Usporedba globalne i lokalne analize osjetljivosti na primjeru potrošnje energije u zgradama

master's thesis

• Title (english): A comparison of the global and the local sensitivity analysis illustrated with special reference to energy use in buildings
• Author: Nikola Bađun
• Mentor: Tea Žakula (mentor)
• Committee member: Hrvoje Jasak (predsjednik povjerenstva)
• Committee member: Marino Grozdek (član povjerenstva)
• Committee member: Tea Žakula (član povjerenstva)
• Committee member: Damir Dović (član povjerenstva)
• Granter: University of Zagreb; Faculty of Mechanical Engineering and Naval Architecture; Zagreb
• Defense date and country: 2020-04-10, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Mechanical Engineering; Process Energy Engineering

Abstract

U okviru ovog rada provedena je usporedba lokalne i globalne metode analize osjetljivosti s obzirom na potrebnu godišnju energiju za grijanje i hlađenje referentne zgrade škole za kontinentalnu klimu u Republici Hrvatskoj. Na temelju detaljnog pregleda literature te preporuka za provedeni tip analize kao globalna metoda odabrana je Sobol metoda. Varirane su vrijednosti 12 ulaznih varijabli energijskog modela zgrade: toplinskog kapaciteta i koeficijenta prolaza topline vanjskih zidova, koeficijenta prolaza topline i stupnja propuštanja Sunčevog zračenja prozora, udjela površine prozora u sjevernoj, istočnoj, južnoj i zapadnoj fasadi, unutarnjih toplinskih dobitaka, konvektivnog udjela toplinskih dobitaka, infiltracije te vremena prekida rada grijanja i hlađenja. Dinamička simulacija energijskih tokova zgrade provedena je koristeći računalni program za energijsko modeliranje TRNSYS, dok je generiranje uzorka i računanje indeksa osjetljivosti provedeno u računalnom programu SIMLAB. Rezultati dviju metoda uspoređeni su s obzirom na rangiranje ulaznih varijabli po utjecaju te na kvantifikaciju tog utjecaja pripadajućim indeksima osjetljivosti. Kao najutjecajnije varijable na potrebnu energiju za grijanje prepoznate su infiltracija i koeficijent prolaza topline zida, dok su za hlađenje to unutarnji toplinski dobici i stupanj propuštanja prozora. Međutim, redoslijed rangiranja utjecaja varijabli razlikuje se s obzirom na dvije metode. Također, lokalna metoda zahtjeva puno manji broj simulacija (73 kod lokalne naspram 13.312 kod globalne u okviru ovog rada) te su njeni rezultati vizualno jednostavni i pregledni. Međutim, rezultati lokalnom metodom pokazuju izrazitu ovisnost o odabranom referentnom slučaju i ne mogu kvantificirati utjecaj interakcija između varijabli. S druge strane, globalna je metoda neovisna o modelu i referentnom slučaju, a prikazuje rezultate koji uključuju sve moguće kombinacije vrijednosti ulaznih varijabli te se stoga smatra pouzdanijom i sveobuhvatnijom metodom.

Abstract (english)

In this work, a comparison between the local and the global sensitivity analysis method has been carried out with regard to the annual energy need for heating and cooling of the reference school building for continental climate in Croatia. The values of 12 energy building model inputs have been varied: heat capacity and heat transfer coefficient of the external walls, heat transfer coefficient and solar heat gain coefficient of windows, window-to-wall ratio of the north, east, south and west façades, internal heat gains, convective part of the internal heat gains, infiltration and duration of night hours without heating or cooling. Dynamic simulation of building energy flows has been calculated using TRNSYS energy modelling software, whereas sample generation and the calculation of sensitivity indices has been conducted using SIMLAB software. The results of the two methods have been compared concerning input factor influence ranking and quantification of that influence using normalized sensitivity indices. Infiltration and heat transfer coefficient of the external walls have been recognized as the two most influential inputs concerning heating energy need, whereas internal heat gains and SHGC were most influential inputs concerning cooling energy need. However, the ranking order differs with respect to the two methods. A smaller number of simulations is needed for the local method (73 for the local method compared to 13.312 for the global method) and the results are visually easy to interpret. However, the results of the local method show a distinct dependence on a base case, and cannot quantify the interaction effects among inputs. On the other hand, the Sobol method is model-independent and does not depend upon a base case. Moreover, its results include all possible combinations of input values so it is therefore considered to be a more reliable and robust method.

• Keywords: analiza osjetljivosti; lokalna metoda; globalna Sobol metoda; energijski model zgrade
• Keywords (english): sensitivity analysis; local method; global Sobol method; building energy model
• Language: croatian
• URN:NBN: urn:nbn:hr:235:884357
• Study programme: Title: Mechanical Engineering; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka strojarstva (magistar/magistra inženjer/inženjerka strojarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2020-04-03 14:17:10