Wyniki wyszukiwania - Biblioteka Nauki

1

Cold forming techniques for the manufacture of steel structual sections

100%

Macdonald M. , Heiyantuduwa M. , Harrison D. K.

|

tom nr 2

40-43

2

Product quality monitoring using artificial neural networks

100%

Wood C. H. , Harrison D. K. , Szpytko J.

|

tom Vol. 53, nr 7

173-183

PL

Rozwój automatyzacji maszyn i urządzeń wymaga nowego spojrzenia na współdziałanie w układzie człowiek/operator - urządzenie. Automatyka przyczynia się do poprawy bezpieczeństwa i niezawodności układów antropocentrycznych, jednakże wiedza i umiejętności operatora są nadal istotne w procesie eksploatacji maszyn i urządzeń. Rozwijane są działania zorientowane na ciągłe monitorowanie stanu technicznego procesu typu CM (ang. Condition Monitoring). Przedmiotem artykułu są wyniki badań noża tokarki podczas operacji skrawania. Rejestrowane podczas skrawania wibracje są analizowane w odniesieniu do uzyskanej jakości obrabianej powierzchni. Jakość obrabianej powierzchni była przedmiotem klasyfikacji dla potrzeb sieci neuronowych. Na podstawie przeprowadzonych badań sformułowano korelacje pomiędzy poziomem wibracji noża skrawającego tokarki a jakością obrabianego przedmiotu. Otrzymane z badań wyniki umożliwiły opracowanie narzędzia uniemożliwiającego realizację operacji w rezultacie, której jakość obrabianej powierzchni nie będzie zadawalająca. Opracowane narzędzie wykorzystuje w procesie decyzyjnym sieci neuronowe.

EN

With the trend towards the use of automated systems apparent and the removal of direct human contact in the manufacturing of components, the expertise associated with human operators is also being eroded. While automation can reduce human errors, improve safety and economy of operation, it is felt that the loss of benefits felt due to the human expertise might be recovered. The process of condition monitoring (CM) involves monitoring the condition of a particular machine to attempt to detect the adverse changes that would indicate that the performance of the machine is failing. This paper reports the results of acquiring vibration readings during the cutting operation of a centre lathe and analysing the data off-line with a view to determining the state of the Surface Finish produced during the cutting procedure. The process of using the Artificial Neural Networks to classify the measured signature analysis data into distinguished classes of Surface Finish quality will be discussed. This is part of a body of on-going work which aims to show how automated techniques employing Artificial Intelligence are preferred if CM is to make a real impact in the manufacturing industry.

3

A novel technique for driveline assembly applications

100%

Coban H. , Harrison D. K. , De Silva A. K. M.

|

tom Vol. 18, No. 4

55-63

EN

A novel technique for drivetrain assembly - Mill-Knurling and Press-Fitting (MKPF) is projected as a substitute to laser welding or bolting. This joining practice involves the press fitting of two mating surfaces, one with mill-knurled teeth and the other which is of a comparatively softer material, enabling it to stream over the teeth making a joint. This process has been applied within an automobile rear axle differential which is subjected to random torque loads. Experimental analysis and simulation has been used to evaluate the serviceable viability and the latent benefits of mill knurled joints with both laser welded and bolted joints currently used by BMW. Assumptions such as total cost of planning, research and development, total investment, total specific resources, raw material and manufacturing costs were used to evaluate mill knurling as an alternative to laser welding and bolted assemblies. The MKPF method has been successfully applied to assemble rear axle differential cases to the bevel gears. The costs, weight and size estimations are very positive in comparison to the competitive methods of laser welding and bolting. The noteworthy weight saving will strengthen the efforts by the automotive industry to reduce the emission levels of vehicles. The pitch flank deviation values which are critical for the life cycle duty of the bevel gear need to be further investigated in order to achieve comparable results with laser welding.

4

An efficient information system for automated process planning

100%

Pang J. , Harrison D. K. , MacDonald M.

|

tom Vol. 6, nr 2

49-63

EN

Process planning is one of the most important components in manufacturing industry. It may be considered as a bridge between design and manufacturing. Tremendous efforts have been made to develop automated process planning systems during the last three decades. However, their effectiveness is stili far from satisfactory. Process planning is a very complicated and complex task. It not only requires a good deal of technique specific expertise and knowledge but also is very dependent on dynamic manufacturing resources. With the recent development in computer information systems, especially Artificial Intelligence (Al) and advanced modelling techniques, Computer Aided Process Planning (CAPP) has greatly benefited from the new information processing capabilities. By efficiently combining some of these techniques such as object-oriented techniques, deductive mechanisms, logic programming and fuzzy logic, an integrated and intelligent information representation is introduced to facilitate automatic process planning in this paper. It can effectively satisfy the requirements of representing the process planning information and also provide an open structure for the information exchange within the Computer Integrated Manufacturing (CIM) environment.

5

The identification of on-line surface finish via an ANN application

100%

Wood C. H. , Harrison D. K. , Sihra T. S. , Szpytko J.

|

tom Vol. 6, nr 2

87-93

EN

In this paper, a brief description is provided of a comprehensive experiment conducted to investigate the effect of the vibration characteristics of a lathe as the tool tip is worn through repeated machining operations. The characteristics are measured as vibration spectra and visual pattern techniques are employed to detect trends in these spectra, which indicate the variance in surface finish quality. Artificial Neural Networks (ANNs) are then employed to provide a more powerful tool for feature extraction and combined with surface finish measurement taken during the cutting operation, a correlation between these two parameters is undertaken. The results show that ANNs are capable at complementing existing tool wear monitoring techniques and potentially they could be applied to a wider spectrum of condition monitoring strategies developed to date.

PL

Przedmiotem artykułu są badania relacji pomiędzy jakością powierzchni (gładkość) przedmiotu obrabianego z użyciem tokarki, a wzbudzanymi drganiami obserwowanymi na jej nożu skrawającym (dla powtarzalnych cyklicznie operacji). W rezultacie przeprowadzonych badań sformułowano relacje pomiędzy gładkością powierzchni i drganiami na nożu, które umożliwiły zastosowanie sieci neuronowych ANN do prognozowania tendencji zmian w stanie technicznym tokarki. Opracowane narzędzie ANN umożliwia prowadzenie w układzie on-line oceny jakości wykończenia powierzchni obrabianego przedmiotu.