Search results for: numerical weather prediction

Authors: Neoclis Hadjigeorgiou, Spyridon Angelopoulos, Evangelos V. Hristoforou, Paul P. Sotiriadis

Abstract:

The increasing demand for accurate and reliable magnetic measurements over the past decades has paved the way for the development of different types of magnetic sensing systems as well as of more advanced measurement techniques. Anisotropic Magneto Resistance (AMR) sensors have emerged as a promising solution for applications requiring high resolution, providing an ideal balance between performance and cost. However, certain issues of AMR sensors such as non-linear response and measurement noise are rarely discussed in the relevant literature. In this work, an analog closed loop compensation system is proposed, developed and tested as a means to eliminate the non-linearity of AMR response, reduce the 1/f noise and enhance the sensitivity of magnetic sensor. Additional performance aspects, such as cross-axis and hysteresis effects are also examined. This system was analyzed using an analytical model and a P-Spice model, considering both the sensor itself as well as the accompanying electronic circuitry. In addition, a commercial closed loop architecture Flux-Gate sensor (calibrated and certified), has been used for comparison purposes. Three different experimental setups have been constructed for the purposes of this work, each one utilized for DC magnetic field measurements, AC magnetic field measurements and Noise density measurements respectively. The DC magnetic field measurements have been conducted in laboratory environment employing a cubic Helmholtz coil setup in order to calibrate and characterize the system under consideration. A high-accuracy DC power supply has been used for providing the operating current to the Helmholtz coils. The results were recorded by a multichannel voltmeter The AC magnetic field measurements have been conducted in laboratory environment employing a cubic Helmholtz coil setup in order to examine the effective bandwidth not only of the proposed system but also for the Flux-Gate sensor. A voltage controlled current source driven by a function generator has been utilized for the Helmholtz coil excitation. The result was observed by the oscilloscope. The third experimental apparatus incorporated an AC magnetic shielding construction composed of several layers of electric steel that had been demagnetized prior to the experimental process. Each sensor was placed alone and the response was captured by the oscilloscope. The preliminary experimental results indicate that closed loop AMR response presented a maximum deviation of 0.36% with respect to the ideal linear response, while the corresponding values for the open loop AMR system and the Fluxgate sensor reached 2% and 0.01% respectively. Moreover, the noise density of the proposed close loop AMR sensor system remained almost as low as the noise density of the AMR sensor itself, yet considerably higher than that of the Flux-Gate sensor. All relevant numerical data are presented in the paper.

Keywords: AMR sensor, chopper, closed loop, electronic noise, magnetic noise, memory effects, flux-gate sensor, linearity improvement, sensitivity improvement

Procedia PDF Downloads 407

Authors: Corentin Jankowiak, Aggelos Zacharopoulos, Caterina Brandoni

Abstract:

In a context of increasing stress put on the electricity network by the decarbonization of many sectors, energy storage is likely to be the key mitigating element, by acting as a buffer between production and demand. In particular, the highest potential for storage is when connected closer to the loads. Yet, low voltage storage struggles to penetrate the market at a large scale due to the novelty and complexity of the solution, and the competitive advantage of fossil fuel-based technologies regarding regulations. Strong and reliable numerical simulations are required to show the benefits of storage located near loads and promote its development. The present study was restrained from excluding aggregated control of storage: it is assumed that the storage units operate independently to one another without exchanging information – as is currently mostly the case. A computationally light battery model is presented in detail and validated by direct comparison with a domestic battery operating in real conditions. This model is then used to develop Peak-Shaving (PS) control strategies as it is the decentralized service from which beneficial impacts are most likely to emerge. The aggregation of flatter, peak- shaved consumption profiles is likely to lead to flatter and arbitraged profile at higher voltage layers. Furthermore, voltage fluctuations can be expected to decrease if spikes of individual consumption are reduced. The crucial part to achieve PS lies in the charging pattern: peaks depend on the switching on and off of appliances in the dwelling by the occupants and are therefore impossible to predict accurately. A performant PS strategy must, therefore, include a smart charge recovery algorithm that can ensure enough energy is present in the battery in case it is needed without generating new peaks by charging the unit. Three categories of PS algorithms are introduced in detail. First, using a constant threshold or power rate for charge recovery, followed by algorithms using the State Of Charge (SOC) as a decision variable. Finally, using a load forecast – of which the impact of the accuracy is discussed – to generate PS. A performance metrics was defined in order to quantitatively evaluate their operating regarding peak reduction, total energy consumption, and self-consumption of domestic photovoltaic generation. The algorithms were tested on load profiles with a 1-minute granularity over a 1-year period, and their performance was assessed regarding these metrics. The results show that constant charging threshold or power are far from optimal: a certain value is not likely to fit the variability of a residential profile. As could be expected, forecast-based algorithms show the highest performance. However, these depend on the accuracy of the forecast. On the other hand, SOC based algorithms also present satisfying performance, making them a strong alternative when the reliable forecast is not available.

Keywords: decentralised control, domestic integrated batteries, electricity network performance, peak-shaving algorithm

Procedia PDF Downloads 102

Authors: Mozhdeh Shirinzadeh, Richard Stroetmann

Abstract:

Orthotropic steel deck bridge consists of a deck plate, longitudinal stiffeners under the deck plate, cross beams and the main longitudinal girders. Due to the several advantages, Orthotropic Steel Deck (OSD) systems have been utilized in many bridges worldwide. The significant feature of this structural system is its high load-bearing capacity while having relatively low dead weight. In addition, cost efficiency and the ability of rapid field erection have made the orthotropic steel deck a popular type of bridge worldwide. However, OSD bridges are highly susceptible to fatigue damage. A large number of welded joints can be regarded as the main weakness of this system. This problem is, in particular, evident in the bridges which were built before 1994 when the fatigue design criteria had not been introduced in the bridge design codes. Recently, an Orthotropic-composite slab (OCS) for road bridges has been experimentally and numerically evaluated and developed at Technische Universität Dresden as a part of AIF-FOSTA research project P1265. The results of the project have provided a solid foundation for the design and analysis of Orthotropic-composite decks with dowel strips as a durable alternative to conventional steel or reinforced concrete decks. In continuation, while using the achievements of that project, the application of a modified Ortho-composite deck for an existing typical OSD bridge is investigated. Composite action is obtained by using rows of dowel strips in a clothoid (CL) shape. Regarding Eurocode criteria for different fatigue detail categories of an OSD bridge, the effect of the proposed modification approach is assessed. Moreover, a numerical parametric study is carried out utilizing finite element software to determine the impact of different variables, such as the size and arrangement of dowel strips, the application of transverse or longitudinal rows of dowel strips, and local wheel loads. For the verification of the simulation technique, experimental results of a segment of an OCS deck are used conducted in project P1265. Fatigue assessment is performed based on the last draft of Eurocode 1993-2 (2024) for the most probable detail categories (Hot-Spots) that have been reported in the previous statistical studies. Then, an analytical comparison is provided between the typical orthotropic steel deck and the modified Ortho-composite deck bridge in terms of fatigue issues and durability. The load-bearing capacity of the bridge, the critical deflections, and the composite behavior are also evaluated and compared. Results give a comprehensive overview of the efficiency of the rehabilitation method considering the required design service life of the bridge. Moreover, the proposed approach is assessed with regard to the construction method, details and practical aspects, as well as the economic point of view.

Keywords: composite action, fatigue, finite element method, steel deck, bridge

Procedia PDF Downloads 58

Authors: Anna Makuch, Grzegorz Kokot, Konstanty Skalski, Jakub Banczorowski

Abstract:

Cancellous bone is a porous composite of a hierarchical structure and anisotropic properties. The biological tissue is considered to be a viscoelastic material, but many studies based on a nanoindentation method have focused on their elasticity and microhardness. However, the response of many organic materials depends not only on the load magnitude, but also on its duration and time course. Depth Sensing Indentation (DSI) technique has been used for examination of creep in polymers, metals and composites. In the indentation tests on biological samples, the mechanical properties are most frequently determined for animal tissues (of an ox, a monkey, a pig, a rat, a mouse, a bovine). However, there are rare reports of studies of the bone viscoelastic properties on microstructural level. Various rheological models were used to describe the viscoelastic behaviours of bone, identified in the indentation process (e. g Burgers model, linear model, two-dashpot Kelvin model, Maxwell-Voigt model). The goal of the study was to determine the influence of creep effect on the mechanical properties of human cancellous bone in indentation tests. The aim of this research was also the assessment of the material properties of bone structures, having in mind the energy aspects of the curve (penetrator loading-depth) obtained in the loading/unloading cycle. There was considered how the different holding times affected the results within trabecular bone.As a result, indentation creep (CIT), hardness (HM, HIT, HV) and elasticity are obtained. Human trabecular bone samples (n=21; mean age 63±15yrs) from the femoral heads replaced during hip alloplasty were removed and drained from alcohol of 1h before the experiment. The indentation process was conducted using CSM Microhardness Tester equipped with Vickers indenter. Each sample was indented 35 times (7 times for 5 different hold times: t1=0.1s, t2=1s, t3=10s, t4=100s and t5=1000s). The indenter was advanced at a rate of 10mN/s to 500mN. There was used Oliver-Pharr method in calculation process. The increase of hold time is associated with the decrease of hardness parameters (HIT(t1)=418±34 MPa, HIT(t2)=390±50 MPa, HIT(t3)= 313±54 MPa, HIT(t4)=305±54 MPa, HIT(t5)=276±90 MPa) and elasticity (EIT(t1)=7.7±1.2 GPa, EIT(t2)=8.0±1.5 GPa, EIT(t3)=7.0±0.9 GPa, EIT(t4)=7.2±0.9 GPa, EIT(t5)=6.2±1.8 GPa) as well as with the increase of the elastic (Welastic(t1)=4.11∙10-7±4.2∙10-8Nm, Welastic(t2)= 4.12∙10-7±6.4∙10-8 Nm, Welastic(t3)=4.71∙10-7±6.0∙10-9 Nm, Welastic(t4)= 4.33∙10-7±5.5∙10-9Nm, Welastic(t5)=5.11∙10-7±7.4∙10-8Nm) and inelastic (Winelastic(t1)=1.05∙10-6±1.2∙10-7 Nm, Winelastic(t2) =1.07∙10-6±7.6∙10-8 Nm, Winelastic(t3)=1.26∙10-6±1.9∙10-7Nm, Winelastic(t4)=1.56∙10-6± 1.9∙10-7 Nm, Winelastic(t5)=1.67∙10-6±2.6∙10-7)) reaction of materials. The indentation creep increased logarithmically (R2=0.901) with increasing hold time: CIT(t1) = 0.08±0.01%, CIT(t2) = 0.7±0.1%, CIT(t3) = 3.7±0.3%, CIT(t4) = 12.2±1.5%, CIT(t5) = 13.5±3.8%. The pronounced impact of creep effect on the mechanical properties of human cancellous bone was observed in experimental studies. While the description elastic-inelastic, and thus the Oliver-Pharr method for data analysis, may apply in few limited cases, most biological tissues do not exhibit elastic-inelastic indentation responses. Viscoelastic properties of tissues may play a significant role in remodelling. The aspect is still under an analysis and numerical simulations. Acknowledgements: The presented results are part of the research project founded by National Science Centre (NCN), Poland, no.2014/15/B/ST7/03244.

