complications in individual patients.

Keywords: immune dysregulation, postoperative complications, surgical trauma, flow cytometry

Procedia PDF Downloads 65

Authors: Geng Xiangren, Liu Lei, Gui Ye-Wei, Tang Wei, Wang An-ling

Abstract:

The malfunction of thermal protection system (TPS) caused by aerodynamic heating is a latent trouble to aircraft structure safety. Accurately predicting the structure temperature field is quite important for the TPS design of hypersonic vehicle. Since Thornton’s work in 1988, the coupled method of aerodynamic heating and heat transfer has developed rapidly. However, little attention has been paid to the influence of structural deformation on aerodynamic heating and structural temperature field. In the flight, especially the long-endurance flight, the structural deformation, caused by the aerodynamic heating and temperature rise, has a direct impact on the aerodynamic heating and structural temperature field. Thus, the coupled interaction cannot be neglected. In this paper, based on the method of static aero-thermo-elasticity, considering the influence of aero-thermo-elasticity deformation, the aerodynamic heating and heat transfer coupled results of hypersonic vehicle wing model were calculated. The results show that, for the low-curvature region, such as fuselage or center-section wing, structure deformation has little effect on temperature field. However, for the stagnation region with high curvature, the coupled effect is not negligible. Thus, it is quite important for the structure temperature prediction to take into account the effect of elastic deformation. This work has laid a solid foundation for improving the prediction accuracy of the temperature distribution of aircraft structures and the evaluation capacity of structural performance.

Keywords: aerothermoelasticity, elastic deformation, structural temperature, multi-field coupling

Procedia PDF Downloads 320

Authors: Nadish Anand, Richard D. Gould

Abstract:

A simplistic model is introduced for determining the thermal characteristics of a Low-rise Residential (LRR) building and then predicts the energy usage by its Heating Ventilation & Air Conditioning (HVAC) system according to changes in weather conditions which are reflected in the Ambient Temperature (Outside Air Temperature). The LRR buildings are treated as a simple lump for solving the heat transfer problem and the model is derived using the lumped capacitance model of transient conduction heat transfer from bodies. Since most contemporary HVAC systems have a thermostat control which will have an offset temperature and user defined set point temperatures which define when the HVAC system will switch on and off. The aim is to predict without any error the Body Temperature (i.e. the Inside Air Temperature) which will estimate the switching on and off of the HVAC system. To validate the mathematical model derived from lumped capacitance we have used EnergyPlus simulation engine, which simulates Buildings with considerable accuracy. We have predicted through the low order model the Inside Air Temperature of a single house kept in three different climate zones (Detroit, Raleigh & Austin) and different orientations for summer and winter seasons. The prediction error from the model for the same day as that of model parameter calculation has showed an error of < 10% in winter for almost all the orientations and climate zones. Whereas the prediction error is only <10% for all the orientations in the summer season for climate zone at higher latitudes (Raleigh & Detroit). Possible factors responsible for the large variations are also noted in the work, paving way for future research.

Keywords: building energy, energy consumption, energy+, HVAC, low order model, lumped capacitance

Procedia PDF Downloads 249

Authors: Said Alshukri, Mazhar Hussain Malik

Abstract:

Green hydrogen is hydrogen produced using renewable energy sources. In the last few years, Oman aimed to reduce its dependency on fossil fuels. Recently, the hydrogen economy has become a global trend, and many countries have started to investigate the feasibility of implementing this sector. Oman created an alliance to establish the policy and rules for this sector. With motivation coming from both global and local interest in green hydrogen, this paper investigates the potential of producing hydrogen from wind and solar energies in three different locations in Oman, namely Duqm, Salalah, and Sohar. By using machine learning-based software “WEKA” and local metrological data, the project was designed to figure out which location has the highest wind and solar energy potential. First, various supervised models were tested to obtain their prediction accuracy, and it was found that the Random Forest (RF) model has the best prediction performance. The RF model was applied to 2021 metrological data for each location, and the results indicated that Duqm has the highest wind and solar energy potential. The system of one wind turbine in Duqm can produce 8335 MWh/year, which could be utilized in the water electrolysis process to produce 88847 kg of hydrogen mass, while a solar system consisting of 2820 solar cells is estimated to produce 1666.223 MWh/ year which is capable of producing 177591 kg of hydrogen mass.

Keywords: green hydrogen, machine learning, wind and solar energies, WEKA, supervised models, random forest

Procedia PDF Downloads 50

Authors: R. O. Osaseri, A. R. Usiobaifo

Abstract:

The power transformer which is responsible for the voltage transformation is of great relevance in the power system and oil-immerse transformer is widely used all over the world. A prompt and proper maintenance of the transformer is of utmost importance. The dissolved gasses content in power transformer, oil is of enormous importance in detecting incipient fault of the transformer. There is a need for accurate prediction of the incipient fault in transformer oil in order to facilitate the prompt maintenance and reducing the cost and error minimization. Study on fault prediction and diagnostic has been the center of many researchers and many previous works have been reported on the use of artificial intelligence to predict incipient failure of transformer faults. In this study machine learning technique was employed by using gradient decent algorithms and Support Vector Machine (SVM) in predicting incipient fault diagnosis of transformer. The method focuses on creating a system that improves its performance on previous result and historical data. The system design approach is basically in two phases; training and testing phase. The gradient decent algorithm is trained with a training dataset while the learned algorithm is applied to a set of new data. This two dataset is used to prove the accuracy of the proposed model. In this study a transformer fault diagnostic model based on Support Vector Machine (SVM) and gradient decent algorithms has been presented with a satisfactory diagnostic capability with high percentage in predicting incipient failure of transformer faults than existing diagnostic methods.

Keywords: diagnostic model, gradient decent, machine learning, support vector machine (SVM), transformer fault

Procedia PDF Downloads 296

Authors: Saba Gachpaz, Hamid Reza Heidari

Abstract:

The sharp increase in population growth leads to more pressure on agricultural areas to satisfy the food supply. To achieve this, more resources should be consumed and, besides other environmental concerns, highlight sustainable agricultural development. Land-use management is a crucial factor in obtaining optimum productivity. Machine learning is a widely used technique in the agricultural sector, from yield prediction to customer behavior. This method focuses on learning and provides patterns and correlations from our data set. In this study, nine physical control factors, namely, soil classification, electrical conductivity, normalized difference water index (NDWI), groundwater level, elevation, annual precipitation, pH of water, annual mean temperature, and slope in the alluvial plain in Khuzestan (an agricultural hotspot in Iran) are used to decide the best agricultural land use for both rainfed and irrigated agriculture for ten different crops. For this purpose, each variable was imported into Arc GIS, and a raster layer was obtained. In the next level, by using training samples, all layers were imported into the python environment. A random forest model was applied, and the weight of each variable was specified. In the final step, results were visualized using a digital elevation model, and the importance of all factors for each one of the crops was obtained. Our results show that despite 62% of the study area being allocated to agricultural purposes, only 42.9% of these areas can be defined as a suitable class for cultivation purposes.

