Linköping Electronic Conference Proceedings
Not a member yet
1113 research outputs found
Sort by
Machine learning techniques for modeling chemical absorption in CO2 capture process
Get PDFPost-combustion carbon capture (PCC) technologies play an important role in the reduction of CO2 emissions to address climate challenges. This process is usually simulated in process simulation software based on first-principle models, which calculate physical properties directly from basic physical quantities such as mass and temperature. Using first-principle models usually requires a long computation time, which makes optimization and control difficult. In this study, machine learning algorithms, such as eXtreme Gradient Boosting (XGBoost) and Support Vector Regression (SVR), are investigated as potential alternative modeling approaches. XGBoost is an ensemble algorithm that is based on the decision tree and optimized by gradient boosting. SVR fits the best line within a predefined or threshold error value. These two algorithms are used to build models to predict the CO2 capture rate (CR) and specific reboiler duty (SRD) in a monoethanolamine-based PCC process. By using the XGBoost, the verification result shows R2 (a statistical measure that represents the fitness of the model) in predicting CR is 91.7% and in predicting SRD is 80.8%, while by using SVR the R2 in predicting CR and SRD is 87.9% and 87.2% individually. In addition, XGBoost and SVR take 0.022 seconds and 0.317 seconds respectively to predict CR and SRD of 1318 cases, while the first-principal process simulation model needs 3.15 seconds to calculate 1 case. The data-driven models built using the XGBoost algorithm are employed for further optimization, which aims to find an operating point to have a higher CR and lower SRD. Particle swarm optimization (PSO), a stochastic optimization technique based on the movement and intelligence of swarms, is implemented for the optimization. The CR and SRD for optimal operating conditions are 72.2% and 4.3 MJ/kg each. The computations are faster with the data-driven models incorporated in the optimization technique. Thus, the application of machine learning techniques in carbon capture technologies is demonstrated successfully
Simulation of Flow in the Human Upper Airways Modeled as a Piping System Using the Hydraulic Diameter
Get PDFObstructive sleep apnea (OSA) is a medical condition characterized by repetitive obstructions in the human upper airways during sleep. Recent estimates from the United States show that the condition impacts 15% to 20% of the adult population. OSA treatment can be subdivided into surgical and non-surgical approaches. Non-surgical approaches such as continuous positive airway pressure (CPAP) devices have the highest success rates when used correctly. However, these approaches have low patient compliance due to the invasive nature of the devices during sleep, leaving surgery as a viable alternative for many. Predicting the outcome of OSA surgery is difficult due to the complex nature of both the airways and the surgeries themselves. CFD modeling of the airways is a helpful way to gain valuable insights into the flow structures and the impact of individual surgeries on the airways. However, CFD is not a viable approach for each patient-specific case due to its time-consuming nature. A pragmatic model has been created to predict the outcome of OSA surgery on a patient-specific basis to produce valid surgical estimates fast to be used by non-CFD engineers. The model transforms the human upper airways into a piping system by applying the hydraulic diameter equation on geometries created from CT scans. This paper aims to validate the use of the hydraulic diameter given by Dh = 4 · (A / Pe), where A is the cross-sectional area and Pe is the wetted perimeter, on the complex geometries of the nasal cavity and to provide a novel equation for the hydraulic diameter in the nasal cavity. The proposed hydraulic diameter equation is given by Dh = CDh · (A / Pe) where CDh is the hydraulic diameter coefficient. Airflow has been simulated through a simplified geometry using CFD to validate the hydraulic diameter and find an updated equation. Pragmatic model simulations using the hydraulic diameter have been compared to the results from CFD simulations to assess the pragmatic model’s accuracy. The results showed that the original hydraulic diameter did not give entirely accurate results and that the novel equation using CDh = 3.71 gave the pragmatic model better accuracy for the validation cases. Tuning the parameter CDh for flow in an OSA patient’s upper airways, the pragmatic model succeeded in quite accurately reproducing the area-averaged pressure in the patient’s upper airways
An improved delay and state observer for SISO LTI systems with known delay lower bound
Get PDFIt is known that the presence of delays hinders the performance achievable by a feedback control system, and it can even lead to closed-loop instability if not considered during the design. For this reason, predictors are often included in the loop, although they typically require the knowledge of the exact value of the delay, which in some applications is hard to obtain in practice. This paper presents a method to design an observer that simultaneously estimates the unknown state and the time-varying input delay of a plant based on an available model and the measurements coming from the sensors. In particular, the main contribution of this paper is to show that by accounting for a known lower bound of the input delay, it is possible to improve the observer’s performance when compared to state-of-the-art approaches encountered in the literature. Simulations are used to illustrate the efficiency of the proposed design method
Eulerian-Lagrangian simulation of air-steam biomass gasification in a bubbling fluidized bed gasifier