Keywords: bone, creep, indentation, mechanical properties

Procedia PDF Downloads 150

Authors: Nicola Milani, Carla Bellissimo, Davide Pogliana, Davide Panzin, Luca Garlaschelli, Giulia Facchinetti, Claudia Casson, Luca Marazzina, Andrea Sartori, Simone Rivaroli, Jeff Konin

Abstract:

The patient is a 26 year-old male amateur football player from Milan, Italy; (81kg; 185cm; BMI 23.6 kg/m²). He sustained a non-contact anterior cruciate ligament tear to his right knee in June 2021. In September 2021, his right knee ligament was reconstructed using a semitendinosus graft. The injury occurred during a football match on natural grass with typical shoes on a warm day (32 degrees celsius). Playing as a defender he sustained the injury during a change of direction, where the foot was fixated on the grass. He felt pain and was unable to continue playing the match. The surgeon approved his rehabilitation to begin two weeks post-operative. The initial physiotherapist assessment determined performing two training sessions per day within the first three months. In the first three weeks, the pain was 4/10 on Numerical Rating Scale (NRS), no swelling, a range of motion was 0-110°, with difficulty fully extending his knee and minimal quadriceps activation. Crutches were discontinued at four weeks with improved walking. Active exercise, electrostimulator, physical therapy, massages, osteopathy, and passive motion were initiated. At week 6, he completed his first functional movement screen; the score was 16/21 with no pain and no swelling. At week 8, the isokinetic test showed a 23% differential deficit between the two legs in maximum strength (at 90°/s). At week 10, he improved to 15% of injury-induced deficit which suggested he was ready to start running. At week 12, the athlete sustained his first threshold test. At week 16, he performed his first return to sports movement assessment, which revealed a 10% stronger difference between the legs. At week 16, he had his second threshold test. At week 17, his first on-field test revealed a 5% differential deficit between the two legs in the hop test. At week 18, isokinetic test demonstrates that the uninjured leg was 7% stronger than the recovering leg in maximum strength (at 90°/s). At week 20, his second on-field test revealed a 2% difference in hop test; at week 21, his third isokinetic test demonstrated a difference of 5% in maximum strength (at 90°/s). At week 21, he performed his second return to sports movement assessment which revealed a 2% difference between the limbs. Since it was the end of the championship, the team asked him to partake in the playoffs; moreover the player was very motivated to participate in the playoffs also because he was the captain of the team. Together with the player and the team, we decided to let him play even though we were aware of a heightened risk of injury than what is reported in the literature because of two factors: biological recovery times and the results of the tests we performed. In the decision making process about the athlete’s recovery time, it is important to balance the information available from the literature with the desires of the patient to avoid frustration.

Keywords: ACL, football, rehabilitation, return to play

Procedia PDF Downloads 98

Authors: Anita Bajpai, Sarah Duan, Ashlee Erskine, Shehzein Khan, Raymond Kramer

Abstract:

Introduction: This research focuses on exploring the beneficial effects if any, of Vitamin-D in reducing the need for supplemental oxygen among hospitalized COVID-19 patients. Two questions are investigated – Q1)Doeshaving a healthy level of baselineVitamin-D 25-OH (≥ 30ng/ml) help,andQ2) does administering Vitamin-D therapy after-the-factduring inpatient hospitalization help? Methods/Study Design: This is a comprehensive, retrospective, observational study of all inpatients at RUHS from March through December 2020 who tested positive for COVID-19 based on real-time reverse transcriptase–polymerase chain reaction assay of nasal and pharyngeal swabs and rapid assay antigen test. To address Q1, we looked atall N1=182 patients whose baseline plasma Vitamin-D 25-OH was known and who needed supplemental oxygen. Of this, a total of 121 patients had a healthy Vitamin-D level of ≥30 ng/mlwhile the remaining 61 patients had low or borderline (≤ 29.9ng/ml)level. Similarly, for Q2, we looked at a total of N2=893 patients who were given supplemental oxygen, of which713 were not given Vitamin-D and 180 were given Vitamin-D therapy. The numerical value of the maximum amount of oxygen flow rate(dependent variable) administered was recorded for each patient. The mean values and associated standard deviations for each group were calculated. Thesetwo sets of independent data served as the basis for independent, two-sample t-Test statistical analysis. To be accommodative of any reasonable benefitof Vitamin-D, ap-value of 0.10(α< 10%) was set as the cutoff point for statistical significance. Results: Given the large sample sizes, the calculated statistical power for both our studies exceeded the customary norm of 80% or better (β< 0.2). For Q1, the mean value for maximumoxygen flow rate for the group with healthybaseline level of Vitamin-D was 8.6 L/min vs.12.6L/min for those with low or borderline levels, yielding a p-value of 0.07 (p < 0.10) with the conclusion that those with a healthy level of baseline Vitamin-D needed statistically significant lower levels of supplemental oxygen. ForQ2, the mean value for a maximum oxygen flow rate for those not administered Vitamin-Dwas 12.5 L/min vs.12.8L/min for those given Vitamin-D, yielding a p-valueof 0.87 (p > 0.10). We thereforeconcludedthat there was no statistically significant difference in the use of oxygen therapy between those who were or were not administered Vitamin-D after-the-fact in the hospital. Discussion/Conclusion: We found that patients who had healthy levels of Vitamin-D at baseline needed statistically significant lower levels of supplemental oxygen. Vitamin-D is well documented, including in a recent article in the Lancet, for its anti-inflammatory role as an adjuvant in the regulation of cytokines and immune cells. Interestingly, we found no statistically significant advantage for giving Vitamin-D to hospitalized patients. It may be a case of “too little too late”. A randomized clinical trial reported in JAMA also did not find any reduction in hospital stay of patients given Vitamin-D. Such conclusions come with a caveat that any delayed marginal benefits may not have materialized promptly in the presence of a significant inflammatory condition. Since Vitamin-D is a low-cost, low-risk option, it may still be useful on an inpatient basis until more definitive findings are established.

Keywords: COVID-19, vitamin-D, supplemental oxygen, vitamin-D in primary care

Procedia PDF Downloads 135

Authors: Jiujie Cai, Fengxia LI, Haibo Wang

Abstract:

After years of research, offshore oil and gas development now are shifted to unconventional reservoirs, where multi-stage hydraulic fracturing technology has been widely used. However, the simulation of complex hydraulic fractures in tight reservoirs is faced with geological and engineering difficulties, such as large burial depths, sand-shale interbeds, and complex stress barriers. The objective of this work is to simulate the hydraulic fracture propagation in the tight sandstone matrix of the marine-continental transitional reservoirs, where the Shanxi Formation in Tianhuan syncline of the Dongsheng gas field was used as the research target. The characteristic parameters of the vertical rock samples with rich beddings were clarified through rock mechanics experiments. The influence of rock mechanical parameters, vertical stress difference of pay-zone and bedding layer, and fracturing parameters (such as injection rates, fracturing fluid viscosity, and number of perforation clusters within single stage) on fracture initiation and propagation were investigated. In this paper, a 3-D fracture propagation model was built to investigate the complex fracture propagation morphology by boundary element method, considering the strength of bonding surface between layers, vertical stress difference and fracturing parameters (such as injection rates, fluid volume and viscosity). The research results indicate that on the condition of vertical stress difference (3 MPa), the fracture height can break through and enter the upper interlayer when the thickness of the overlying bedding layer is 6-9 m, considering effect of the weak bonding surface between layers. The fracture propagates within the pay zone when overlying interlayer is greater than 13 m. Difference in fluid volume distribution between clusters could be more than 20% when the stress difference of each cluster in the segment exceeds 2MPa. Fracture cluster in high stress zones cannot initiate when the stress difference in the segment exceeds 5MPa. The simulation results of fracture height are much higher if the effect of weak bonding surface between layers is not involved. By increasing the injection rates, increasing fracturing fluid viscosity, and reducing the number of clusters within single stage can promote the fracture height propagation through layers. Optimizing the perforation position and reducing the number of perforations can promote the uniform expansion of fractures. Typical curves of fracture height estimation were established for the tight sandstone of the Lower Permian Shanxi Formation. The model results have good consistency with micro-seismic monitoring results of hydraulic fracturing in Well 1HF.