Keywords: land suitability, machine learning, random forest, sustainable agriculture

Procedia PDF Downloads 59

Authors: Ahmed Alghurabi, Mysara Mohyaldinn, Shiferaw Jufar, Obai Younis, Abdullah Abduljabbar

Abstract:

Erosion modeling equations were typically acquired from regulated experimental trials for solid particles entrained in single-phase or multi-phase flows. Evidently, those equations were later employed to predict the erosion damage caused by the continuous impacts of solid particles entrained in streamflow. It is also well-known that the particle impact angle and velocity do not change drastically in gas-sand flow erosion prediction; hence an accurate prediction of erosion can be projected. On the contrary, high-density fluid flows, such as water flow, through complex geometries, such as sand screens, greatly affect the sand particles’ trajectories/tracks and consequently impact the erosion rate predictions. Particle tracking models and erosion equations are frequently applied simultaneously as a method to improve erosion visualization and estimation. In the present work, computational fluid dynamic (CFD)-based erosion modeling was performed using a commercially available software; ANSYS Fluent. The continuous phase (water flow) behavior was simulated using the realizable K-epsilon model, and the secondary phase (solid particles), having a 5% flow concentration, was tracked with the help of the discrete phase model (DPM). To accomplish a successful erosion modeling, three erosion equations from the literature were utilized and introduced to the ANSYS Fluent software to predict the screen wire-slot velocity surge and estimate the maximum erosion rates on the screen surface. Results of turbulent kinetic energy, turbulence intensity, dissipation rate, the total pressure on the screen, screen wall shear stress, and flow velocity vectors were presented and discussed. Moreover, the particle tracks and path-lines were also demonstrated based on their residence time, velocity magnitude, and flow turbulence. On one hand, results from the utilized erosion equations have shown similarities in screen erosion patterns, locations, and DPM concentrations. On the other hand, the model equations estimated slightly different values of maximum erosion rates of the wire-wrapped screen. This is solely based on the fact that the utilized erosion equations were developed with some assumptions that are controlled by the experimental lab conditions.

Keywords: CFD simulation, erosion rate prediction, material loss due to erosion, water-sand flow

Procedia PDF Downloads 142

Authors: Silvia Arrate, Waldo Salud, Eloy París

Abstract:

The wear of cutting tools is one of the most decisive elements when planning tunneling works, programming the maintenance stops and saving the optimum stock of spare parts during the evolution of the excavation. Being able to predict the behavior of cutting tools can give a very competitive advantage in terms of costs and excavation performance, optimized to the needs of the TBM itself. The incredible evolution of data science in recent years gives the option to implement it at the time of analyzing the key and most critical parameters related to machinery with the purpose of knowing how the cutting head is performing in front of the excavated ground. Taking this as a case study, Metro Line 3 of Guadalajara in Mexico will develop the feasibility of using Specific Energy versus data science applied over parameters of Torque, Penetration, and Contact Force, among others, to predict the behavior and status of cutting tools. The results obtained through both techniques are analyzed and verified in the function of the wear and the field situations observed in the excavation in order to determine its effectiveness regarding its predictive capacity. In conclusion, the possibilities and improvements offered by the application of digital tools and the programming of calculation algorithms for the analysis of wear of cutting head elements compared to purely empirical methods allow early detection of possible damage to cutting tools, which is reflected in optimization of excavation performance and a significant improvement in costs and deadlines.

Keywords: cutting tools, data science, prediction, TBM, wear

Procedia PDF Downloads 28

Authors: Sophia Liang Zhou

Abstract:

For investors, businesses, consumers, and governments, an accurate assessment of future housing prices is crucial to critical decisions in resource allocation, policy formation, and investment strategies. Previous studies are contradictory about macroeconomic determinants of housing price and largely focused on one or two areas using point prediction. This study aims to develop data-driven models to accurately predict future housing market trends in different markets. This work studied five different metropolitan areas representing different market trends and compared three-time lagging situations: no lag, 6-month lag, and 12-month lag. Linear regression (LR), random forest (RF), and artificial neural network (ANN) were employed to model the real estate price using datasets with S&P/Case-Shiller home price index and 12 demographic and macroeconomic features, such as gross domestic product (GDP), resident population, personal income, etc. in five metropolitan areas: Boston, Dallas, New York, Chicago, and San Francisco. The data from March 2005 to December 2018 were collected from the Federal Reserve Bank, FBI, and Freddie Mac. In the original data, some factors are monthly, some quarterly, and some yearly. Thus, two methods to compensate missing values, backfill or interpolation, were compared. The models were evaluated by accuracy, mean absolute error, and root mean square error. The LR and ANN models outperformed the RF model due to RF’s inherent limitations. Both ANN and LR methods generated predictive models with high accuracy ( > 95%). It was found that personal income, GDP, population, and measures of debt consistently appeared as the most important factors. It also showed that technique to compensate missing values in the dataset and implementation of time lag can have a significant influence on the model performance and require further investigation. The best performing models varied for each area, but the backfilled 12-month lag LR models and the interpolated no lag ANN models showed the best stable performance overall, with accuracies > 95% for each city. This study reveals the influence of input variables in different markets. It also provides evidence to support future studies to identify the optimal time lag and data imputing methods for establishing accurate predictive models.

Keywords: linear regression, random forest, artificial neural network, real estate price prediction

Procedia PDF Downloads 81

Authors: Ahmed Z. Mazen, Nejat Rahmanian, Iqbal Mujtaba, Ali Hassanpour

Abstract:

In oil and gas well drilling, the drill bit is an important part of the Bottom Hole Assembly (BHA), which is installed and designed to drill and produce a hole by several mechanisms. The efficiency of the bit depends on many drilling parameters such as weight on bit, rotary speed, and mud properties. When the bit is pulled out of the hole, the evaluation of the bit damage must be recorded very carefully to guide engineers in order to select the bits for further planned wells. Having a worn bit for hole drilling may cause severe damage to bit leading to cutter or cone losses in the bottom of hole, where a fishing job will have to take place, and all of these will increase the operating cost. The main factor to reduce the cost of drilling operation is to maximize the rate of penetration by analyzing real-time data to predict the drill bit wear while drilling. There are numerous models in the literature for prediction of the rate of penetration based on drilling parameters, mostly based on empirical approaches. One of the most commonly used approaches is Bourgoyne and Young model, where the rate of penetration can be estimated by the drilling parameters as well as a wear index using an empirical correlation, provided all the constants and coefficients are accurately determined. This paper introduces a new methodology to estimate the eight coefficients for Bourgoyne and Young model using the gPROMS parameters estimation GPE (Version 4.2.0). Real data collected form similar formations (12 ¼’ sections) in two different fields in Libya are used to estimate the coefficients. The estimated coefficients are then used in the equations and applied to nearby wells in the same field to predict the bit wear.

