Title Razvoj hibridnog estimatora trošenja alata i metoda vođenja alatnog stroja
Title (english) Development of a hybrid tool wear monitoring system and machine tool control methods
Author Danko Brezak
Mentor Dubravko Majetić (mentor)
Committee member Branko Novaković (predsjednik povjerenstva)
Committee member Dubravko Majetić (član povjerenstva)
Committee member Mladen Crneković (član povjerenstva)
Committee member Toma Udiljak (član povjerenstva)
Committee member Jože Balič (član povjerenstva) VIAF: 61785268
Granter University of Zagreb Faculty of Mechanical Engineering and Naval Architecture Zagreb
Defense date and country 2007-05-10, Croatia
Scientific / art field, discipline and subdiscipline TECHNICAL SCIENCES Mechanical Engineering Production Mechanical Engineering
Universal decimal classification (UDC ) 007 GENERALLY Cybernetics. Self-acting systems
Universal decimal classification (UDC ) 004 GENERALLY Computer science and technology. Computing. Data processing
Universal decimal classification (UDC ) 621 APPLIED SCIENCES. MEDICINE. TECHNOLOGY Mechanical engineering. Nuclear technology. Machinery
Abstract U radu je razmatrana problematika razvoja sustava za nadzor trošenja reznih alata i metoda adaptivnog vođenja alatnih strojeva prema postojanosti rezne oštrice, imajući uvidu njihov značaj u projektiranju suvremenih inteligentnih obradnih sustava. U tom je smislu, u prvom dijelu rada, detaljno opisan i analiziran predloženi model hibridnog estimatora parametra trošenja zasnovan na dva serijski povezana modula. Najprije je formiran klasifikacijski modul zasnovan na analitičkom konceptu neizrazite logike bez pravila ponašanja. Time je omogućena primjena neizrazitog odlučivanja bez ograničenja u broju značajki trošenja, čime se podiže stupanj pouzdanosti i robusnosti navedenog modula kao nužnih uvjeta preciznije procjene stupnja istrošenosti. Do konačne se vrijednosti parametra trošenja iz klasificiranog područja trošenja dolazi primjenom modula za estimaciju. On je izveden primjenom regresijskog algoritma metode vektorski podržanog učenja, čime se nastojalo osigurati konačno rješenje u optimalnoj formi s obzirom na odabranu strukturu modula.U drugome je dijelu rada analiziran koncept adaptivnog vođenja alatnog stroja, pri čemu je željeni stupanj istrošenosti u prethodno definiranom vremenu obrade realiziran vođenjem preko parametra brzine rezanja. Osim navedenog kriterija, dodatnu je funkciju cilja predstavljala i maksimalna produktivnost procesa ograničena njegovim tehnološkim karakteristikama. Algoritam vođenja realiziran je umjetnom neuronskom mrežom zasnovanom na radijalnim baznim funkcijama, a šum generiran pogreškom u estimaciji iznosa parametra trošenja filtriran je korištenjem modificirane dinamičke neuronske mreže. Algoritmi umjetnih neuronskih mreža posebno su pogodni u vođenju onih procesa čija je dinamika modelirana nekom od metoda učenja s podacima dobivenim iz snimljenih signala procesa. U ovu se skupinu ubraja i trošenje reznih alata,pri čemu se podaci za strukturiranje sustava za nadzor trošenja mogu iskoristiti i za određivanje strukture ostalih elemenata regulacijskog kruga realiziranih primjenom umjetnih neuronskih mreža.
Abstract (english) In this dissertation two main issues have been taken under consideration having in mind
their significant role in designing of modern intelligent machine tools – development of a
tool wear monitoring system and adaptive machine control algorithm for maintaining tool
wear rate in the predefined cutting time. In the first part of the work a flank wear hybrid
estimator based on two serially connected modules is presented and analyzed. Firstly,
a classification module is designed using analytical fuzzy logic concept without rule
base. Thereby, it is possible to utilize fuzzy logic decision-making without any
constraints in the number of tool wear features in order to raise the module reliability
and robustness as a necessary conditions in precise tool wear parameter estimation.
The estimated wear parameter value is then obtained from the second estimation
module. It is structured using Support Vector Machines regression algorithm which
assures an optimal estimation regarding the structure of the module.
In the second part of the work an adaptive machine tool control algorithms are
analyzed whereat the desired wear level in the predefined machining time is achieved
by adapting the cutting speed. Besides the mentioned criterion, the maximization of the
process productivity, which is constrained by its technological characteristics, is also
taken as an additional cost function. The controller is structured using the algorithm
based on artificial neural network with radial basis activation functions and the noise
generated by the wear parameter estimation error is filtered using modified recurrent
type of neural network. Artificial neural networks are very suitable for controlling the
processes which dynamics is modeled by one of the learning methods using the data
obtained from the measured process signals. Tool wearing is one of those processes
where the data used for the structuring of the monitoring system can also be used for
the modeling of other elements of the control loop realized by artificial neural networks.
Keywords
nadzor trošenja reznih alata
hibridni estimator parametra trošenja
neizrazita logika bez pravila ponašanja
metoda vektorski podržanog učenja
adaptivno vođenje alatnih strojeva
vođenje prema postojanosti reznih alata
Keywords (english)
cutting tool wear monitoring
flank wear hybrid estimator
fuzzy logic without rule base
support vector machines
adaptive machine control
tool durability control
artificial neural networks
Language croatian
URN:NBN urn:nbn:hr:235:405881
Study programme Title: Mechanical Engineering and Naval Architecture Study programme type: university Study level: postgraduate Academic / professional title: doktor/doktorica znanosti, područje tehničkih znanosti (doktor/doktorica znanosti, područje tehničkih znanosti)
Type of resource Text
File origin Born digital
Access conditions Open access
Terms of use
Repository Repository of Faculty of Mechanical Engineering and Naval Architecture University of Zagreb
Created on 2020-12-01 11:58:48