Keywords: fracture propagation, boundary element method, fracture height, offshore oil and gas, marine-continental transitional reservoirs, rock mechanics experiment

Procedia PDF Downloads 101

Authors: Jelvin Tom Sebastian, Vinod Yeldho Baby

Abstract:

In space, electronic devices are constantly attacked with radiation, which causes certain parts to fail or behave in unpredictable ways. To advance the thermal controllability for microsatellites, we need a new approach and thermal control system that is smaller than that on conventional satellites and that demand no electric power. Heat exchange inside the microsatellites is not that easy as conventional satellites due to the smaller size. With slight mass gain and no electric power, accommodating heat using phase change materials (PCMs) is a strong candidate for solving micro satellites' thermal difficulty. In other words, PCMs can absorb or produce heat in the form of latent heat, changing their phase and minimalizing the temperature fluctuation around the phase change point. The main restriction for these systems is thermal conductivity weakness of common PCMs. As PCM is having low thermal conductivity, it increases the melting and solidification time, which is not suitable for specific application like electronic cooling. In order to increase the thermal conductivity nanoparticles are introduced. Adding the nanoparticles in base PCM increases the thermal conductivity. Increase in weight concentration increases the thermal conductivity. This paper numerically investigates the thermal energy storage panel with nanoparticle enhanced phase change material. Silver nanostructure have increased the thermal properties of the base PCM, eicosane. Different weight concentration (1, 2, 3.5, 5, 6.5, 8, 10%) of silver enhanced phase change material was considered. Both steady state and transient analysis was performed to compare the characteristics of nanoparticle enhanced phase material at different heat loads. Results showed that in steady state, the temperature near the front panel reduced and temperature on NePCM panel increased as the weight concentration increased. With the increase in thermal conductivity more heat was absorbed into the NePCM panel. In transient analysis, it was found that the effect of nanoparticle concentration on maximum temperature of the system was reduced as the melting point of the material reduced with increase in weight concentration. But for the heat load of maximum 20W, the model with NePCM did not attain the melting point temperature. Therefore it showed that the model with NePCM is capable of holding more heat load. In order to study the heat load capacity double the load is given, maximum of 40W was given as first half of the cycle and the other is given constant OW. Higher temperature was obtained comparing the other heat load. The panel maintained a constant temperature for a long duration according to the NePCM melting point. In both the analysis, the uniformity of temperature of the TESP was shown. Using Ag-NePCM it allows maintaining a constant peak temperature near the melting point. Therefore, by altering the weight concentration of the Ag-NePCM it is possible to create an optimum operating temperature required for the effective working of the electronics components.

Keywords: carbon-fiber-reinforced polymer, micro/nano-satellite, nanoparticle phase change material, thermal energy storage

Procedia PDF Downloads 191

Authors: Kalaumari Mayoral-Peña, Ana I. Montejano-Montelongo, Josué Reyes-Muñoz, Gonzalo A. Ortiz-Mancilla, Mayrin Rodríguez-Cruz, Víctor Hernández-Villalobos, Jesús A. Guzmán-López, Santiago García-Jacobo, Iván Licona-Vázquez, Grisel Fierros-Romero, Rosario Flores-Vallejo

Abstract:

The group B-lactamic antibiotics include some of the most frequently used small drug molecules against bacterial infections. Nevertheless, an alarming decrease in their efficacy has been reported due to the emergence of antibiotic-resistant bacteria. Infections caused by bacteria expressing extended Spectrum B-lactamases (ESBLs) are difficult to treat and account for higher morbidity and mortality rates, delayed recovery, and high economic burden. According to the Global Report on Antimicrobial Resistance Surveillance, it is estimated that mortality due to resistant bacteria will ascend to 10 million cases per year worldwide. These facts highlight the importance of developing low-cost and readily accessible detection methods of drug-resistant ESBLs bacteria to prevent their spread and promote accurate and fast diagnosis. Bacterial detection is commonly done using molecular diagnostic techniques, where PCR stands out for its high performance. However, this technique requires specialized equipment not available everywhere, is time-consuming, and has a high cost. Loop-Mediated Isothermal Amplification (LAMP) is an alternative technique that works at a constant temperature, significantly decreasing the equipment cost. It yields double-stranded DNA of several lengths with repetitions of the target DNA sequence as a product. Although positive and negative results from LAMP can be discriminated by colorimetry, fluorescence, and turbidity, there is still a large room for improvement in the point-of-care implementation. DNA origami is a technique that allows the formation of 3D nanometric structures by folding a large single-stranded DNA (scaffold) into a determined shape with the help of short DNA sequences (staples), which hybridize with the scaffold. This research aimed to generate DNA origami structures using LAMP products as scaffolds to improve the sensitivity to detect ESBLs in point-of-care diagnosis. For this study, the coding sequence of the CTM-X-15 ESBL of E. coli was used to generate the LAMP products. The set of LAMP primers were designed using PrimerExplorerV5. As a result, a target sequence of 200 nucleotides from CTM-X-15 ESBL was obtained. Afterward, eight different DNA origami structures were designed using the target sequence in the SDCadnano and analyzed with CanDo to evaluate the stability of the 3D structures. The designs were constructed minimizing the total number of staples to reduce costs and complexity for point-of-care applications. After analyzing the DNA origami designs, two structures were selected. The first one was a zig-zag flat structure, while the second one was a wall-like shape. Given the sequence repetitions in the scaffold sequence, both were able to be assembled with only 6 different staples each one, ranging between 18 to 80 nucleotides. Simulations of both structures were performed using scaffolds of different sizes yielding stable structures in all the cases. The generation of the LAMP products were tested by colorimetry and electrophoresis. The formation of the DNA structures was analyzed using electrophoresis and colorimetry. The modeling of novel detection methods through bioinformatics tools allows reliable control and prediction of results. To our knowledge, this is the first study that uses LAMP products and DNA-origami in combination to delect ESBL-producing bacterial strains, which represent a promising methodology for diagnosis in the point-of-care.

Keywords: beta-lactamases, antibiotic resistance, DNA origami, isothermal amplification, LAMP technique, molecular diagnosis

Procedia PDF Downloads 193

Authors: Philipp Sommer, Amgad Agoub

Abstract:

The construction sector significantly impacts global CO₂ emissions, particularly through the energy usage of residential buildings. To address this, various governments, including Germany's, are focusing on reducing emissions via sustainable refurbishment initiatives. This study examines the application of machine learning (ML) to estimate energy demands dynamically in residential buildings and enhance the potential for large-scale sustainable refurbishment. A major challenge in Germany is the lack of extensive publicly labeled datasets for energy performance, as energy performance certificates, which provide critical data on building-specific energy requirements and consumption, are not available for all buildings or require on-site inspections. Conversely, England and other countries in the European Union (EU) have rich public datasets, providing a viable alternative for analysis. This research adapts insights from these English datasets to the German context by developing a comprehensive data schema and calibration dataset capable of predicting building energy demand effectively. The study proposes a minimal feature set, determined through feature importance analysis, to optimize the ML model. Findings indicate that ML significantly improves the scalability and accuracy of energy demand forecasts, supporting more effective emissions reduction strategies in the construction industry. Integrating energy performance certificates into municipal heat planning in Germany highlights the transformative impact of data-driven approaches on environmental sustainability. The goal is to identify and utilize key features from open data sources that significantly influence energy demand, creating an efficient forecasting model. Using Extreme Gradient Boosting (XGB) and data from energy performance certificates, effective features such as building type, year of construction, living space, insulation level, and building materials were incorporated. These were supplemented by data derived from descriptions of roofs, walls, windows, and floors, integrated into three datasets. The emphasis was on features accessible via remote sensing, which, along with other correlated characteristics, greatly improved the model's accuracy. The model was further validated using SHapley Additive exPlanations (SHAP) values and aggregated feature importance, which quantified the effects of individual features on the predictions. The refined model using remote sensing data showed a coefficient of determination (R²) of 0.64 and a mean absolute error (MAE) of 4.12, indicating predictions based on efficiency class 1-100 (G-A) may deviate by 4.12 points. This R² increased to 0.84 with the inclusion of more samples, with wall type emerging as the most predictive feature. After optimizing and incorporating related features like estimated primary energy consumption, the R² score for the training and test set reached 0.94, demonstrating good generalization. The study concludes that ML models significantly improve prediction accuracy over traditional methods, illustrating the potential of ML in enhancing energy efficiency analysis and planning. This supports better decision-making for energy optimization and highlights the benefits of developing and refining data schemas using open data to bolster sustainability in the building sector. The study underscores the importance of supporting open data initiatives to collect similar features and support the creation of comparable models in Germany, enhancing the outlook for environmental sustainability.

Keywords: machine learning, remote sensing, residential building, energy performance certificates, data-driven, heat planning

Procedia PDF Downloads 26

Authors: Sara Rejeb, Catherine Duveau, Tabea Rebafka

Abstract:

To monitor the production process of turbofan aircraft engines, multiple measurements of various geometrical parameters are systematically recorded on manufactured parts. Engine parts are subject to extremely high standards as they can impact the performance of the engine. Therefore, it is essential to analyze these databases to better understand the influence of the different parameters on the engine's performance. Self-organizing maps are unsupervised neural networks which achieve two tasks simultaneously: they visualize high-dimensional data by projection onto a 2-dimensional map and provide clustering of the data. This technique has become very popular for data exploration since it provides easily interpretable results and a meaningful global view of the data. As such, self-organizing maps are usually applied to aircraft engine condition monitoring. As databases in this field are huge and complex, they naturally contain multiple missing entries for various reasons. The classical Kohonen algorithm to compute self-organizing maps is conceived for complete data only. A naive approach to deal with partially observed data consists in deleting items or variables with missing entries. However, this requires a sufficient number of complete individuals to be fairly representative of the population; otherwise, deletion leads to a considerable loss of information. Moreover, deletion can also induce bias in the analysis results. Alternatively, one can first apply a common imputation method to create a complete dataset and then apply the Kohonen algorithm. However, the choice of the imputation method may have a strong impact on the resulting self-organizing map. Our approach is to address simultaneously the two problems of computing a self-organizing map and imputing missing values, as these tasks are not independent. In this work, we propose an extension of self-organizing maps for partially observed data, referred to as missSOM. First, we introduce a criterion to be optimized, that aims at defining simultaneously the best self-organizing map and the best imputations for the missing entries. As such, missSOM is also an imputation method for missing values. To minimize the criterion, we propose an iterative algorithm that alternates the learning of a self-organizing map and the imputation of missing values. Moreover, we develop an accelerated version of the algorithm by entwining the iterations of the Kohonen algorithm with the updates of the imputed values. This method is efficiently implemented in R and will soon be released on CRAN. Compared to the standard Kohonen algorithm, it does not come with any additional cost in terms of computing time. Numerical experiments illustrate that missSOM performs well in terms of both clustering and imputation compared to the state of the art. In particular, it turns out that missSOM is robust to the missingness mechanism, which is in contrast to many imputation methods that are appropriate for only a single mechanism. This is an important property of missSOM as, in practice, the missingness mechanism is often unknown. An application to measurements on one type of part is also provided and shows the practical interest of missSOM.