Get PDFTo numerically study biomass gasification in a three-dimensional bubbling fluidized bed, a CFD-DEM (computational fluid dynamics – discrete element method) model with heat transfer and homogeneous and heterogeneous chemical reactions is implemented. An ideal reactor model is used for the air-steam bubbling fluidized bed (BFB) gasification reactor assuming perfectly mixed solids and plug flow. A validated computational particle fluid dynamics (CPFD) model has been applied to investigate the sensitivity analysis of mesh grids as well as to find the optimum number of grids. The result shows that 7452 grid cells are the optimal number of cells for the existence BFB gasifier. The effects of key process operating parameters such as steam to biomass ratio (SB), as well as temperature shows that by enhancing the SB ratio or reactor temperature, gas yields increase. H2 and CO2 concentrations promote by increasing the steam to biomass ratio while CO and CH4 production drop. The optimal value of SB for the gasification process can be found in the range of 0.3 to 1
Citizens’ use of Health Information Technology between 2013-2021 in Denmark: A longitudinal study
Get PDFAn increasing number of citizens with multiple chronic conditions and technological innovations enabling new types of treatments pressure the Danish healthcare sector economically. The solution so far has been increased patient responsibility and the application of digital healthcare solutions. This longitudinal study examines how Danish citizens between 2013-2021 interact with Health Information Technology (HIT) and digital data. Results show that the Danes' use of HIT and digital data has increased over the period. Additionally, the numbers reveal that education, gender, age and chronic conditions influence how HIT and digital health data are used, which is relevant from a health inequity perspective
Data collection and smart nudging to promote physical activity and a healthy lifestyle using wearable devices
Get PDFNudge principles and techniques can motivate and improve personal health through emerging digital devices, such as activity trackers. Tracking people's health and well-being using such devices have earned widespread interest. These devices can continuously capture and analyze health-related data from individuals and communities in their everyday environment. Providing context-aware nudges can help individuals to self-manage and improve their health. In this study, we discuss how a consumer-based activity tracker can be used to track different variables for physical activity (PA) and how it has the potential to be an important source of data for future smart nudging
Exploring Digital Psychosocial Follow-up for Survivors of Childhood Critical Illness
Get PDFThis extended abstract describes the plan and status of a PhD project using a design science research approach to explore how to design digital psychosocial follow-up for survivors of childhood critical illness
Score and Venue Adjustment on Transition Data in Hockey
Get PDFAre zone exits and entries influenced by score and venue the same way shots and goals are? Using our proprietary database of over 120,000 transition events, we analyzed how score and venue can impact how much you control your transitions and your success percentage. Playing at home or on the road does not seem to have much impact overall, especially compared to the influence the score of the game has. Trailing teams appear to be able to make more controlled zone exits, with greater success, probably due to a lesser pressure. On the other hand, leading teams tend to dump the puck out of their defensive zone more often. A trailing team would also try more zone entries but the split between controlled and dump attempts surprisingly remains stable, contradicting a common idea that defenses make it harder to enter the offensive zone when protecting a lead. The “play a simple game on the road” mantra, with less controlled transitions, does not seem to hold either, when looking at the data
Evaluating deep tracking models for player tracking in broadcast ice hockey video
Get PDFTracking and identifying players is an important problem in computer vision based ice hockey analytics. Player tracking is a challenging problem since the motion of players in hockey is fast-paced and non-linear. There is also significant player-player and player-board occlusion, camera panning and zooming in hockey broadcast video. Prior published research perform player tracking with the help of handcrafted features for player detection and re-identification. Although commercial solutions for hockey player tracking exist, to the best of our knowledge, no network architectures used, training data or performance metrics are publicly reported. There is currently no published work for hockey player tracking making use of the recent advancements in deep learning while also reporting the current accuracy metrics used in literature. Therefore, in this paper we compare and contrast several state-of-the-art tracking algorithms and analyze their performance and failure modes in ice hockey
The Nature of Icelandic as a Second Language: An Insight from the Learner Error Corpus for Icelandic
Get PDFThe Icelandic L2 Error Corpus is an expanding collection of texts written by users of Icelandic as a second language, published on CLARIN. It currently consisting of 22,705 manually-annotated errors in different categories pertaining to grammar, spelling, lexical and other issues. The corpus was used to perform a contrastive interlanguage analysis, first using a native speaker reference corpus – the Icelandic Error Corpus, then analysing the corpus internally based on linguistic features relevant to second language acquisition. This paper presents the corpus and first results of the analysis