@misc{Shrestha, type = {Master Thesis}, author = {Shrestha, Manjil}, title = {Development of a Language Model for Medical Domain}, url = {http://nbn-resolving.de/urn:nbn:de:hbz:1383-opus4-7405}, school = {Hochschule Rhein-Waal}, pages = {141}, abstract = {Language models are widely used as a representation of written language in various machine learning tasks, with the most commonly used model being Bidirectional Encoder Representations from Transformers (BERT). It was shown that the prediction quality strongly benefits from language model pre-training on domain-specific data. The publicly available models, though are always trained on Wikipedia, news or legal data, thereby missing the domain specific knowledge about medical terms. In this thesis, we will train a BERT language model on medical data and compare performance with domain-unspecific language models. The dataset used for this purpose is the Non-technical Summaries - International Statistical Classification of Diseases (NTS-ICD) task of classification of animal experiment descriptions into International Statistical Classification of Diseases (ICD) categories.}, language = {en} } @misc{AjitkumarShah, type = {Master Thesis}, author = {Ajitkumar Shah, Karan}, title = {Development of Decision-making Model for Cost Optimization of Heat Exchanger on the Basis of Spreadsheet Programs}, url = {http://nbn-resolving.de/urn:nbn:de:hbz:1383-opus4-9114}, school = {Hochschule Rhein-Waal}, pages = {62}, abstract = {The current advancements in the field of heat exchangers have expanded their range of applications in many industries. The most important factor while investing in the heat exchanger is sizing them for a particular application. The sizing of a shell and tube heat exchanger is always a challenge, often causing customers to invest in either an oversized or undersized heat exchanger. The purpose of this study is to solve the customer's conundrum of investing in an appropriate size shell and tube heat exchanger. The objectives of this study were (a) to develop and implement a spreadsheet program for optimization of overall heat transfer coefficient of shell and tube heat exchanger, (b) to determine profitability analysis using the Net Present Value (NPV) method, (c) to analyse the price sensitivity using the Monte Carlo simulation. The iteration and optimization are based on Kern's method. The user defines the process parameters such as the temperature of hot and cold streams, mass flow rates and fluid densities. The user also assumes the tube related properties. The final sizing of the shell side is determined along with Reynolds number and pressure losses on the shell and tube side. The pressure losses are then converted into the operating cost of the heat exchanger to determine NPV. The Monte Carlo simulation calculates 1000 different NPVs for a given scenario, thus facilitating the user's crucial decision-making process. This study is probably a first instance to combine preliminary sizing of the shell and tube heat exchanger with price sensitivity analysis using the Monte Carlo method.}, language = {en} } @misc{O'Shea, type = {Master Thesis}, author = {O'Shea, Brendan}, title = {Evaluation of Solidworks on how to integrate part preparation for Additive Manufacturing in the CAD modeling process}, url = {http://nbn-resolving.de/urn:nbn:de:hbz:1383-opus4-13823}, school = {Hochschule Rhein-Waal}, pages = {71}, abstract = {Solidworks is a CAD software that facilitates the construction of 3D models and drawings to enable production. These systems are usually designed to be used in conjunction with Subtractive Manufacturing (SM) and forming processes. This is due to the enormous portion of world-wide production in which these two processes are used. Because of this, design guidelines are customarily based on the abilities and constraints of conventional manufacturing processes. Additive Manufacturing (AM) replaces these with a new set of advantages and limitations that should be considered to assist in the design of parts. Design for Additive Manufacturing (DfAM) represents a methodology which closely resembles Design for Assembly (DFA) and Design for Manufacturing (DFM), as far as reducing component count and allow for easier manufacturing. The aim of this research is to establish DfAM criteria and to evaluate the core Solidworks software in its ability to support design using these criteria. DfAM criteria is sub-divided into two categories. Product DfAM criteria are those derived from the advantages offered by AM technology, while Process DfAM are those determined by the limitations of the AM technique used. Three example models, each addressing multiple criteria, are created in Solidworks and optimized for printing. Through this process, helpful features in the software are identified and analyzed. Embodiment of these examples by means of Fused Deposition Modeling (FDM) validates these features in an exemplary way and explores the integration of AM design practices in Solidworks. The results of this evaluation will allow further research into CAD support for AM, specifically with respect to Product DfAM.}, language = {en} }