Keywords: imputation method of missing data, partially observed data, robustness to missingness mechanism, self-organizing maps

Procedia PDF Downloads 131

Authors: De-Shin Liu, Po-Chun Wen, Zhen-Wei Zhuang, Hsueh-Chih Liu, Pei-Chen Huang

Abstract:

Wafer probing tests play an important role in semiconductor manufacturing procedures in accordance with the yield and reliability requirement of the wafer after the backend-of-the-line process. Accordingly, the stable physical and electrical contact between the probe and the tested wafer during wafer probing is regarded as an essential issue in identifying the known good die. The probe card can be integrated with multiple probe needles, which are classified as vertical, cantilever and micro-electro-mechanical systems type probe selections. Among all potential probe types, the vertical probe has several advantages as compared with other probe types, including maintainability, high probe density and feasibility for high-speed wafer testing. In the present study, the mechanical response of the wafer probing test with the vertical cobra probe on 720 μm thick silicon (Si) substrate with a 1.4 μm thick aluminum (Al) pad is investigated by the experiment and transient dynamic simulation approach. Because the deformation mechanism of the vertical cobra probe is determined by both bending and buckling mechanisms, the stable correlation between contact forces and overdrive (OD) length must be carefully verified. Moreover, the decent OD length with corresponding contact force contributed to piercing the native oxide layer of the Al pad and preventing the probing test-induced damage on the interconnect system. Accordingly, the scratch depth of the Al pad under various OD lengths is estimated by the atomic force microscope (AFM) and simulation work. In the wafer probing test configuration, the contact phenomenon between the probe needle and the tested object introduced large deformation and twisting of mesh gridding, causing the subsequent numerical divergence issue. For this reason, the arbitrary Lagrangian-Eulerian method is utilized in the present simulation work to conquer the aforementioned issue. The analytic results revealed a slight difference when the OD is considered as 40 μm, and the simulated is almost identical to the measured scratch depths of the Al pad under higher OD lengths up to 70 μm. This phenomenon can be attributed to the unstable contact of the probe at low OD length with the scratch depth below 30% of Al pad thickness, and the contact status will be being stable when the scratch depth over 30% of pad thickness. The splash of the Al pad is observed by the AFM, and the splashed Al debris accumulates on a specific side; this phenomenon is successfully simulated in the transient dynamic simulation. Thus, the preferred testing OD lengths are found as 45 μm to 70 μm, and the corresponding scratch depths on the Al pad are represented as 31.4% and 47.1% of Al pad thickness, respectively. The investigation approach demonstrated in this study contributed to analyzing the mechanical response of wafer probing test configuration under large strain conditions and assessed the geometric designs and material selections of probe needles to meet the requirement of high resolution and high-speed wafer-level probing test for thinned wafer application.

Keywords: wafer probing test, vertical probe, probe mark, mechanical response, FEA simulation

Procedia PDF Downloads 37

Authors: Denis Jordan, Daniel Golkowski, Mathias Lukas, Katharina Merz, Caroline Mlynarcik, Max Maurer, Valentin Riedl, Stefan Foerster, Eberhard F. Kochs, Andreas Bender, Ruediger Ilg

Abstract:

Introduction: So far, clinical assessments that rely on behavioral responses to differentiate coma states or even predict outcome in coma patients are unreliable, e.g. because of some patients’ motor disabilities. The present study was aimed to provide prognosis in coma patients using markers from electroencephalogram (EEG), blood oxygen level dependent (BOLD) functional magnetic resonance imaging (fMRI) and [18F]-fluorodeoxyglucose (FDG) positron emission tomography (PET). Unsuperwised principal component analysis (PCA) was used for multimodal integration of markers. Methods: Approved by the local ethics committee of the Technical University of Munich (Germany) 20 patients (aged 18-89) with severe brain damage were acquired through intensive care units at the Klinikum rechts der Isar in Munich and at the Therapiezentrum Burgau (Germany). At the day of EEG/fMRI/PET measurement (date I) patients (<3.5 month in coma) were grouped in the minimal conscious state (MCS) or vegetative state (VS) on the basis of their clinical presentation (coma recovery scale-revised, CRS-R). Follow-up assessment (date II) was also based on CRS-R in a period of 8 to 24 month after date I. At date I, 63 channel EEG (Brain Products, Gilching, Germany) was recorded outside the scanner, and subsequently simultaneous FDG-PET/fMRI was acquired on an integrated Siemens Biograph mMR 3T scanner (Siemens Healthineers, Erlangen Germany). Power spectral densities, permutation entropy (PE) and symbolic transfer entropy (STE) were calculated in/between frontal, temporal, parietal and occipital EEG channels. PE and STE are based on symbolic time series analysis and were already introduced as robust markers separating wakefulness from unconsciousness in EEG during general anesthesia. While PE quantifies the regularity structure of the neighboring order of signal values (a surrogate of cortical information processing), STE reflects information transfer between two signals (a surrogate of directed connectivity in cortical networks). fMRI was carried out using SPM12 (Wellcome Trust Center for Neuroimaging, University of London, UK). Functional images were realigned, segmented, normalized and smoothed. PET was acquired for 45 minutes in list-mode. For absolute quantification of brain’s glucose consumption rate in FDG-PET, kinetic modelling was performed with Patlak’s plot method. BOLD signal intensity in fMRI and glucose uptake in PET was calculated in 8 distinct cortical areas. PCA was performed over all markers from EEG/fMRI/PET. Prognosis (persistent VS and deceased patients vs. recovery to MCS/awake from date I to date II) was evaluated using the area under the curve (AUC) including bootstrap confidence intervals (CI, *: p<0.05). Results: Prognosis was reliably indicated by the first component of PCA (AUC=0.99*, CI=0.92-1.00) showing a higher AUC when compared to the best single markers (EEG: AUC<0.96*, fMRI: AUC<0.86*, PET: AUC<0.60). CRS-R did not show prediction (AUC=0.51, CI=0.29-0.78). Conclusion: In a multimodal analysis of EEG/fMRI/PET in coma patients, PCA lead to a reliable prognosis. The impact of this result is evident, as clinical estimates of prognosis are inapt at time and could be supported by quantitative biomarkers from EEG, fMRI and PET. Due to the small sample size, further investigations are required, in particular allowing superwised learning instead of the basic approach of unsuperwised PCA.

Keywords: coma states and prognosis, electroencephalogram, entropy, functional magnetic resonance imaging, machine learning, positron emission tomography, principal component analysis

Procedia PDF Downloads 319

Authors: B. Neethu, Diptesh Das

Abstract:

The present study investigates the performance of a semi-active controller using magneto-rheological dampers (MR) for seismic response reduction of a multi-span bridge. The application of structural control to the structures during earthquake excitation involves numerous challenges such as proper formulation and selection of the control strategy, mathematical modeling of the system, uncertainty in system parameters and noisy measurements. These problems, however, need to be tackled in order to design and develop controllers which will efficiently perform in such complex systems. A control algorithm, which can accommodate un-certainty and imprecision compared to all the other algorithms mentioned so far, due to its inherent robustness and ability to cope with the parameter uncertainties and imprecisions, is the sliding mode algorithm. A sliding mode control algorithm is adopted in the present study due to its inherent stability and distinguished robustness to system parameter variation and external disturbances. In general a semi-active control scheme using an MR damper requires two nested controllers: (i) an overall system controller, which derives the control force required to be applied to the structure and (ii) an MR damper voltage controller which determines the voltage required to be supplied to the damper in order to generate the desired control force. In the present study a sliding mode algorithm is used to determine the desired optimal force. The function of the voltage controller is to command the damper to produce the desired force. The clipped optimal algorithm is used to find the command voltage supplied to the MR damper which is regulated by a semi active control law based on sliding mode algorithm. The main objective of the study is to propose a robust semi active control which can effectively control the responses of the bridge under real earthquake ground motions. Lumped mass model of the bridge is developed and time history analysis is carried out by solving the governing equations of motion in the state space form. The effectiveness of MR dampers is studied by analytical simulations by subjecting the bridge to real earthquake records. In this regard, it may also be noted that the performance of controllers depends, to a great extent, on the characteristics of the input ground motions. Therefore, in order to study the robustness of the controller in the present study, the performance of the controllers have been investigated for fourteen different earthquake ground motion records. The earthquakes are chosen in such a way that all possible characteristic variations can be accommodated. Out of these fourteen earthquakes, seven are near-field and seven are far-field. Also, these earthquakes are divided into different frequency contents, viz, low-frequency, medium-frequency, and high-frequency earthquakes. The responses of the controlled bridge are compared with the responses of the corresponding uncontrolled bridge (i.e., the bridge without any control devices). The results of the numerical study show that the sliding mode based semi-active control strategy can substantially reduce the seismic responses of the bridge showing a stable and robust performance for all the earthquakes.