Keywords: Bourgoyne and Young model, bit wear, gPROMS, rate of penetration

Procedia PDF Downloads 130

Authors: Alexander Heckman, Rohan Goswami, Zachi Attia, Paul Friedman, Peter Noseworthy, Demilade Adedinsewo, Pablo Moreno-Franco, Rickey Carter, Tathagat Narula

Abstract:

Background: Postoperative atrial fibrillation (POAF) is associated with adverse health consequences, higher costs, and longer hospital stays. Utilizing existing predictive models that rely on clinical variables and circulating biomarkers, multiple societies have published recommendations on the treatment and prevention of POAF. Although reasonably practical, there is room for improvement and automation to help individualize treatment strategies and reduce associated complications. Methods and Results: In this retrospective cohort study of solid organ transplant recipients, we evaluated the diagnostic utility of a previously developed AI-based ECG prediction for silent AF on the development of POAF within 30 days of transplant. A total of 2261 non-cardiac transplant patients without a preexisting diagnosis of AF were found to have a 5.8% (133/2261) incidence of POAF. While there were no apparent sex differences in POAF incidence (5.8% males vs. 6.0% females, p=.80), there were differences by race and ethnicity (p<0.001 and 0.035, respectively). The incidence in white transplanted patients was 7.2% (117/1628), whereas the incidence in black patients was 1.4% (6/430). Lung transplant recipients had the highest incidence of postoperative AF (17.4%, 37/213), followed by liver (5.6%, 56/1002) and kidney (3.6%, 32/895) recipients. The AUROC in the sample was 0.62 (95% CI: 0.58-0.67). The relatively low discrimination may result from undiagnosed AF in the sample. In particular, 1,177 patients had at least 1 AI-ECG screen for AF pre-transplant above .10, a value slightly higher than the published threshold of 0.08. The incidence of POAF in the 1104 patients without an elevated prediction pre-transplant was lower (3.7% vs. 8.0%; p<0.001). While this supported the hypothesis that potentially undiagnosed AF may have contributed to the diagnosis of POAF, the utility of the existing AI-ECG screening algorithm remained modest. When the prediction for POAF was made using the first postoperative ECG in the sample without an elevated screen pre-transplant (n=1084 on account of n=20 missing postoperative ECG), the AUROC was 0.66 (95% CI: 0.57-0.75). While this discrimination is relatively low, at a threshold of 0.08, the AI-ECG algorithm had a 98% (95% CI: 97 – 99%) negative predictive value at a sensitivity of 66% (95% CI: 49-80%). Conclusions: This study's principal finding is that the incidence of POAF is rare, and a considerable fraction of the POAF cases may be latent and undiagnosed. The high negative predictive value of AI-ECG screening suggests utility for prioritizing monitoring and evaluation on transplant patients with a positive AI-ECG screening. Further development and refinement of a post-transplant-specific algorithm may be warranted further to enhance the diagnostic yield of the ECG-based screening.

Keywords: artificial intelligence, atrial fibrillation, cardiology, transplant, medicine, ECG, machine learning

Procedia PDF Downloads 105

Authors: Koe Han Beng, Khoo Boo Cheong

Abstract:

The rising interest of stepped hull design has been led by the demand of more efficient high-speed boat. At the same time, the need of accurate prediction method for stepped planing hull is getting more important. By understanding the flow at high Froude number is the key in designing a practical step hull, the study surrounding stepped hull has been done mainly in the towing tank which is time-consuming and costly for initial design phase. Here the feasibility of predicting hydrodynamics of high-speed planing hull both with and without step using computational fluid dynamics (CFD) with the volume of fluid (VOF) methodology is studied in this work. First the flow around the prismatic body is analyzed, the force generated and its center of pressure are compared with available experimental and empirical data from the literature. The wake behind the transom on the keel line as well as the quarter beam buttock line are then compared with the available data, this is important since the afterbody flow of stepped hull is subjected from the wake of the forebody. Finally the calm water performance prediction of a conventional planing hull and its stepped version is then analyzed. Overset mesh methodology is employed in solving the dynamic equilibrium of the hull. The resistance, trim, and heave are then compared with the experimental data. The resistance is found to be predicted well and the dynamic equilibrium solved by the numerical method is deemed to be acceptable. This means that computational fluid dynamics will be very useful in further study on the complex flow around stepped hull and its potential usage in the design phase.

Keywords: planing hulls, stepped hulls, wake shape, numerical simulation, hydrodynamics

Procedia PDF Downloads 265

Authors: Muhammad Luthfi, Sutikno Sutikno, Purhadi Purhadi

Abstract:

Weather forecast has necessarily been improved to provide the communities an accurate and objective prediction as well. To overcome such issue, the numerical-based weather forecast was extensively developed to reduce the subjectivity of forecast. Yet the Numerical Weather Predictions (NWPs) outputs are unfortunately issued without taking dynamical weather behavior and local terrain features into account. Thus, NWPs outputs are not able to accurately forecast the weather quantities, particularly for medium and long range forecast. The aim of this research is to aid and extend the development of ensemble forecast for Meteorology, Climatology, and Geophysics Agency of Indonesia. Ensemble method is an approach combining various deterministic forecast to produce more reliable one. However, such forecast is biased and uncalibrated due to its underdispersive or overdispersive nature. As one of the parametric methods, Bayesian Model Averaging (BMA) generates the calibrated ensemble forecast and constructs predictive PDF for specified period. Such method is able to utilize ensemble of any size but does not take spatial correlation into account. Whereas space dependencies involve the site of interest and nearby site, influenced by dynamic weather behavior. Meanwhile, Geostatistical Output Perturbation (GOP) reckons the spatial correlation to generate future weather quantities, though merely built by a single deterministic forecast, and is able to generate an ensemble of any size as well. This research conducts both BMA and GOP to generate the calibrated ensemble forecast for the daily temperature at few meteorological sites nearby Indonesia international airport.

