Listing 1 - 10 of 165 | << page >> |
Sort by
|
Choose an application
Choose an application
Choose an application
Computer logic --- Rule-based programming --- Rule-based methods (Computer science) --- Computer programming
Choose an application
Doelstelling: Deze scriptie heeft als doel een vergelijking te maken van twee benaderingen van automatische vertaling. Als voorbeeld van regelgebaseerde automatische vertaling kozen we Systran en Google Translate als voorbeeld van statistische automatische vertaling. Meer specifiek vergeleken we de prestaties van deze systemen voor het Engels en het Russisch. Middelen of methode: Deze twee vertaalsystemen werden getest aan de hand van een specifieke testsuite waar zeven moeilijke punten centraal staan. Door korte zinnetjes met deze moeilijke punten te laten vertalen door beide systemen, zowel voor Engels-Russisch als Russisch-Engels konden we achterhalen welke systemen beter overweg kunnen met welke moeilijkheden. Daarna lieten we ook nog een algemene journalistieke tekst vertalen om te zien hoe deze moeilijkheden aangepakt worden. Resultaten: Het is niet mogelijk om één van beide systeem als het 'beste systeem' uit te roepen. Beide maken veel fouten.Het is wel zo dat het ene systeem minder fouten maakt tegen bepaalde moeilijkheden dan het ander systeem en omgekeerd. Afhankelijk van de categorie presteert Google Translate beter dan Systran, maar in anderecategorieën maakt Systran dan toch minder fouten.
English. --- Google Translate. --- Machine translation. --- Rule-based machine translation. --- Russian. --- Statistical machine translation. --- Systran.
Choose an application
This thesis aims at developing, optimizing and controling a hybrid solar microgrid for rural electrification and heat supply in sub-Saharan Africa. The considered system includes PV, Parabolic Trough Collectors, Organic Rankine Cycle and LPG generator, as well as chemical battery storage and thermal energy storage. The work focuses on multiple aspects of the ongoing development of solar hybrid microgrids for the rural electrification of remote areas in Lesotho. These aspects range from very specific improvements (the mechanical design of a high expansion ratio expander) to the more global evaluation of their impact once included into a complex micro-grid system. Special attention has also been paid to the links between thermal and electrical demands. The main contributions of this thesis are: - The mechanical design of a high expansion ratio scroll expander, involving drawing, machining and assembly of the parts. - The detailed model of an organic Rankine cycle with the purpose of evaluating the improvement brought by the high expansion ratio scroll expander and mapping the ORC performance. - A building model developed to predict the thermal loads of a health clinic in rural communities of Lesotho. The developed lumped-parameter model can be used for various building typologies and communities. The model is designed to be as generic and simple as possible, and contrasts with the data-intensive models generally proposed in the literature. - The gathering of monitoring and weather data relative to a health clinic in Lesotho, and their use for the calibration of the building model. - A microgrid model built by interconnecting all of its subcomponent models. A rule-based control strategy is developed, accounting for interactions between thermal and electrical loads, and dispatching heat and power flows of each component in order to cover the demand while minimizing the fuel consumption. - A particle-swarm optimization model used to optimize the microgrid under different cost assumptions and control strategies The above models prove that the system performs better with the developed high expansion ratio expander. The maximum output power of the ORC is increased by 33 %, and the fuel consumption of the microgrid is reduced by 25 %. For the studied community of Ha Nkau in Lesotho, the determined optimal system infrastructure is composed of PV (65 kW) and batteries (259 kWh) only, and the optimum control strategy achieves a levelized cost of electricity of 0.202 USD/kWh. Fuel consumption is mainly due to the burner, which supplies all the thermal load because no other heating system is selected by the optimization.
modeling --- optimization --- microgrid --- rural electrification --- ORC --- scroll expander --- heat supply --- control --- rule based control --- Lesotho --- Ingénierie, informatique & technologie > Energie
Choose an application
Expert systems (Computer science) --- Expertsystemen (Informatica) --- Rule-based programming --- Systèmes experts (Informatique) --- Rule-based programming. --- 681.3*I20 --- Rule-based methods (Computer science) --- Computer programming --- Knowledge-based systems (Computer science) --- Systems, Expert (Computer science) --- Artificial intelligence --- Computer systems --- Soft computing --- Artificial intelligence (AI) in general; cognitive simulation; philosophical foundations --- 681.3*I20 Artificial intelligence (AI) in general; cognitive simulation; philosophical foundations
Choose an application
681.3*H20 --- 681.3*I24 --- 681.3*I24 Knowledge representation formalisms and methods: frames and scripts predicate logic relation systems representation languages procedural and rule-based representations semantic networks (Artificial intelligence) --- Knowledge representation formalisms and methods: frames and scripts predicate logic relation systems representation languages procedural and rule-based representations semantic networks (Artificial intelligence) --- Computerwetenschap--?*H20 --- 681.3*I24 Knowledge representation formalisms and methods: frames and scripts; predicate logic; relation systems; representation languages; procedural and rule-based representations; semantic networks (Artificial intelligence) --- Knowledge representation formalisms and methods: frames and scripts; predicate logic; relation systems; representation languages; procedural and rule-based representations; semantic networks (Artificial intelligence) --- Artificial intelligence. Robotics. Simulation. Graphics