Keywords: bridge, semi active control, sliding mode control, MR damper

Procedia PDF Downloads 114

Authors: Md Tanvir Rahman, Mahbube Subhani, Mahmud Ashraf, Paul Kremer

Abstract:

Cross-laminated timber (CLT) is made from layers of timber boards orthogonally oriented in the thickness direction, and due to this, CLT can withstand bi-axial bending in contrast with most other engineered wood products such as laminated veneer lumber (LVL) and glued laminated timber (GLT). Wood is cylindrically anisotropic in nature and is characterized by significantly lower elastic modulus and shear modulus in the planes perpendicular to the fibre direction, and is therefore classified as orthotropic material and is thus characterized by 9 elastic constants which are three elastic modulus in longitudinal direction, tangential direction and radial direction, three shear modulus in longitudinal tangential plane, longitudinal radial plane and radial tangential plane and three Poisson’s ratio. For simplification, timber materials are generally assumed to be transversely isotropic, reducing the number of elastic properties characterizing it to 5, where the longitudinal plane and radial planes are assumed to be planes of symmetry. The validity of this assumption was investigated through numerical modelling of CLT with both orthotropic mechanical properties and transversely isotropic material properties for three softwood species, which are Norway spruce, Douglas fir, Radiata pine, and three hardwood species, namely Victorian ash, Beech wood, and Aspen subjected to uniformly distributed loading under simply supported boundary condition. It was concluded that assuming the timber to be transversely isotropic results in a negligible error in the order of 1 percent. It was also observed that along with longitudinal elastic modulus, ratio of longitudinal shear modulus (GL) and rolling shear modulus (GR) has a significant effect on a deflection for CLT panels of lower span to depth ratio. For softwoods such as Norway spruce and Radiata pine, the ratio of longitudinal shear modulus, GL to rolling shear modulus GR is reported to be in the order of 12 to 15 times in literature. This results in shear flexibility in transverse layers leading to increased deflection under out-of-plane loading. The rolling shear modulus of hardwoods has been found to be significantly higher than those of softwoods, where the ratio between longitudinal shear modulus to rolling shear modulus as low as 4. This has resulted in a significant rise in research into the manufacturing of CLT from entirely from hardwood, as well as from a combination of softwood and hardwoods. The commonly used beam theory to analyze the performance of CLT panels under out-of-plane loads are the Shear analogy method, Gamma method, and k-method. The shear analogy method has been found to be the most effective method where shear deformation is significant. The effect of the ratio of longitudinal shear modulus and rolling shear modulus of cross-layer on the deflection of CLT under uniformly distributed load with respect to its length to depth ratio was investigated using shear analogy method. It was observed that shear deflection is reduced significantly as the ratio of the shear modulus of the longitudinal layer and rolling shear modulus of cross-layer decreases. This indicates that there is significant room for improvement of the bending performance of CLT through developing hybrid CLT from a mix of softwood and hardwood.

Keywords: rolling shear modulus, shear deflection, ratio of shear modulus and rolling shear modulus, timber

Procedia PDF Downloads 103

Authors: Noura Al Akkari, Aurélie Foucquier, Sylvain Lespinats

Abstract:

Load monitoring is a management process for energy consumption towards energy savings and energy efficiency. Non Intrusive Load Monitoring (NILM) is one method of load monitoring used for disaggregation purposes. NILM is a technique for identifying individual appliances based on the analysis of the whole residence data retrieved from the main power meter of the house. Our NILM framework starts with data acquisition, followed by data preprocessing, then event detection, feature extraction, then general appliance modeling and identification at the final stage. The event detection stage is a core component of NILM process since event detection techniques lead to the extraction of appliance features. Appliance features are required for the accurate identification of the household devices. In this research work, we aim at developing a new event detection methodology with accurate load disaggregation to extract appliance features. Time-domain features extracted are used for tuning general appliance models for appliance identification and classification steps. We use unsupervised algorithms such as Dynamic Time Warping (DTW). The proposed method relies on detecting areas of operation of each residential appliance based on the power demand. Then, detecting the time at which each selected appliance changes its states. In order to fit with practical existing smart meters capabilities, we work on low sampling data with a frequency of (1/60) Hz. The data is simulated on Load Profile Generator software (LPG), which was not previously taken into consideration for NILM purposes in the literature. LPG is a numerical software that uses behaviour simulation of people inside the house to generate residential energy consumption data. The proposed event detection method targets low consumption loads that are difficult to detect. Also, it facilitates the extraction of specific features used for general appliance modeling. In addition to this, the identification process includes unsupervised techniques such as DTW. To our best knowledge, there exist few unsupervised techniques employed with low sampling data in comparison to the many supervised techniques used for such cases. We extract a power interval at which falls the operation of the selected appliance along with a time vector for the values delimiting the state transitions of the appliance. After this, appliance signatures are formed from extracted power, geometrical and statistical features. Afterwards, those formed signatures are used to tune general model types for appliances identification using unsupervised algorithms. This method is evaluated using both simulated data on LPG and real-time Reference Energy Disaggregation Dataset (REDD). For that, we compute performance metrics using confusion matrix based metrics, considering accuracy, precision, recall and error-rate. The performance analysis of our methodology is then compared with other detection techniques previously used in the literature review, such as detection techniques based on statistical variations and abrupt changes (Variance Sliding Window and Cumulative Sum).

Keywords: electrical disaggregation, DTW, general appliance modeling, event detection

Procedia PDF Downloads 57

Authors: Rob Leslie, Taher Karimian

Abstract:

The blockage of cross drainage structures is a risk that needs to be understood and managed or lessened through the design. A blockage is a random event, influenced by site-specific factors, which needs to be quantified for design. Under and overestimation of blockage can have major impacts on flood risk and cost associated with drainage structures. The importance of this matter is heightened for those projects located within sensitive lands. It is a particularly complex problem for large linear infrastructure projects (e.g., rail corridors) located within floodplains where blockage factors can influence flooding upstream and downstream of the infrastructure. The selection of the appropriate blockage factors for hydraulic modeling has been subject to extensive research by hydraulic engineers. This paper has been prepared to review the current Australian Rainfall and Runoff 2019 (ARR 2019) methodology for blockage assessment by applying this method to a transport corridor brownfield upgrade case study in New South Wales. The results of applying the method are also validated against asset data and maintenance records. ARR 2019 – Book 6, Chapter 6 includes advice and an approach for estimating the blockage of bridges and culverts. This paper concentrates specifically on the blockage of cross drainage structures. The method has been developed to estimate the blockage level for culverts affected by sediment or debris due to flooding. The objective of the approach is to evaluate a numerical blockage factor that can be utilized in a hydraulic assessment of cross drainage structures. The project included an assessment of over 200 cross drainage structures. In order to estimate a blockage factor for use in the hydraulic model, a process has been advanced that considers the qualitative factors (e.g., Debris type, debris availability) and site-specific hydraulic factors that influence blockage. A site rating associated with the debris potential (i.e., availability, transportability, mobility) at each crossing was completed using the method outlined in ARR 2019 guidelines. The hydraulic results inputs (i.e., flow velocity, flow depth) and qualitative factors at each crossing were developed into an advanced spreadsheet where the design blockage level for cross drainage structures were determined based on the condition relating Inlet Clear Width and L10 (average length of the longest 10% of the debris reaching the site) and the Adjusted Debris Potential. Asset data, including site photos and maintenance records, were then reviewed and compared with the blockage assessment to check the validity of the results. The results of this assessment demonstrate that the estimated blockage factors at each crossing location using ARR 2019 guidelines are well-validated with the asset data. The primary finding of the study is that the ARR 2019 methodology is a suitable approach for culvert blockage assessment that has been validated against a case study spanning a large geographical area and multiple sub-catchments. The study also found that the methodology can be effectively coded within a spreadsheet or similar analytical tool to automate its application.

Keywords: ARR 2019, blockage, culverts, methodology

Procedia PDF Downloads 317

Authors: R. R. Galiev, I. I. Lykov, A. E. Shitikov, I. A. Bilenko

Abstract:

Dual-comb interferometry is based on the mixing of two optical frequency combs with slightly different lines spacing which results in the mapping of the optical spectrum into the radio-frequency domain for future digitizing and numerical processing. The dual-comb approach enables diverse applications, including metrology, fast high-precision spectroscopy, and distance range. Ordinary frequency-modulated continuous-wave (FMCW) laser-based Light Identification Detection and Ranging systems (LIDARs) suffer from two main disadvantages: slow and unreliable mechanical, spatial scan and a rather wide linewidth of conventional lasers, which limits speed measurement resolution. Dual-comb distance measurements with Allan deviations down to 12 nanometers at averaging times of 13 microseconds, along with ultrafast ranging at acquisition rates of 100 megahertz, allowing for an in-flight sampling of gun projectiles moving at 150 meters per second, was previously demonstrated. Nevertheless, pump lasers with EDFA amplifiers made the device bulky and expensive. An alternative approach is a direct coupling of the laser to a reference microring cavity. Backscattering can tune the laser to the eigenfrequency of the cavity via the so-called self-injection locked (SIL) effect. Moreover, the nonlinearity of the cavity allows a solitonic frequency comb generation in the very same cavity. In this work, we developed a fully integrated, power-efficient, electrically driven dual-micro comb source based on the semiconductor lasers SIL to high-quality integrated Si3N4 microresonators. We managed to obtain robust 1400-1700 nm combs generation with a 150 GHz or 1 THz lines spacing and measure less than a 1 kHz Lorentzian withs of stable, MHz spaced beat notes in a GHz band using two separated chips, each pumped by its own, self-injection locked laser. A deep investigation of the SIL dynamic allows us to find out the turn-key operation regime even for affordable Fabry-Perot multifrequency lasers used as a pump. It is important that such lasers are usually more powerful than DFB ones, which were also tested in our experiments. In order to test the advantages of the proposed techniques, we experimentally measured a minimum detectable speed of a reflective object. It has been shown that the narrow line of the laser locked to the microresonator provides markedly better velocity accuracy, showing velocity resolution down to 16 nm/s, while the no-SIL diode laser only allowed 160 nm/s with good accuracy. The results obtained are in agreement with the estimations and open up ways to develop LIDARs based on compact and cheap lasers. Our implementation uses affordable components, including semiconductor laser diodes and commercially available silicon nitride photonic circuits with microresonators.