Keywords: Bayesian Model Averaging, ensemble forecast, geostatistical output perturbation, numerical weather prediction, temperature

Procedia PDF Downloads 257

Authors: Manuela Villani, Anila Xhahysa, Christopher Brooks, Marco Pagani

Abstract:

The geological structure of Western Balkans is strongly affected by the collision between Adria microplate and the southwestern Euroasia margin, resulting in a considerably active seismic region. The Harmonization of Seismic Hazard Maps in the Western Balkan Countries Project (BSHAP) (2007-2011, 2012-2015) by NATO supported the preparation of new seismic hazard maps of the Western Balkan, but when inspecting the seismic hazard models produced later by these countries on a national scale, significant differences in design PGA values are observed in the border, for instance, North Albania-Montenegro, South Albania- Greece, etc. Considering the fact that the catalogues were unified and seismic sources were defined within BSHAP framework, obviously, the differences arise from the Ground Motion Prediction Equations selection, which are generally the component with highest impact on the seismic hazard assessment. At the time of the project, a modest database was present, namely 672 three-component records, whereas nowadays, this strong motion database has increased considerably up to 20,939 records with Mw ranging in the interval 3.7-7 and epicentral distance distribution from 0.47km to 490km. Statistical analysis of the strong motion database showed the lack of recordings in the moderate-to-large magnitude and short distance ranges; therefore, there is need to re-evaluate the Ground Motion Prediction Equation in light of the recently updated database and the new generations of GMMs. In some cases, it was observed that some events were more extensively documented in one database than the other, like the 1979 Montenegro earthquake, with a considerably larger number of records in the BSHAP Analogue SM database when compared to ESM23. Therefore, the strong motion flat-file provided from the Harmonization of Seismic Hazard Maps in the Western Balkan Countries Project was merged with the ESM23 database for the polygon studied in this project. After performing the preliminary residual analysis, the candidate GMPE-s were identified. This process was done using the GMPE performance metrics available within the SMT in the OpenQuake Platform. The Likelihood Model and Euclidean Distance Based Ranking (EDR) were used. Finally, for this study, a GMPE logic tree was selected and following the selection of candidate GMPEs, model weights were assigned using the average sample log-likelihood approach of Scherbaum.

Keywords: residual analysis, GMPE, western balkan, strong motion, openquake

Procedia PDF Downloads 54

Authors: Abdolghani Ebrahimi, Diego Klabjan, Chenxi Ge, Daniela Ladner, Parker Stride

Abstract:

In this work, we use machine learning and novel data analysis techniques to predict the one-year mortality of cirrhotic patients. Data from 2,322 patients with liver cirrhosis are collected at a single medical center. Different machine learning models are applied to predict one-year mortality. A comprehensive feature space including demographic information, comorbidity, clinical procedure and laboratory tests is being analyzed. A temporal pattern mining technic called Frequent Subgraph Mining (FSM) is being used. Model for End-stage liver disease (MELD) prediction of mortality is used as a comparator. All of our models statistically significantly outperform the MELD-score model and show an average 10% improvement of the area under the curve (AUC). The FSM technic itself does not improve the model significantly, but FSM, together with a machine learning technique called an ensemble, further improves the model performance. With the abundance of data available in healthcare through electronic health records (EHR), existing predictive models can be refined to identify and treat patients at risk for higher mortality. However, due to the sparsity of the temporal information needed by FSM, the FSM model does not yield significant improvements. To the best of our knowledge, this is the first work to apply modern machine learning algorithms and data analysis methods on predicting one-year mortality of cirrhotic patients and builds a model that predicts one-year mortality significantly more accurate than the MELD score. We have also tested the potential of FSM and provided a new perspective of the importance of clinical features.

Keywords: machine learning, liver cirrhosis, subgraph mining, supervised learning

Procedia PDF Downloads 113

Authors: Yan Zhang

Abstract:

Predictive maintenance is a technique to predict when an in-service machine will fail so that maintenance can be planned in advance. Analytics-driven predictive maintenance is gaining increasing attention in many industries such as manufacturing, utilities, aerospace, etc., along with the emerging demand of Internet of Things (IoT) applications and the maturity of technologies that support Big Data storage and processing. This study aims to build an end-to-end analytics solution that includes both real-time machine condition monitoring and machine learning based predictive analytics capabilities. The goal is to showcase a general predictive maintenance solution architecture, which suggests how the data generated from field machines can be collected, transmitted, stored, and analyzed. We use a publicly available aircraft engine run-to-failure dataset to illustrate the streaming analytics component and the batch failure prediction component. We outline the contributions of this study from four aspects. First, we compare the predictive maintenance problems from the view of the traditional reliability centered maintenance field, and from the view of the IoT applications. When evolving to the IoT era, predictive maintenance has shifted its focus from ensuring reliable machine operations to improve production/maintenance efficiency via any maintenance related tasks. It covers a variety of topics, including but not limited to: failure prediction, fault forecasting, failure detection and diagnosis, and recommendation of maintenance actions after failure. Second, we review the state-of-art technologies that enable a machine/device to transmit data all the way through the Cloud for storage and advanced analytics. These technologies vary drastically mainly based on the power source and functionality of the devices. For example, a consumer machine such as an elevator uses completely different data transmission protocols comparing to the sensor units in an environmental sensor network. The former may transfer data into the Cloud via WiFi directly. The latter usually uses radio communication inherent the network, and the data is stored in a staging data node before it can be transmitted into the Cloud when necessary. Third, we illustrate show to formulate a machine learning problem to predict machine fault/failures. By showing a step-by-step process of data labeling, feature engineering, model construction and evaluation, we share following experiences: (1) what are the specific data quality issues that have crucial impact on predictive maintenance use cases; (2) how to train and evaluate a model when training data contains inter-dependent records. Four, we review the tools available to build such a data pipeline that digests the data and produce insights. We show the tools we use including data injection, streaming data processing, machine learning model training, and the tool that coordinates/schedules different jobs. In addition, we show the visualization tool that creates rich data visualizations for both real-time insights and prediction results. To conclude, there are two key takeaways from this study. (1) It summarizes the landscape and challenges of predictive maintenance applications. (2) It takes an example in aerospace with publicly available data to illustrate each component in the proposed data pipeline and showcases how the solution can be deployed as a live demo.

Keywords: Internet of Things, machine learning, predictive maintenance, streaming data

Procedia PDF Downloads 364

Authors: Cristian Pauna

Abstract:

Trading the financial markets is a widespread activity today. A large number of investors, companies, public of private funds are buying and selling every day in order to make profit. Algorithmic trading is the prevalent method to make the trade decisions after the electronic trading release. The orders are sent almost instantly by computers using mathematical models. This paper will present a price prediction methodology based on a multi-dimensional neural network. Using the Fisher transform, the neural network will be instructed for a low-latency auto-adaptive process in order to predict the price evolution for the next period of time. The model is designed especially for algorithmic trading and uses the real-time price series. It was found that the characteristics of the Fisher function applied at the nodes scale level can generate reliable trading signals using the neural network methodology. After real time tests it was found that this method can be applied in any timeframe to trade the financial markets. The paper will also include the steps to implement the presented methodology into an automated trading system. Real trading results will be displayed and analyzed in order to qualify the model. As conclusion, the compared results will reveal that the neural network methodology applied together with the Fisher transform at the nodes level can generate a good price prediction and can build reliable trading signals for algorithmic trading.

