Unapređenje procesa proizvodnje primenom modifikovane PFMEA analize i metoda optimizacije

Abstract

Organizacija i upravljanje procesom proizvodnje u preduzećima automobilske industrije su veoma važni i kompleksni zadaci koji zahtevaju primenu različitih metoda i tehnika. Na pouzdanost ovog poslovnog procesa mogu uticati brojni faktori, kao što su nepovoljni i nebezbedni radni uslovi, neispravnost proizvodne i kontrolne opreme, neadekvatan materijal itd. Postupak identifikacije i analize potencijalnih načina otkaza koji mogu da nastanu u procesu proizvodnje u automobilskoj industriji zasnovan je na primeni Procesne analize efekata i načina otkaza (engl. Process Failure Mode and Effect Analysis, skr. PFMEA). U ovoj doktorskoj disertaciji razvijeno je ukupno pet modela. Prva dva modela zasnivaju na modifikaciji i unapređenju PFMEA analize, primenom metoda višeatributivnog odlučivanja, proširenih intervalnim fazi brojevima tipa-2 (engl. Interval type-2 fuzzy Multi-Attribute Decision Making, skr. IT2FMADM). U prvom modelu je izvršena modifikacija tradicionalne PFMEA analize kroz primenu Metode najbolji-najlošiji proširenom sa intervalnim fazi brojevima tipa-2 (engl. Interval type-2 fuzzy Best-Worst Method, skr. IT2FBWM). Osim toga, rangiranje načina otkaza izvršeno je prema uputstvima iz najnovijeg PFMEA priručnika. U drugom modelu, razvijena je metodologija za rangiranje načina otkaza koja se zasniva na kombinaciji IT2FBWM metode i metode Višekriterijumska optimizacija i kompromisno rešenje proširenom sa intervalnim fazi brojevima tipa-2 (skr. IT2FVIKOR). Treći razvijeni model je model za analizu i izbor dobavljača koji se zasniva na primeni metode Analitičkog hijerarhijskog procesa (engl. Interval type-2 fuzzy Analytic hierarchy process, skr. IT2FAHP) i metode Višeatributivnog poređenja granica oblasti aproksimacije (engl. Interval type-2 fuzzy Multi-Attributive Border Approximation area Comparison, skr. IT2FMABAC), obe proširene sa intervalnim fazi brojevima tipa-2. Četvrti model razvijen je za analizu faktora koji utiču na bezbednost i zdravlje zaposlenih na radnom mestu. Model integriše primenu Delfi metode (engl. Delphi method) i dve metode za rangiranje načina otkaza proširene sa intervalnim fazi brojevima tipa-2, a to su Tehnika za određivanje redosleda preferencija na osnovu sličnosti sa idealnim rešenjem (engl. Interval type-2 fuzzy Technique for Order Preference by Similarity to Ideal Solution, skr. IT2FTOPSIS) i Metoda jednostavnog otežavanja (engl. Interval type-2 fuzzy Simple Additive Weighting, skr. IT2FSAW). Peti model se odnosi na primenu dve heuristike, Genetskog algoritma (engl. Genetic algorithm, skr. GA) i Metode promenljivih okolina (engl. Variable Neighborhood Search, skr. VNS), za određivanje prioriteta načina otkaza. Predloženi modeli testirani su na podacima koji potiču iz preduzeća automobilske industrije. Razmatrana preduzeća nalaze se na različitim nivoima u lancu snabdevanja automobilske industrije i proizvode različite vrste proizvoda.

Organization and management of the production process in automotive industry companies are very important and complex tasks that require the application of different methods and techniques. The reliability of this business process can be affected by numerous factors, such as unfavorable and unsafe working conditions, malfunctioning of production and control equipment, inadequate materials, etc. The process of identifying and analyzing potential failure modes that may occur in the automotive industry production process is based on the application of the Process Failure Mode and Effect Analysis (PFMEA) method. In this doctoral dissertation, a total of five models have been developed. The first two models are based on the modification and improvement of the PFMEA analysis, using Interval type-2 fuzzy Multi-Attribute Decision Making (IT2FMADM) methods. In the first model, a modification of the traditional PFMEA analysis was carried out through the application of the Interval type-2 fuzzy Best-Worst Method (IT2FBWM) with interval type-2 fuzzy numbers. In addition, the ranking of failure modes was performed according to the instructions from the latest PFMEA handbook. In the second model, a methodology for ranking failure modes was developed based on the combination of the IT2FBWM method and the Multi-Criteria Optimization and Compromise Solution method extended with interval type-2 fuzzy numbers (IT2FVIKOR). The third developed model is a model for the analysis and selection of suppliers, based on the application of the Interval type-2 fuzzy Analytic Hierarchy Process (IT2FAHP) method and the Interval type-2 fuzzy Multi-Attributive Border Approximation area Comparison (IT2FMABAC) method. The fourth model was developed for the analysis of factors affecting the safety and health of employees in the workplace. The model integrates the application of the Delphi method and two methods for ranking failure modes extended with interval type-2 fuzzy numbers, which are the Interval type-2 fuzzy Technique for Order Preference by Similarity to Ideal Solution (IT2FTOPSIS) method and the Interval type-2 fuzzy Simple Additive Weighting (IT2FSAW) method. The fifth model refers to the application of two heuristics, the Genetic Algorithm (GA) and the Variable Neighborhood Search (VNS), to determine the priorities of failure modes. The proposed models were tested on data originating from companies in the automotive industry. The considered companies are located at different levels in the automotive supply chain and produce different types of products.