Choose an application
Doelstelling: Deze masterproef beoordeelt de vertaalkwaliteit van de online vertaalsystemen Google Translate (GT) en Babel Fish (BF) van het Engels naar het Nederlands. Daarbij werd onderzocht in welke opzichten GT beter of slechter presteert dan BF en wat hiervoor de verklaring(en) zou(den) kunnen zijn. Middelen of methode: Als teksten werden een Engels- en Nederlandstalige handleiding en wetenschappelijke bijsluiter geselecteerd. De machinevertalingen werden beoordeeld aan de hand van vier methodes. Eerst werd met de BLEU-methode gezocht naar lexicale overlapping tussen de automatische en menselijke vertalingen. Daarnaast werd een foutenanalyse uitgevoerd waarvan de resultaten gebruikt werden om de accuraatheid en begrijpelijkheid van de machinevertalingen te bespreken. Tot slot toetsten twaalf studenten-vertalers en vier professionele vertalers de vertalingen aan hun eigen criteria en de criteria accuraatheid en begrijpelijkheid. Bij het zoeken naar verklaringen voor kwaliteitsverschillen werden Google en het online vertaalgeheugen van TAUS geraadpleegd om de hypothese te toetsen dat GT voor moeilijke bronsegmenten een beroep doet op bestaande menselijke vertalingen. Resultaten: In het onderzoek kwamen twee constanten naar voren: GT scoorde op alle evaluaties beter dan BF en presteerde vooral heel goed op de bijsluiter. Voor BF was het kwaliteitsverschil tussen beide vertalingen verwaarloosbaar. De vergelijking van de vertalingen van GT met Google en TAUS liet zien dat er voor de medische vertaling veel en voor de technische vertaling praktisch geen overeenkomsten werden gevonden. Bijgevolg kan worden geconcludeerd dat er een positieve correlatie is tussen het aantal beschikbare teksten en de vertaalkwaliteit van GT.
Automatische vertaling. --- Babel Fish. --- Google Translate. --- Kwaliteitsbeoordeling. --- Machinevertaling. --- Rule-based machine translation (RBMT). --- Statistical machine translation (SMT). --- Vertaalkundige studie – Onderzoek van vertalingen – schriftelijke taal. --- Vertaalkwaliteit. --- Vertaalsysteem.
Choose an application
Optimization generally involve picking the optimum solution to a problem considering all the factors or design variables. HOLISHIP (HOLIstic optimisation of SHIP design and operation for life-cycle) is a European Union research project which is a system based approach aimed at developing optimized designs for the future, structure of which is divided into clusters and into several work packages (WP). Structural optimization of RoRo & RoPax vessels pose several challenges due to the unique design features of these type of vessels. This Thesis work concentrates on structural optimization of midship section transverse frame of RoRo & RoPax hulls for minimum weight thus achieving reduction of lightship weight which is one of the major technical requirements during conceptual design phase as part of WP4&WP7 of HOLISHIP project. Rule based tool called ‘STEEL’ by Bureau Veritas is used for the structural & load modeling and further structural analysis of the transverse frame and then optimization loop is established using ‘modeFRONTIER’ and ‘CAESES’ tools to study effect of different design variables. Also the structural optimization loop involving STEEL tool is to couple with a parametric hull in order to enable study of coupled structural analysis for different parametric hull variations. Structural weight is kept as the objective to minimize and design constraints are considered as per applicable Bureau Veritas rules for classification of steel ships. Then surrogate models are generated to replace the optimization loop using Response surface methodology and results obtained with different algorithms like polynomial regression, artificial neural networks etc. are studied further which would reduce complexity associated compared to conventional direct methods
Choose an application
In this thesis, the main objective is to do scantling, weight optimization and possible cost optimization for a platform supply vessel (PSV) for different polar classes with respect to the rules of Llyod’s Register. Taking polar class 2 as an example, scantling of hull structure was calculated based on equations in the rules (rules for PSV, for general cargo ships, for polar class vessel). According to the calculation and data first processed in Excel, all parameters affecting the scantling can be pointed out, which is critical for the later weight optimization. Regarding to the limitation of the rules, the structural optimization methodology has to be limited in the relationship of a variety of parameters in the equations. The main weight optimization operation was performed by MALAB according to all equations and constrains needed. The optimized result including dimension of plate, deck, framing and other elements was obtained in the final comparison. In optimization process, panel based optimization was programmed and the combination of serial and parallel algorithm was selected to optimize CPU time. After finding the relationship of variables by using Excel, quantitative relationship and trend between variables was analyzed by using MATLAB. The top appropriate solution corresponding with the criteria was founded based on the database created by MATLAB. As for some detailed design, this research will specially consider about. In this research, several important trends were concluded which could significantly influence the design process. Cost optimization is a complicated topic for any kind of ship involving variety of aspects. In this research, several main parameters including weight, welding length and part number were identified and taken into account for multi-objective optimization, which was made with the same optimization method. According to Pareto Frontier chart, suitable solution could be found with respect to specific weight of parameters. To check the optimized design, Rules Calc was selected to prove that the design satisfies the requirement of rules. According to plastic design theory, FEM was performed to prove the design is acceptable in strength.
Listing 1 - 10 of 165 | << page >> |
Sort by
|