Keywords: algorithmic trading, automated trading systems, financial markets, high-frequency trading, neural network

Procedia PDF Downloads 138

Authors: Raptis Sotirios

Abstract:

Health and social care (HSc) services planning and scheduling are facing unprecedented challenges due to the pandemic pressure and also suffer from unplanned spending that is negatively impacted by the global financial crisis. Data-driven can help to improve policies, plan and design services provision schedules using algorithms assist healthcare managers’ to face unexpected demands using fewer resources. The paper discusses services packing using statistical significance tests and machine learning (ML) to evaluate demands similarity and coupling. This is achieved by predicting the range of the demand (class) using ML methods such as CART, random forests (RF), and logistic regression (LGR). The significance tests Chi-Squared test and Student test are used on data over a 39 years span for which HSc services data exist for services delivered in Scotland. The demands are probabilistically associated through statistical hypotheses that assume that the target service’s demands are statistically dependent on other demands as a NULL hypothesis. This linkage can be confirmed or not by the data. Complementarily, ML methods are used to linearly predict the above target demands from the statistically found associations and extend the linear dependence of the target’s demand to independent demands forming, thus groups of services. Statistical tests confirm ML couplings making the prediction also statistically meaningful and prove that a target service can be matched reliably to other services, and ML shows these indicated relationships can also be linear ones. Zero paddings were used for missing years records and illustrated better such relationships both for limited years and in the entire span offering long term data visualizations while limited years groups explained how well patients numbers can be related in short periods or can change over time as opposed to behaviors across more years. The prediction performance of the associations is measured using Receiver Operating Characteristic(ROC) AUC and ACC metrics as well as the statistical tests, Chi-Squared and Student. Co-plots and comparison tables for RF, CART, and LGR as well as p-values and Information Exchange(IE), are provided showing the specific behavior of the ML and of the statistical tests and the behavior using different learning ratios. The impact of k-NN and cross-correlation and C-Means first groupings is also studied over limited years and the entire span. It was found that CART was generally behind RF and LGR, but in some interesting cases, LGR reached an AUC=0 falling below CART, while the ACC was as high as 0.912, showing that ML methods can be confused padding or by data irregularities or outliers. On average, 3 linear predictors were sufficient, LGR was found competing RF well, and CART followed with the same performance at higher learning ratios. Services were packed only if when significance level(p-value) of their association coefficient was more than 0.05. Social factors relationships were observed between home care services and treatment of old people, birth weights, alcoholism, drug abuse, and emergency admissions. The work found that different HSc services can be well packed as plans of limited years, across various services sectors, learning configurations, as confirmed using statistical hypotheses.

Keywords: class, cohorts, data frames, grouping, prediction, prob-ability, services

Procedia PDF Downloads 207

Authors: Erika Yamaguchi, Sirawadee Arunyanrt, Shunichi Ohmori, Kazuho Yoshimoto

Abstract:

In this paper, we suggest a simulation tool to make a decision of monthly production planning for maximizing a profit of Specialty store retailer of Private label Apparel (SPA) firms. Most of SPA firms are fabless and make outsourcing deals for productions with factories of their subcontractors. Every month, SPA firms make a booking for production lines and manpower in the factories. The booking is conducted a few months in advance based on a demand prediction and a monthly production planning at that time. However, the demand prediction is updated month by month, and the monthly production planning would change to meet the latest demand prediction. Then, SPA firms have to change the capacities initially booked within a certain range to suit to the monthly production planning. The booking system is called “capacity-booking”. These days, though it is an issue for SPA firms to make precise monthly production planning, many firms are still conducting the production planning by empirical rules. In addition, it is also a challenge for SPA firms to match their products and factories with considering their demand predictabilities and regulation abilities. In this paper, we suggest a model for considering these two issues. An objective is to maximize a total profit of certain periods, which is sales minus costs of production, inventory, and capacity-booking penalty. To make a better monthly production planning at SPA firms, these points should be considered: demand predictabilities by random trends, previous and next month’s production planning of the target month, and regulation abilities of the capacity-booking. To decide matching products and factories for outsourcing, it is important to consider seasonality, volume, and predictability of each product, production possibility, size, and regulation ability of each factory. SPA firms have to consider these constructions and decide orders with several factories per one product. We modeled these issues as a linear programming. To validate the model, an example of several computational experiments with a SPA firm is presented. We suppose four typical product groups: basic, seasonal (Spring / Summer), seasonal (Fall / Winter), and spot product. As a result of the experiments, a monthly production planning was provided. In the planning, demand predictabilities from random trend are reduced by producing products which are different product types. Moreover, priorities to produce are given to high-margin products. In conclusion, we developed a simulation tool to make a decision of monthly production planning which is useful when the production planning is set every month. We considered the features of capacity-booking, and matching of products and factories which have different features and conditions.

Keywords: capacity-booking, SPA, monthly production planning, linear programming

Procedia PDF Downloads 501

Authors: Aid Abdelkrim

Abstract:

A method of fatigue damage accumulation based upon application of energy parameters of the fatigue process is proposed in the paper. Using this model is simple, it has no parameter to be determined, it requires only the knowledge of the curve W–N (W: strain energy density N: number of cycles at failure) determined from the experimental Wöhler curve. To examine the performance of nonlinear models proposed in the estimation of fatigue damage and fatigue life of components under random loading, a batch of specimens made of 6082 T 6 aluminium alloy has been studied and some of the results are reported in the present paper. The paper describes an algorithm and suggests a fatigue cumulative damage model, especially when random loading is considered. This work contains the results of uni-axial random load fatigue tests with different mean and amplitude values performed on 6082T6 aluminium alloy specimens. The proposed model has been formulated to take into account the damage evolution at different load levels and it allows the effect of the loading sequence to be included by means of a recurrence formula derived for multilevel loading, considering complex load sequences. It is concluded that a ‘damaged stress interaction damage rule’ proposed here allows a better fatigue damage prediction than the widely used Palmgren–Miner rule, and a formula derived in random fatigue could be used to predict the fatigue damage and fatigue lifetime very easily. The results obtained by the model are compared with the experimental results and those calculated by the most fatigue damage model used in fatigue (Miner’s model). The comparison shows that the proposed model, presents a good estimation of the experimental results. Moreover, the error is minimized in comparison to the Miner’s model.