Keywords: dual-comb spectroscopy, LIDAR, optical microresonator, self-injection locking

Procedia PDF Downloads 53

Authors: Noura Al Akkari, Aurélie Foucquier, Sylvain Lespinats

Abstract:

Load monitoring is a management process for energy consumption towards energy savings and energy efficiency. Non Intrusive Load Monitoring (NILM) is one method of load monitoring used for disaggregation purposes. NILM is a technique for identifying individual appliances based on the analysis of the whole residence data retrieved from the main power meter of the house. Our NILM framework starts with data acquisition, followed by data preprocessing, then event detection, feature extraction, then general appliance modeling and identification at the final stage. The event detection stage is a core component of NILM process since event detection techniques lead to the extraction of appliance features. Appliance features are required for the accurate identification of the household devices. In this research work, we aim at developing a new event detection methodology with accurate load disaggregation to extract appliance features. Time-domain features extracted are used for tuning general appliance models for appliance identification and classification steps. We use unsupervised algorithms such as Dynamic Time Warping (DTW). The proposed method relies on detecting areas of operation of each residential appliance based on the power demand. Then, detecting the time at which each selected appliance changes its states. In order to fit with practical existing smart meters capabilities, we work on low sampling data with a frequency of (1/60) Hz. The data is simulated on Load Profile Generator software (LPG), which was not previously taken into consideration for NILM purposes in the literature. LPG is a numerical software that uses behaviour simulation of people inside the house to generate residential energy consumption data. The proposed event detection method targets low consumption loads that are difficult to detect. Also, it facilitates the extraction of specific features used for general appliance modeling. In addition to this, the identification process includes unsupervised techniques such as DTW. To our best knowledge, there exist few unsupervised techniques employed with low sampling data in comparison to the many supervised techniques used for such cases. We extract a power interval at which falls the operation of the selected appliance along with a time vector for the values delimiting the state transitions of the appliance. After this, appliance signatures are formed from extracted power, geometrical and statistical features. Afterwards, those formed signatures are used to tune general model types for appliances identification using unsupervised algorithms. This method is evaluated using both simulated data on LPG and real-time Reference Energy Disaggregation Dataset (REDD). For that, we compute performance metrics using confusion matrix based metrics, considering accuracy, precision, recall and error-rate. The performance analysis of our methodology is then compared with other detection techniques previously used in the literature review, such as detection techniques based on statistical variations and abrupt changes (Variance Sliding Window and Cumulative Sum).

Keywords: general appliance model, non intrusive load monitoring, events detection, unsupervised techniques;

Procedia PDF Downloads 58

Authors: Aleksandr S. Demin, Anastasiia V. Ivanova

Abstract:

Purpose: The current economy tends to increase the influence of digital technologies and diminish the human role in management. However, it is impossible to deny that a person still leads a business with its own set of values and priorities. The article presented aims to incorporate the peculiarities of the national culture and the characteristics of the supply chain using the quantitative values of the national culture obtained by the scholars of comparative management (Hofstede, House, and others). Design/Methodology/Approach: The conducted research is based on the secondary data in the field of cross-country comparison achieved by Prof. Hofstede and received in the GLOBE project. The data mentioned are used to design different aspects of the supply chain both on the cross-functional and inter-organizational levels. The connection between a range of principles in general (roles assignment, customer service prioritization, coordination of supply chain partners) and in comparative management (acknowledgment of the national peculiarities of the country in which the company operates) is shown over economic and mathematical models, mainly linear programming models. Findings: The combination of the team management wheel concept, the business processes of the global supply chain, and the national culture characteristics let a transnational corporation to form a supply chain crew balanced in costs, functions, and personality. To elaborate on an effective customer service policy and logistics strategy in goods and services distribution in the country under review, two approaches are offered. The first approach relies exceptionally on the customer’s interest in the place of operation, while the second one takes into account the position of the transnational corporation and its previous experience in order to accord both organizational and national cultures. The effect of integration practice on the achievement of a specific supply chain goal in a specific location is advised to assess via types of correlation (positive, negative, non) and the value of national culture indices. Research Limitations: The models developed are intended to be used by transnational companies and business forms located in several nationally different areas. Some of the inputs to illustrate the application of the methods offered are simulated. That is why the numerical measurements should be used with caution. Practical Implications: The research can be of great interest for the supply chain managers who are responsible for the engineering of global supply chains in a transnational corporation and the further activities in doing business on the international area. As well, the methods, tools, and approaches suggested can be used by top managers searching for new ways of competitiveness and can be suitable for all staff members who are keen on the national culture traits topic. Originality/Value: The elaborated methods of decision-making with regard to the national environment suggest the mathematical and economic base to find a comprehensive solution.

Keywords: logistics integration, logistics services, multinational corporation, national culture, team management, service policy, supply chain management

Procedia PDF Downloads 88

Authors: Jiahui You, Kyung Jae Lee

Abstract:

Hydraulic fracturing fluid (HFF) is widely used in shale reservoir productions. HFF contains diverse chemical additives, which result in the dissolution and precipitation of minerals through multiple chemical reactions. In this study, a new pore-scale Darcy–Brinkman–Stokes (DBS) model coupled with Level Set Method (LSM) is developed to address the microscopic phenomena occurring during the iron–HFF interaction, by numerically describing mass transport, chemical reactions, and pore structure evolution. The new model is developed based on OpenFOAM, which is an open-source platform for computational fluid dynamics. Here, the DBS momentum equation is used to solve for velocity by accounting for the fluid-solid mass transfer; an advection-diffusion equation is used to compute the distribution of injected HFF and iron. The reaction–induced pore evolution is captured by applying the LSM, where the solid-liquid interface is updated by solving the level set distance function and reinitialized to a signed distance function. Then, a smoothened Heaviside function gives a smoothed solid-liquid interface over a narrow band with a fixed thickness. The stated equations are discretized by the finite volume method, while the re-initialized equation is discretized by the central difference method. Gauss linear upwind scheme is used to solve the level set distance function, and the Pressure–Implicit with Splitting of Operators (PISO) method is used to solve the momentum equation. The numerical result is compared with 1–D analytical solution of fluid-solid interface for reaction-diffusion problems. Sensitivity analysis is conducted with various Damkohler number (DaII) and Peclet number (Pe). We categorize the Fe (III) precipitation into three patterns as a function of DaII and Pe: symmetrical smoothed growth, unsymmetrical growth, and dendritic growth. Pe and DaII significantly affect the location of precipitation, which is critical in determining the injection parameters of hydraulic fracturing. When DaII<1, the precipitation uniformly occurs on the solid surface both in upstream and downstream directions. When DaII>1, the precipitation mainly occurs on the solid surface in an upstream direction. When Pe>1, Fe (II) transported deeply into and precipitated inside the pores. When Pe<1, the precipitation of Fe (III) occurs mainly on the solid surface in an upstream direction, and they are easily precipitated inside the small pore structures. The porosity–permeability relationship is subsequently presented. This pore-scale model allows high confidence in the description of Fe (II) dissolution, transport, and Fe (III) precipitation. The model shows fast convergence and requires a low computational load. The results can provide reliable guidance for injecting HFF in shale reservoirs to avoid clogging and wellbore pollution. Understanding Fe (III) precipitation, and Fe (II) release and transport behaviors give rise to a highly efficient hydraulic fracture project.

Keywords: reactive-transport , Shale, Kerogen, precipitation

Procedia PDF Downloads 150

Authors: Pawel Martynowicz

Abstract:

Vibrations are a crucial problem for slender structures such as towers, masts, chimneys, wind turbines, bridges, high buildings, etc., that is why most of them are equipped with vibration attenuation or fatigue reduction solutions. In this work, a slender structure (i.e., wind turbine tower-nacelle model) equipped with nonlinear, semiactive tuned vibration absorber(s) is analyzed. For this study purposes, magnetorheological (MR) dampers are used as semiactive actuators. Several optimal-based approaches to structural vibration attenuation are investigated against the standard ‘ground-hook’ law and passive tuned vibration absorber(s) implementations. The common approach to optimal control of nonlinear systems is offline computation of the optimal solution, however, so determined open loop control suffers from lack of robustness to uncertainties (e.g., unmodelled dynamics, perturbations of external forces or initial conditions), and thus perturbation control techniques are often used. However, proper linearization may be an issue for highly nonlinear systems with implicit relations between state, co-state, and control. The main contribution of the author is the development as well as numerical and experimental verification of the Pontriagin maximum-principle-based vibration control concepts that produce directly actuator control input (not the demanded force), thus force tracking algorithm that results in control inaccuracy is entirely omitted. These concepts, including one-step optimal control, quasi-optimal control, and optimal-based modified ‘ground-hook’ law, can be directly implemented in online and real-time feedback control for periodic (or semi-periodic) disturbances with invariant or time-varying parameters, as well as for non-periodic, transient or random disturbances, what is a limitation for some other known solutions. No offline calculation, excitations/disturbances assumption or vibration frequency determination is necessary, moreover, all of the nonlinear actuator (MR damper) force constraints, i.e., no active forces, lower and upper saturation limits, hysteresis-type dynamics, etc., are embedded in the control technique, thus the solution is optimal or suboptimal for the assumed actuator, respecting its limitations. Depending on the selected method variant, a moderate or decisive reduction in the computational load is possible compared to other methods of nonlinear optimal control, while assuring the quality and robustness of the vibration reduction system, as well as considering multi-pronged operational aspects, such as possible minimization of the amplitude of the deflection and acceleration of the vibrating structure, its potential and/or kinetic energy, required actuator force, control input (e.g. electric current in the MR damper coil) and/or stroke amplitude. The developed solutions are characterized by high vibration reduction efficiency – the obtained maximum values of the dynamic amplification factor are close to 2.0, while for the best of the passive systems, these values exceed 3.5.

