Planiranje razvoja proizvoda u automobilskoj industriji

master's thesis

• Title (english): A tool for product development planning in the automotive industry
• Author: Marko Mandić
• Mentor: Dorian Marjanović (mentor)
• Granter: University of Zagreb; Faculty of Mechanical Engineering and Naval Architecture; Zagreb
• Defense date and country: 2016-05-20, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Mechanical Engineering

Abstract

Automobilska industrija jedno je od najzahtjevnijih okruženja u kontekstu razvoja novih proizvoda. Karakterizirana je konstantnim napretkom u proizvodnji i razvojnim procesima, što za dobavljače stvara izuzetno izazovno okruženje te ih prisiljava na kontinuiranu prilagodbu zahtjevima tržišta. Razvoj i proizvodnja spremnika za gorivo je dio industrije koja je pod utjecajem strogih emisijskih normi i uvjetovana je visokim zahtjevima proizvođača automobila. U tako dinamičnom okruženju koje ovisi o globalnim ekonomskim prilikama, od izrazite je važnosti da se projekti provedu efikasno i unutar određenog vremenskog perioda. Projekt se može nazvati efikasnim ako su ciljevi ostvareni na vrijeme i unutar određenog financijskog okvira pri tome da kvaliteta proizvoda ostane na traženom nivou. Tvrtke koje imaju dobro isplanirane i osmišljene projekte imaju prednost nad konkurencijom te imaju su veće šanse da provedu uspješnije projekte. Jedan od najvažnijih aspekata planiranja projekta je izrada financijskog plana te kreiranje kompetitivne ponude prema krajnjem kupcu. Da bi se uspješno izradio financijski plan projekta potrebno je napraviti preciznu procjenu troškova projekta. Postoji nekoliko metoda koje se mogu koristiti za predviđanje troškova, ovisno o kakvom tipu projekta se radi i u kojoj fazi se očekuje izrada procjene. Metode koje se najčešće koriste su intuitivne i spadaju u grupu kvalitativnih metoda. Međutim, za tvrtke koje imaju dugogodišnje iskustvo u proizvodnji specifičnih proizvoda, moguće je primijeniti kvantitativne parametarske metode predviđanja koje se oslanjaju na podatke o prethodnim projektima. Za promatrani slučaj, proizvodnju spremnika za gorivo, parametarsko predviđanje troškova se pokazalo metodom izbora. Parametarske metode obično koriste regresijsku analizu, iako je moguće koristiti i napredne algoritme poput neizrazitih metoda. Izrada alata za predviđanje troškova motivirana je inicijalnom analizom postojećih metoda i uočavanjem nepravilnosti u procjenama. Nadalje, činjenica da odjel program managementa nema standardne procedure u predviđanju troškova, samo doprinosi potrebi za izradu takvog alata. Fokus rada je na analizi metoda za predviđanje i povijesnih podataka te ispitivanje mogućnosti za unaprijeđenje procesa razvoja. Također, cilj je bio napraviti alat za predviđanje troškova primarno za program management odjel te ispitati mogućnosti korištenja regresijskih tehnika u usporedbi s neizrazitim metodama. Predviđanje pomoću alata omogućeno je kombinacijom parametarske metode i utjecaja stručnjaka, iz razloga što pozitivan utjecaj stručnjaka s iskustvom u planiranju projekta može doprinijeti točnosti procjene s obzirom na ograničen uzorak podataka.Alat je napravljen na način da, ukoliko bude mogućnosti za daljnju analizu podataka i povećanje uzorka, omogući poboljšanje procjene jednostavnim editiranjem programskog koda.

Abstract (english)

Automotive industry has proven to be one of the most challenging environments for new product development. Constant advancements in manufacturing technologies and development processes put out by the OEMs, makes it very demanding for the suppliers to keep up the pace with the ever changing market needs. Fuel tank development and manufacturing is considered to be part of the industry impacted by the strict emission rules for automotive industry thus being heavily conditioned by the OEMs emission requirements. In industry that dynamic and with market that depends on global economy, it is of crucial importance to execute projects with great efficiency in a reasonable period of time. Project could be called efficient if the goals are met in time and within planned financial framework while quality of the product is not compromised. Companies that have well planned and thought out projects have a competitive advantage over others, and they are consequently more likely to have more positive project outcomes. One of the most important aspects of project planning is establishing the financial framework and offering a competitive quote to the customer. Way to achieve that is to estimate total project costs as accurately as possible. There are several methods for cost estimation techniques that could be used, depending on the type of the project and the phase in which the estimation is being done. Most common methods are qualitative intuitive ones. However, for the companies with proven track record and extensive experience with the particular product over the years, it is possible to use quantitative methods in the form of parametric or analytical cost estimations relying on history data of previous projects. For observed case, the fuel tank development, parametric cost estimation method has proven to be the method of choice. Parametric methods usually use regression techniques, although there are also possibilities to use advanced algorithms such as fuzzy methods. Tool creation was motivated by initial observation of the existing estimation methods and the fact that the program management department didn’t have any standardized procedures in the cost estimation process. The focus of this thesis is on analysis of the estimation procedures, history data and the possibilities for improvements. Furthermore, the goal was to create the usable tool for cost estimation, primarily for program management department, and to give the overview of using the well known regression techniques as opposed to the advanced fuzzy algorithms. The tool was created in a way that allows influence of an expert on calculated estimation because positive influence can compensate possibly small sample of data. Moreover, the tool was created in a way that if more data is to be analyzed, it would be easy to improve the estimation accuracy by simple intervention to the code.

• Keywords: predviđanje troškova; parametarska metoda; neuronske mreže; spremnik za gorivo; automobilska industrija; višestruka regresija
• Keywords (english): cost estimation; parametric; neural networks; fuel tank; automotive industry; history data; multiple regression
• Language: english
• URN:NBN: urn:nbn:hr:235:956598
• 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: 2019-05-22 16:39:53