Keywords: damage accumulation, energy model, damage indicator, variable loading, random loading

Procedia PDF Downloads 374

Authors: Mustafa Salman, Nurcan Cetinkaya, Zehra Selcuk, Bugra Genc

Abstract:

The objectives of the present study were to estimate the microbial-N flow to the small intestine and to predict the digestible organic matter intake (DOMI) in grazing Karayaka sheep based on urinary excretion of purine derivatives (xanthine, hypoxanthine, uric acid, and allantoin) by the use of spot urine sampling under field conditions. In the trial, 10 Karayaka sheep from 2 to 3 years of age were used. The animals were grazed in a pasture for ten months and fed with concentrate and vetch plus oat hay for the other two months (January and February) indoors. Highly significant linear and cubic relationships (P<0.001) were found among months for purine derivatives index, purine derivatives excretion, purine derivatives absorption, microbial-N and DOMI. Through urine sampling and the determination of levels of excreted urinary PD and Purine Derivatives / Creatinine ratio (PDC index), microbial-N values were estimated and they indicated that the protein nutrition of the sheep was insufficient. In conclusion, the prediction of protein nutrition of sheep under the field conditions may be possible with the use of spot urine sampling, urinary excreted PD and PDC index. The mean purine derivative levels in spot urine samples from sheep were highest in June, July and October. Protein nutrition of pastured sheep may be affected by weather changes, including rainfall. Spot urine sampling may useful in modeling the feed consumption of pasturing sheep. However, further studies are required under different field conditions with different breeds of sheep to develop spot urine sampling as a model.

Keywords: Karayaka sheep, spot sampling, urinary purine derivatives, PDC index, microbial-N, feed intake

Procedia PDF Downloads 508

Authors: M. D. Haneef, R. B. Randall, Z. Peng

Abstract:

Journal bearings used in IC engines are prone to premature failures and are likely to fail earlier than the rated life due to highly impulsive and unstable operating conditions and frequent starts/stops. Vibration signature extraction and wear debris analysis techniques are prevalent in the industry for condition monitoring of rotary machinery. However, both techniques involve a great deal of technical expertise, time and cost. Limited literature is available on the application of these techniques for fault detection in reciprocating machinery, due to the complex nature of impact forces that confounds the extraction of fault signals for vibration based analysis and wear prediction. This work is an extension of a previous study, in which an engine simulation model was developed using a MATLAB/SIMULINK program, whereby the engine parameters used in the simulation were obtained experimentally from a Toyota 3SFE 2.0 litre petrol engines. Simulated hydrodynamic bearing forces were used to estimate vibrations signals and envelope analysis was carried out to analyze the effect of speed, load and clearance on the vibration response. Three different loads 50/80/110 N-m, three different speeds 1500/2000/3000 rpm, and three different clearances, i.e., normal, 2 times and 4 times the normal clearance were simulated to examine the effect of wear on bearing forces. The magnitude of the squared envelope of the generated vibration signals though not affected by load, but was observed to rise significantly with increasing speed and clearance indicating the likelihood of augmented wear. In the present study, the simulation model was extended further to investigate the bearing wear behavior, resulting as a consequence of different operating conditions, to complement the vibration analysis. In the current simulation, the dynamics of the engine was established first, based on which the hydrodynamic journal bearing forces were evaluated by numerical solution of the Reynold’s equation. Also, the essential outputs of interest in this study, critical to determine wear rates are the tangential velocity and oil film thickness between the journal and bearing sleeve, which if not maintained appropriately, have a detrimental effect on the bearing performance. Archard’s wear prediction model was used in the simulation to calculate the wear rate of bearings with specific location information as all determinative parameters were obtained with reference to crank rotation. Oil film thickness obtained from the model was used as a criterion to determine if the lubrication is sufficient to prevent contact between the journal and bearing thus causing accelerated wear. A limiting value of 1 µm was used as the minimum oil film thickness needed to prevent contact. The increased wear rate with growing severity of operating conditions is analogous and comparable to the rise in amplitude of the squared envelope of the referenced vibration signals. Thus on one hand, the developed model demonstrated its capability to explain wear behavior and on the other hand it also helps to establish a correlation between wear based and vibration based analysis. Therefore, the model provides a cost-effective and quick approach to predict the impending wear in IC engine bearings under various operating conditions.

Keywords: condition monitoring, IC engine, journal bearings, vibration analysis, wear prediction

Procedia PDF Downloads 294

Authors: Muhsin Kizhisseri, Jorg Schluter, Saleh Gharie

Abstract:

Atherosclerosis is the main reason of stroke, which is one of the deadliest diseases in the world. The carotid artery in the brain is the prominent location for atherosclerotic progression, which hinders the blood flow into the brain. The inclusion of computational fluid dynamics (CFD) into the diagnosis cycle to understand the hemodynamics of the patient-specific carotid artery can give insights into stroke prediction. Realistic outlet boundary conditions are an inevitable part of the numerical simulations, which is one of the major factors in determining the accuracy of the CFD results. The Windkessel model-based outlet boundary conditions can give more realistic characteristics of the distal vascular branches of the carotid artery, such as the resistance to the blood flow and compliance of the distal arterial walls. This study aims to find the most influential distal branches of the carotid artery by using the Windkessel model parameters in the outlet boundary conditions. The parametric study approach to Windkessel model parameters can include the geometrical features of the distal branches, such as radius and length. The incorporation of the variations of the geometrical features of the major distal branches such as the middle cerebral artery, anterior cerebral artery, and ophthalmic artery through the Windkessel model can aid in identifying the most influential distal branch in the carotid artery. The results from this study can help physicians and stroke neurologists to have a more detailed and accurate judgment of the patient's condition.

Keywords: stroke, carotid artery, computational fluid dynamics, patient-specific, Windkessel model, distal vascular branches

Procedia PDF Downloads 190

Authors: Pei Yi Lin

Abstract:

Objective: The objective of this study is to use machine learning methods to build an early prediction classifier model for acute delirium to improve the quality of medical care for intensive care patients. Background: Delirium is a common acute and sudden disturbance of consciousness in critically ill patients. After the occurrence, it is easy to prolong the length of hospital stay and increase medical costs and mortality. In 2021, the incidence of delirium in the intensive care unit of internal medicine was as high as 59.78%, which indirectly prolonged the average length of hospital stay by 8.28 days, and the mortality rate is about 2.22% in the past three years. Therefore, it is expected to build a delirium prediction classifier through big data analysis and machine learning methods to detect delirium early. Method: This study is a retrospective study, using the artificial intelligence big data database to extract the characteristic factors related to delirium in intensive care unit patients and let the machine learn. The study included patients aged over 20 years old who were admitted to the intensive care unit between May 1, 2022, and December 31, 2022, excluding GCS assessment <4 points, admission to ICU for less than 24 hours, and CAM-ICU evaluation. The CAMICU delirium assessment results every 8 hours within 30 days of hospitalization are regarded as an event, and the cumulative data from ICU admission to the prediction time point are extracted to predict the possibility of delirium occurring in the next 8 hours, and collect a total of 63,754 research case data, extract 12 feature selections to train the model, including age, sex, average ICU stay hours, visual and auditory abnormalities, RASS assessment score, APACHE-II Score score, number of invasive catheters indwelling, restraint and sedative and hypnotic drugs. Through feature data cleaning, processing and KNN interpolation method supplementation, a total of 54595 research case events were extracted to provide machine learning model analysis, using the research events from May 01 to November 30, 2022, as the model training data, 80% of which is the training set for model training, and 20% for the internal verification of the verification set, and then from December 01 to December 2022 The CU research event on the 31st is an external verification set data, and finally the model inference and performance evaluation are performed, and then the model has trained again by adjusting the model parameters. Results: In this study, XG Boost, Random Forest, Logistic Regression, and Decision Tree were used to analyze and compare four machine learning models. The average accuracy rate of internal verification was highest in Random Forest (AUC=0.86), and the average accuracy rate of external verification was in Random Forest and XG Boost was the highest, AUC was 0.86, and the average accuracy of cross-validation was the highest in Random Forest (ACC=0.77). Conclusion: Clinically, medical staff usually conduct CAM-ICU assessments at the bedside of critically ill patients in clinical practice, but there is a lack of machine learning classification methods to assist ICU patients in real-time assessment, resulting in the inability to provide more objective and continuous monitoring data to assist Clinical staff can more accurately identify and predict the occurrence of delirium in patients. It is hoped that the development and construction of predictive models through machine learning can predict delirium early and immediately, make clinical decisions at the best time, and cooperate with PADIS delirium care measures to provide individualized non-drug interventional care measures to maintain patient safety, and then Improve the quality of care.

Keywords: critically ill patients, machine learning methods, delirium prediction, classifier model

Procedia PDF Downloads 45

Authors: Vadim A. Kuznetsov, Anna M. Soldatova, Tatyana N. Enina, Elena A. Gorbatenko, Dmitrii V. Krinochkin

Abstract:

The aim of the study was to evaluate potential parameters related with super-response to CRT. Methods: 60 CRT patients (mean age 54.3 ± 9.8 years; 80% men) with congestive heart failure (CHF) II-IV NYHA functional class, left ventricular ejection fraction < 35% were enrolled. At baseline, 1 month, 3 months and each 6 months after implantation clinical, electrocardiographic and echocardiographic parameters, NT-proBNP level were evaluated. According to the best decrease of left ventricular end-systolic volume (LVESV) (mean follow-up period 33.7 ± 15.1 months) patients were classified as super-responders (SR) (n=28; reduction in LVESV ≥ 30%) and non-SR (n=32; reduction in LVESV < 30%). Results: At baseline groups differed in age (58.1 ± 5.8 years in SR vs 50.8 ± 11.4 years in non-SR; p=0.003), gender (female gender 32.1% vs 9.4% respectively; p=0.028), width of QRS complex (157.6 ± 40.6 ms in SR vs 137.6 ± 33.9 ms in non-SR; p=0.044). Percentage of LBBB was equal between groups (75% in SR vs 59.4% in non-SR; p=0.274). All parameters of mechanical dyssynchrony were higher in SR, but only difference in left ventricular pre-ejection period (LVPEP) was statistically significant (153.0 ± 35.9 ms vs. 129.3 ± 28.7 ms p=0.032). NT-proBNP level was lower in SR (1581 ± 1369 pg/ml vs 3024 ± 2431 pg/ml; p=0.006). The survival rates were 100% in SR and 90.6% in non-SR (log-rank test P=0.002). Multiple logistic regression analysis showed that LVPEP (HR 1.024; 95% CI 1.004–1.044; P = 0.017), baseline NT-proBNP level (HR 0.628; 95% CI 0.414–0.953; P=0.029) and age at baseline (HR 1.094; 95% CI 1.009-1.168; P=0.30) were independent predictors for CRT super-response. ROC curve analysis demonstrated sensitivity 71.9% and specificity 82.1% (AUC=0.827; p < 0.001) of this model in prediction of super-response to CRT. Conclusion: Super-response to CRT is associated with better survival in long-term period. Presence of LBBB was not associated with super-response. LVPEP, NT-proBNP level, and age at baseline can be used as independent predictors of CRT super-response.

Keywords: cardiac resynchronisation therapy, superresponse, congestive heart failure, left bundle branch block

Procedia PDF Downloads 372

Authors: Lule Basha, Eralda Gjika

Abstract:

This study is focused on analyzing climate change in Albania and its potential effects on cereal yields. Initially, monthly temperature and rainfalls in Albania were studied for the period 1960-2021. Climacteric variables are important variables when trying to model cereal yield behavior, especially when significant changes in weather conditions are observed. For this purpose, in the second part of the study, linear and nonlinear models explaining cereal yield are constructed for the same period, 1960-2021. The multiple linear regression analysis and lasso regression method are applied to the data between cereal yield and each independent variable: average temperature, average rainfall, fertilizer consumption, arable land, land under cereal production, and nitrous oxide emissions. In our regression model, heteroscedasticity is not observed, data follow a normal distribution, and there is a low correlation between factors, so we do not have the problem of multicollinearity. Machine-learning methods, such as random forest, are used to predict cereal yield responses to climacteric and other variables. Random Forest showed high accuracy compared to the other statistical models in the prediction of cereal yield. We found that changes in average temperature negatively affect cereal yield. The coefficients of fertilizer consumption, arable land, and land under cereal production are positively affecting production. Our results show that the Random Forest method is an effective and versatile machine-learning method for cereal yield prediction compared to the other two methods.

Keywords: cereal yield, climate change, machine learning, multiple regression model, random forest

Procedia PDF Downloads 68

Authors: Marc Solé, Francesc Giné, Magda Valls, Nina Bijedic

Abstract:

What people say on social media has turned into a rich source of information to understand social behavior. Specifically, the growing use of Twitter social media for political communication has arisen high opportunities to know the opinion of large numbers of politically active individuals in real time and predict the global political tendencies of a specific country. It has led to an increasing body of research on this topic. The majority of these studies have been focused on polarized political contexts characterized by only two alternatives. Unlike them, this paper tackles the challenge of forecasting Spanish political trends, characterized by multiple political parties, by means of analyzing the Twitters Users political tendency. According to this, a new strategy, named Tweets Analysis Strategy (TAS), is proposed. This is based on analyzing the users tweets by means of discovering its sentiment (positive, negative or neutral) and classifying them according to the political party they support. From this individual political tendency, the global political prediction for each political party is calculated. In order to do this, two different strategies for analyzing the sentiment analysis are proposed: one is based on Positive and Negative words Matching (PNM) and the second one is based on a Neural Networks Strategy (NNS). The complete TAS strategy has been performed in a Big-Data environment. The experimental results presented in this paper reveal that NNS strategy performs much better than PNM strategy to analyze the tweet sentiment. In addition, this research analyzes the viability of the TAS strategy to obtain the global trend in a political context make up by multiple parties with an error lower than 23%.