Keywords: magnetorheological damper, nonlinear tuned vibration absorber, optimal control, real-time structural vibration attenuation, wind turbines

Procedia PDF Downloads 107

Authors: Mohammad Alipour, Ekin Esen, Riza Kizilel

Abstract:

Lithium-ion batteries are a commonly used type of rechargeable batteries because of their high specific energy and specific power. With the growing popularity of electric vehicles and hybrid electric vehicles, increasing attentions have been paid to rechargeable Lithium-ion batteries. However, safety problems, high cost and poor performance in low ambient temperatures and high current rates, are big obstacles for commercial utilization of these batteries. By proper thermal management, most of the mentioned limitations could be eliminated. Temperature profile of the Li-ion cells has a significant role in the performance, safety, and cycle life of the battery. That is why little temperature gradient can lead to great loss in the performances of the battery packs. In recent years, numerous researchers are working on new techniques to imply a better thermal management on Li-ion batteries. Keeping the battery cells within an optimum range is the main objective of battery thermal management. Commercial Li-ion cells are composed of several electrochemical layers each consisting negative-current collector, negative electrode, separator, positive electrode, and positive current collector. However, many researchers have adopted a single-layer cell to save in computing time. Their hypothesis is that thermal conductivity of the layer elements is so high and heat transfer rate is so fast. Therefore, instead of several thin layers, they model the cell as one thick layer unit. In previous work, we showed that single-layer model is insufficient to simulate the thermal behavior and temperature nonuniformity of the high-capacity Li-ion cells. We also studied the effects of the number of layers on thermal behavior of the Li-ion batteries. In this work, first thermal and electrochemical behavior of the LiFePO₄ battery is modeled with 3D multilayer cell. The model is validated with the experimental measurements at different current rates and ambient temperatures. Real time heat generation rate is also studied at different discharge rates. Results showed non-uniform temperature distribution along the cell which requires thermal management system. Therefore, aluminum plates with mini-channel system were designed to control the temperature uniformity. Design parameters such as channel number and widths, inlet flow rate, and cooling fluids are optimized. As cooling fluids, water and air are compared. Pressure drop and velocity profiles inside the channels are illustrated. Both surface and internal temperature profiles of single cell and battery packs are investigated with and without cooling systems. Our results show that using optimized Mini-channel cooling plates effectively controls the temperature rise and uniformity of the single cells and battery packs. With increasing the inlet flow rate, cooling efficiency could be reached up to 60%.

Keywords: lithium ion battery, 3D multilayer model, mini-channel cooling plates, thermal management

Procedia PDF Downloads 143

Authors: Abdul Rehman, Bo Liu

Abstract:

Axial flow turbines are commonly designed with high loads that generate strong secondary flows and result in high secondary losses. These losses contribute to almost 30% to 50% of the total losses. Non-axisymmetric endwall profiling is one of the passive control technique to reduce the secondary flow loss. In this paper, the non-axisymmetric endwall profile construction and optimization for the stator endwalls are presented to improve the efficiency of a high pressure turbine. The commercial code NUMECA Fine/ Design3D coupled with Fine/Turbo was used for the numerical investigation, design of experiments and the optimization. All the flow simulations were conducted by using steady RANS and Spalart-Allmaras as a turbulence model. The non-axisymmetric endwalls of stator hub and shroud were created by using the perturbation law based on Bezier Curves. Each cut having multiple control points was supposed to be created along the virtual streamlines in the blade channel. For the design of experiments, each sample was arbitrarily generated based on values automatically chosen for the control points defined during parameterization. The Optimization was achieved by using two algorithms i.e. the stochastic algorithm and gradient-based algorithm. For the stochastic algorithm, a genetic algorithm based on the artificial neural network was used as an optimization method in order to achieve the global optimum. The evaluation of the successive design iterations was performed using artificial neural network prior to the flow solver. For the second case, the conjugate gradient algorithm with a three dimensional CFD flow solver was used to systematically vary a free-form parameterization of the endwall. This method is efficient and less time to consume as it requires derivative information of the objective function. The objective function was to maximize the isentropic efficiency of the turbine by keeping the mass flow rate as constant. The performance was quantified by using a multi-objective function. Other than these two classifications of the optimization methods, there were four optimizations cases i.e. the hub only, the shroud only, and the combination of hub and shroud. For the fourth case, the shroud endwall was optimized by using the optimized hub endwall geometry. The hub optimization resulted in an increase in the efficiency due to more homogenous inlet conditions for the rotor. The adverse pressure gradient was reduced but the total pressure loss in the vicinity of the hub was increased. The shroud optimization resulted in an increase in efficiency, total pressure loss and entropy were reduced. The combination of hub and shroud did not show overwhelming results which were achieved for the individual cases of the hub and the shroud. This may be caused by fact that there were too many control variables. The fourth case of optimization showed the best result because optimized hub was used as an initial geometry to optimize the shroud. The efficiency was increased more than the individual cases of optimization with a mass flow rate equal to the baseline design of the turbine. The results of artificial neural network and conjugate gradient method were compared.

Keywords: artificial neural network, axial turbine, conjugate gradient method, non-axisymmetric endwall, optimization

Procedia PDF Downloads 213

Authors: Oibar Martinez, Clara Oliver, Jose Miguel Miranda

Abstract:

In this paper, we discuss the standard improvements which can be made to reduce the earth resistance in difficult terrains for optimum lightning protection, what are the practical limitations, and how the modeling can be refined for accurate diagnostics and ground resistance minimization. Ground resistance minimization can be made via three different approaches: burying vertical electrodes connected in parallel, burying horizontal conductive plates or meshes, or modifying the own terrain, either by changing the entire terrain material in a large volume or by adding earth-enhancing compounds. The use of vertical electrodes connected in parallel pose several practical limitations. In order to prevent loss of effectiveness, it is necessary to keep a minimum distance between each electrode, which is typically around five times larger than the electrode length. Otherwise, the overlapping of the local equipotential lines around each electrode reduces the efficiency of the configuration. The addition of parallel electrodes reduces the resistance and facilitates the measurement, but the basic parallel resistor formula of circuit theory will always underestimate the final resistance. Numerical simulation of equipotential lines around the electrodes overcomes this limitation. The resistance of a single electrode will always be proportional to the soil resistivity. The electrodes are usually installed with a backfilling material of high conductivity, which increases the effective diameter. However, the improvement is marginal, since the electrode diameter counts in the estimation of the ground resistance via a logarithmic function. Substances that are used for efficient chemical treatment must be environmentally friendly and must feature stability, high hygroscopicity, low corrosivity, and high electrical conductivity. A number of earth enhancement materials are commercially available. Many are comprised of carbon-based materials or clays like bentonite. These materials can also be used as backfilling materials to reduce the resistance of an electrode. Chemical treatment of soil has environmental issues. Some products contain copper sulfate or other copper-based compounds, which may not be environmentally friendly. Carbon-based compounds are relatively inexpensive and they do have very low resistivities, but they also feature corrosion issues. Typically, the carbon can corrode and destroy a copper electrode in around five years. These compounds also have potential environmental concerns. Some earthing enhancement materials contain cement, which, after installation acquire properties that are very close to concrete. This prevents the earthing enhancement material from leaching into the soil. After analyzing different configurations, we conclude that a buried conductive ring with vertical electrodes connected periodically should be the optimum baseline solution for the grounding of a large size structure installed on a large resistivity terrain. In order to show this, a practical example is explained here where we simulate the ground resistance of a conductive ring buried in a terrain with a resistivity in the range of 1 kOhm·m.

Keywords: grounding improvements, large scale scientific instrument, lightning risk assessment, lightning standards

Procedia PDF Downloads 117

Authors: Galatee Levadoux, Trevor Benson, Chris Worrall

Abstract:

With the rise of public awareness on climate change comes an increasing demand for renewable sources of energy. As a result, the wind power sector is striving to manufacture longer, more efficient and reliable wind turbine blades. Currently, one of the leading causes of blade failure in service is improper cure of the resin during manufacture. The infusion process creating the main part of the composite blade structure remains a critical step that is yet to be monitored in real time. This stage consists of a viscous resin being drawn into a mould under vacuum, then undergoing a curing reaction until solidification. Successful infusion assumes the resin fills all the voids and cures completely. Given that the electrical properties of the resin change significantly during its solidification, both the filling of the mould and the curing reaction are susceptible to be followed using dieletrometry. However, industrially available dielectrics sensors are currently too small to monitor the entire surface of a wind turbine blade. The aim of the present research project is to scale up the dielectric sensor technology and develop a device able to monitor the manufacturing process of large composite structures, assessing the conformity of the blade before it even comes out of the mould. An array of flat copper wires acting as electrodes are embedded in a polymer matrix fixed in an infusion mould. A multi-frequency analysis from 1 Hz to 10 kHz is performed during the filling of the mould with an epoxy resin and the hardening of the said resin. By following the variations of the complex admittance Y*, the filling of the mould and curing process are monitored. Results are compared to numerical simulations of the sensor in order to validate a virtual cure-monitoring system. The results obtained by drawing glycerol on top of the copper sensor displayed a linear relation between the wetted length of the sensor and the complex admittance measured. Drawing epoxy resin on top of the sensor and letting it cure at room temperature for 24 hours has provided characteristic curves obtained when conventional interdigitated sensor are used to follow the same reaction. The response from the developed sensor has shown the different stages of the polymerization of the resin, validating the geometry of the prototype. The model created and analysed using COMSOL has shown that the dielectric cure process can be simulated, so long as a sufficient time and temperature dependent material properties can be determined. The model can be used to help design larger sensors suitable for use with full-sized blades. The preliminary results obtained with the sensor prototype indicate that the infusion and curing process of an epoxy resin can be followed with the chosen configuration on a scale of several decimeters. Further work is to be devoted to studying the influence of the sensor geometry and the infusion parameters on the results obtained. Ultimately, the aim is to develop a larger scale sensor able to monitor the flow and cure of large composite panels industrially.

Keywords: composite manufacture, dieletrometry, epoxy, resin infusion, wind turbine blades

Procedia PDF Downloads 145

Authors: Bensaid A., Mostephaoui T., Nedjai R.

Abstract:

A considerable area of Algerian lands are threatened by the phenomenon of wind erosion. For a long time, wind erosion and its associated harmful effects on the natural environment have posed a serious threat, especially in the arid regions of the country. In recent years, as a result of increases in the irrational exploitation of natural resources (fodder) and extensive land clearing, wind erosion has particularly accentuated. The extent of degradation in the arid region of the Algerian Mécheriadepartment generated a new situation characterized by the reduction of vegetation cover, the decrease of land productivity, as well as sand encroachment on urban development zones. In this study, we attempt to investigate the potential of remote sensing and geographic information systems for detecting the spatial dynamics of the ancient dune cords based on the numerical processing of PlanetScope PSB.SB sensors images by September 29, 2021. As a second step, we prospect the use of a multi-scale geographic object-based image analysis (GEOBIA) approach to segment the high spatial resolution images acquired on heterogeneous surfaces that vary according to human influence on the environment. We have used the fractal net evolution approach (FNEA) algorithm to segment images (Baatz&Schäpe, 2000). Multispectral data, a digital terrain model layer, ground truth data, a normalized difference vegetation index (NDVI) layer, and a first-order texture (entropy) layer were used to segment the multispectral images at three segmentation scales, with an emphasis on accurately delineating the boundaries and components of the sand accumulation areas (Dune, dunes fields, nebka, and barkhane). It is important to note that each auxiliary data contributed to improve the segmentation at different scales. The silted areas were classified using a nearest neighbor approach over the Naâma area using imagery. The classification of silted areas was successfully achieved over all study areas with an accuracy greater than 85%, although the results suggest that, overall, a higher degree of landscape heterogeneity may have a negative effect on segmentation and classification. Some areas suffered from the greatest over-segmentation and lowest mapping accuracy (Kappa: 0.79), which was partially attributed to confounding a greater proportion of mixed siltation classes from both sandy areas and bare ground patches. This research has demonstrated a technique based on very high-resolution images for mapping sanded and degraded areas using GEOBIA, which can be applied to the study of other lands in the steppe areas of the northern countries of the African continent.

Keywords: land development, GIS, sand dunes, segmentation, remote sensing

Procedia PDF Downloads 90

Authors: Eiji Nakamachi, Tsuyoshi Eguchi, Sayo Yamamoto, Yusuke Morita, H. Sakamoto

Abstract:

In this study, we aim to reveal the relationship between the meniscal tearing and the articular cartilage injury of knee joint by using the dynamic explicit finite element (FE) method. Meniscal injuries reduce its functional ability and consequently increase the load on the articular cartilage of knee joint. In order to prevent the induction of osteoarthritis (OA) caused by meniscal injuries, many medical treatment techniques, such as artificial meniscus replacement and meniscal regeneration, have been developed. However, it is reported that these treatments are not the comprehensive methods. In order to reveal the fundamental mechanism of OA induction, the mechanical characterization of meniscus under the condition of normal and injured states is carried out by using FE analyses. At first, a FE model of the human knee joint in the case of normal state – ‘intact’ - was constructed by using the magnetron resonance (MR) tomography images and the image construction code, Materialize Mimics. Next, two types of meniscal injury models with the radial tears of medial and lateral menisci were constructed. In FE analyses, the linear elastic constitutive law was adopted for the femur and tibia bones, the visco-hyperelastic constitutive law for the articular cartilage, and the visco-anisotropic hyperelastic constitutive law for the meniscus, respectively. Material properties of articular cartilage and meniscus were identified using the stress-strain curves obtained by our compressive and the tensile tests. The numerical results under the normal walking condition revealed how and where the maximum compressive stress occurred on the articular cartilage. The maximum compressive stress and its occurrence point were varied in the intact and two meniscal tear models. These compressive stress values can be used to establish the threshold value to cause the pathological change for the diagnosis. In this study, FE analyses of knee joint were carried out to reveal the influence of meniscal injuries on the cartilage injury. The following conclusions are obtained. 1. 3D FE model, which consists femur, tibia, articular cartilage and meniscus was constructed based on MR images of human knee joint. The image processing code, Materialize Mimics was used by using the tetrahedral FE elements. 2. Visco-anisotropic hyperelastic constitutive equation was formulated by adopting the generalized Kelvin model. The material properties of meniscus and articular cartilage were determined by curve fitting with experimental results. 3. Stresses on the articular cartilage and menisci were obtained in cases of the intact and two radial tears of medial and lateral menisci. Through comparison with the case of intact knee joint, two tear models show almost same stress value and higher value than the intact one. It was shown that both meniscal tears induce the stress localization in both medial and lateral regions. It is confirmed that our newly developed FE analysis code has a potential to be a new diagnostic system to evaluate the meniscal damage on the articular cartilage through the mechanical functional assessment.

Keywords: finite element analysis, hyperelastic constitutive law, knee joint injury, meniscal tear, stress concentration

Procedia PDF Downloads 223

Authors: Ja-Keoung Koo, Kensuke Nakamura, Hyohun Kim, Dongwha Shin, Yeonseok Kim, Ji-Su Ahn, Byung-Woo Hong

Abstract:

The machine learning techniques based on a convolutional neural network (CNN) have been actively developed and successfully applied to a variety of image analysis tasks including reconstruction, noise reduction, resolution enhancement, segmentation, motion estimation, object recognition. The classical visual information processing that ranges from low level tasks to high level ones has been widely developed in the deep learning framework. It is generally considered as a challenging problem to derive visual interpretation from high dimensional imagery data. A CNN is a class of feed-forward artificial neural network that usually consists of deep layers the connections of which are established by a series of non-linear operations. The CNN architecture is known to be shift invariant due to its shared weights and translation invariance characteristics. However, it is often computationally intractable to optimize the network in particular with a large number of convolution layers due to a large number of unknowns to be optimized with respect to the training set that is generally required to be large enough to effectively generalize the model under consideration. It is also necessary to limit the size of convolution kernels due to the computational expense despite of the recent development of effective parallel processing machinery, which leads to the use of the constantly small size of the convolution kernels throughout the deep CNN architecture. However, it is often desired to consider different scales in the analysis of visual features at different layers in the network. Thus, we propose a CNN model where different sizes of the convolution kernels are applied at each layer based on the random projection. We apply random filters with varying sizes and associate the filter responses with scalar weights that correspond to the standard deviation of the random filters. We are allowed to use large number of random filters with the cost of one scalar unknown for each filter. The computational cost in the back-propagation procedure does not increase with the larger size of the filters even though the additional computational cost is required in the computation of convolution in the feed-forward procedure. The use of random kernels with varying sizes allows to effectively analyze image features at multiple scales leading to a better generalization. The robustness and effectiveness of the proposed CNN based on random kernels are demonstrated by numerical experiments where the quantitative comparison of the well-known CNN architectures and our models that simply replace the convolution kernels with the random filters is performed. The experimental results indicate that our model achieves better performance with less number of unknown weights. The proposed algorithm has a high potential in the application of a variety of visual tasks based on the CNN framework. Acknowledgement—This work was supported by the MISP (Ministry of Science and ICT), Korea, under the National Program for Excellence in SW (20170001000011001) supervised by IITP, and NRF-2014R1A2A1A11051941, NRF2017R1A2B4006023.

Keywords: deep learning, convolutional neural network, random kernel, random projection, dimensionality reduction, object recognition

Procedia PDF Downloads 264

Authors: Zenaida Chitu, Roxana Bojariu, Liliana Velea, Roxana Burcea

Abstract:

A number of studies show that extreme air temperature affects mortality related to cardiovascular diseases, particularly among elderly people. In Romania, the summer thermal discomfort expressed by Universal Thermal Climate Index (UTCI) is highest in the Southern part of the country, where Bucharest, the largest Romanian urban agglomeration, is also located. The urban characteristics such as high building density and reduced green areas enhance the increase of the air temperature during summer. In Bucharest, as in many other large cities, the effect of heat urban island is present and determines an increase of air temperature compared to surrounding areas. This increase is particularly important during heat wave periods in summer. In this context, the researchers performed a temperature-mortality analysis based on daily deaths related to cardiovascular diseases, recorded between 2010 and 2019 in Bucharest. The temperature-mortality relationship was modeled by applying distributed lag non-linear model (DLNM) that includes a bi-dimensional cross-basis function and flexible natural cubic spline functions with three internal knots in the 10th, 75th and 90th percentiles of the temperature distribution, for modelling both exposure-response and lagged-response dimensions. Firstly, this study applied this analysis for the present climate. Extrapolation of the exposure-response associations beyond the observed data allowed us to estimate future effects on mortality due to temperature changes under climate change scenarios and specific assumptions. We used future projections of air temperature from five numerical experiments with regional climate models included in the EURO-CORDEX initiative under the relatively moderate (RCP 4.5) and pessimistic (RCP 8.5) concentration scenarios. The results of this analysis show for RCP 8.5 an ensemble-averaged increase with 6.1% of heat-attributable mortality fraction in future in comparison with present climate (2090-2100 vs. 2010-219), corresponding to an increase of 640 deaths/year, while mortality fraction due to the cold conditions will be reduced by 2.76%, corresponding to a decrease by 288 deaths/year. When mortality data is stratified according to the age, the ensemble-averaged increase of heat-attributable mortality fraction for elderly people (> 75 years) in the future is even higher (6.5 %). These findings reveal the necessity to carefully plan urban development in Bucharest to face the public health challenges raised by the climate change. Paper Details: This work is financed by the project URCLIM which is part of ERA4CS, an ERA-NET initiated by JPI Climate, and funded by Ministry of Environment, Romania with co-funding by the European Union (Grant 690462). A part of this work performed by one of the authors has received funding from the European Union’s Horizon 2020 research and innovation programme from the project EXHAUSTION under grant agreement No 820655.

Keywords: cardiovascular diseases, climate change, extreme air temperature, mortality

Procedia PDF Downloads 104