Keywords: political tendency, prediction, sentiment analysis, Twitter

Procedia PDF Downloads 214

Authors: Yuexin Liu, Gordon B. Mills, Russell R. Broaddus, John N. Weinstein

Abstract:

The lethality of endometrioid endometrial cancer (EEC) is primarily attributable to the high-stage diseases. However, there are no available biomarkers that predict EEC patient staging at the time of diagnosis. We aim to develop a predictive scheme to help in this regards. Using reverse-phase protein array expression profiles for 210 EEC cases from The Cancer Genome Atlas (TCGA), we constructed a Protein Scoring of EEC Staging (PSES) scheme for surgical stage prediction. We validated and evaluated its diagnostic potential in an independent cohort of 184 EEC cases obtained at MD Anderson Cancer Center (MDACC) using receiver operating characteristic curve analyses. Kaplan-Meier survival analysis was used to examine the association of PSES score with patient outcome, and Ingenuity pathway analysis was used to identify relevant signaling pathways. Two-sided statistical tests were used. PSES robustly distinguished high- from low-stage tumors in the TCGA cohort (area under the ROC curve [AUC]=0.74; 95% confidence interval [CI], 0.68 to 0.82) and in the validation cohort (AUC=0.67; 95% CI, 0.58 to 0.76). Even among grade 1 or 2 tumors, PSES was significantly higher in high- than in low-stage tumors in both the TCGA (P = 0.005) and MDACC (P = 0.006) cohorts. Patients with positive PSES score had significantly shorter progression-free survival than those with negative PSES in the TCGA (hazard ratio [HR], 2.033; 95% CI, 1.031 to 3.809; P = 0.04) and validation (HR, 3.306; 95% CI, 1.836 to 9.436; P = 0.0007) cohorts. The ErbB signaling pathway was most significantly enriched in the PSES proteins and downregulated in high-stage tumors. PSES may provide clinically useful prediction of high-risk tumors and offer new insights into tumor biology in EEC.

Keywords: endometrial carcinoma, protein, protein scoring of EEC staging (PSES), stage

Procedia PDF Downloads 202

Authors: Anuruddha Jayasuriya, Thanakorn Pheeraphan

Abstract:

In this paper, a review of different mathematical models which can be used as prediction tools to assess the time to crack reinforced concrete (RC) due to corrosion is investigated. This investigation leads to an experimental study to validate a selected prediction model. Most of these mathematical models depend upon the mechanical behaviors, chemical behaviors, electrochemical behaviors or geometric aspects of the RC members during a corrosion process. The experimental program is designed to verify the accuracy of a well-selected mathematical model from a rigorous literature study. Fundamentally, the experimental program exemplifies both one-dimensional chloride diffusion using RC squared slab elements of 500 mm by 500 mm and two-dimensional chloride diffusion using RC squared column elements of 225 mm by 225 mm by 500 mm. Each set consists of three water-to-cement ratios (w/c); 0.4, 0.5, 0.6 and two cover depths; 25 mm and 50 mm. 12 mm bars are used for column elements and 16 mm bars are used for slab elements. All the samples are subjected to accelerated chloride corrosion in a chloride bath of 5% (w/w) sodium chloride (NaCl) solution. Based on a pre-screening of different models, it is clear that the well-selected mathematical model had included mechanical properties, chemical and electrochemical properties, nature of corrosion whether it is accelerated or natural, and the amount of porous area that rust products can accommodate before exerting expansive pressure on the surrounding concrete. The experimental results have shown that the selected model for both one-dimensional and two-dimensional chloride diffusion had ±20% and ±10% respective accuracies compared to the experimental output. The half-cell potential readings are also used to see the corrosion probability, and experimental results have shown that the mass loss is proportional to the negative half-cell potential readings that are obtained. Additionally, a statistical analysis is carried out in order to determine the most influential factor that affects the time to corrode the reinforcement in the concrete due to chloride diffusion. The factors considered for this analysis are w/c, bar diameter, and cover depth. The analysis is accomplished by using Minitab statistical software, and it showed that cover depth is the significant effect on the time to crack the concrete from chloride induced corrosion than other factors considered. Thus, the time predictions can be illustrated through the selected mathematical model as it covers a wide range of factors affecting the corrosion process, and it can be used to predetermine the durability concern of RC structures that are vulnerable to chloride exposure. And eventually, it is further concluded that cover thickness plays a vital role in durability in terms of chloride diffusion.

Keywords: accelerated corrosion, chloride diffusion, corrosion cracks, passivation layer, reinforcement corrosion

Procedia PDF Downloads 197

Authors: Tahani Aljohani, Jialin Yu, Alexandra. I. Cristea

Abstract:

The more an educational system knows about a learner, the more personalised interaction it can provide, which leads to better learning. However, asking a learner directly is potentially disruptive, and often ignored by learners. Especially in the booming realm of MOOC Massive Online Learning platforms, only a very low percentage of users disclose demographic information about themselves. Thus, in this paper, we aim to predict learners’ demographic characteristics, by proposing an approach using linguistically motivated Deep Learning Architectures for Learner Profiling, particularly targeting gender prediction on a FutureLearn MOOC platform. Additionally, we tackle here the difficult problem of predicting the gender of learners based on their comments only – which are often available across MOOCs. The most common current approaches to text classification use the Long Short-Term Memory (LSTM) model, considering sentences as sequences. However, human language also has structures. In this research, rather than considering sentences as plain sequences, we hypothesise that higher semantic - and syntactic level sentence processing based on linguistics will render a richer representation. We thus evaluate, the traditional LSTM versus other bleeding edge models, which take into account syntactic structure, such as tree-structured LSTM, Stack-augmented Parser-Interpreter Neural Network (SPINN) and the Structure-Aware Tag Augmented model (SATA). Additionally, we explore using different word-level encoding functions. We have implemented these methods on Our MOOC dataset, which is the most performant one comparing with a public dataset on sentiment analysis that is further used as a cross-examining for the models' results.

Keywords: deep learning, data mining, gender predication, MOOCs

Procedia PDF Downloads 117