Primjena metode odzivnih površina i neuronskih mreža za modeliranje i procjenu otpornosti na abrazijsko trošenje Poly oxy metilena (original) (raw)
Related papers
2009
Primjena metode odzivnih povr{ina i neuronskih mre`a za modeliranje i procjenu otpornosti na abrazijsko tro{enje poly oxy metilena. U radu je istra`ivano abrazijsko tro{enje poly oxy metilena (POM), za razli~ite uvjete ispitivanja. Za procjenu parametara modela i odziva, primijenjen je centralno kompozitni plan pokusa. Primjenom metode odzivnih povr{ina dobiven je matemati~ki model ovisnosti gubitka mase o primijenjenom optere}enju i putu klizanja. Tako|er, primjenom neuronskih mre`a, razvijen je model za procjenu i testiranje rezultata. Na kraju su uspore|eni rezultati modela, dobiveni metodom odzivnih povr{ina i primjenom neuronskih mre`a.
Comparative Analysis of Abrasive Wear Using Response Surface Method and Artificial Neural Network
Journal of The Institution of Engineers (India): Series D
This research work deals with the application of response surface methodology and artificial neural network-based mathematical modelling of abrasive wear volume for a dry sliding wear of PTFE pin. The experiments were designed based on central composite design. The disc speed, load and sliding distance have been selected as parameters of the process, while the abrasive wear volume has been selected as an output. The ANNOVA test revealed that the disc speed has maximum influence and contributes 28.21% of abrasive wear volume followed by load, which contributes 12.83% of abrasive wear volume. The two models were compared using root mean square error and absolute standard deviation. The artificial neural network-predicted values of abrasive wear volume were found in close agreement with the actual experimental results as compared to response surface methodology predicted results and hence recommended for the similar studies.
Investigation of Abrasive Wear Performances of Different Polyamides by Response Surface Methodology
Tribology in Industry, 2019
This study presents the investigation of abrasive wear performances for polyamides using a Response Surface Methodology (RSM). Tests were carried out in a pin-on-disc using various conditions against SiC abrasive cloth. Box-Behnken Design of RSM was adopted to study the effect of control factors like load, speed and tensile strength of the tested samples on the volumetric wear rate. The experimental results indicated that the volumetric wear rate increased with increasing load, speed and decreased with increasing the tensile strength, but material property was more effective than other factors. Furthermore, the results of analysis of variance showed that tensile strength was predominant factor on the abrasive wear rate, followed by load and spindle speed. The contributions of tensile strength, load and speed were about 33.55 %, 24.45 % and 21.87 %, respectively, while the contribution of square and 2-way of interaction was about 3.04 % and 11.23 %, respectively.
Polymer Composites, 2020
This article presents the abrasion wear behaviour of different industrial wastes filled glass/polyester composites. Two types of fillers such as fly ash and granite dust are chosen with different weight proportions along with polyester and glass fiber for fabricating the composite using hand layup route. Abrasion wear properties of the developed composites are studied in the dry sand abrasion wear test rig as per the ASTM-G-65 standard. The experiments were conducted based on Taguchi design. Between the two fillers, granite filled composite showed better abrasion resistance property as compared with fly ash. Normal load and filler content are found to be the most influential factor for the abrasion loss of the composites followed by abrading distance and sliding velocity. The results are predicted using neural network and also compared with the experimental and regression model data. Abraded surfaces are examined by the SEM to ascertain the different wear mechanisms responsible for abrasion loss of material.
Computational Materials Science, 2010
Inspired by the biological nervous system, an artificial neural network (ANN) approach is a fascinating computational tool, which can be used to simulate a wide variety of complex engineering problems such as tribo-performance of polymer composites. This paper, in this context, reports the implementation of ANN in analyzing the wear performance of a new class of epoxy based composites filled with pine wood dust. Composites of three different compositions (with 0, 5 and 10 wt.% of pine wood dust reinforced in epoxy resin) are prepared. Dry sliding wear trials are conducted following a well planned experimental schedule based on design of experiments (DOE). Significant control factors predominantly influencing the wear rate are identified. An ANN approach taking into account training and test procedure is implemented to predict the dependence of wear behavior on various control factors. This work shows that pine wood dust possesses good filler characteristics as it improves the sliding wear resistance of the polymeric resin and that factors like filler content, sliding velocity and normal load, in this sequence, are the significant factors affecting the specific wear rate. It is further seen that the use of a neural network model to simulate experiments with parametric design strategy is quite effective for prediction of wear response of materials within and beyond the experimental domain.
Journal of Taibah University for Science, 2018
This work employs the T6 heat treatment process to aluminium-clay (Al-Clay) composite consisting of 15 wt% clay. The samples were solutionized at 500°C, 550°C and 600°C, and were quenched in air, oil and water. Selected samples of the heat-treated composite were subjected to wear tests using Denison T62 HS pin-on-disc wear-testing machine in accordance with ASTM: G99-05 standard. The effects of two different loads (4 and 10 N) and three sliding speeds (200, 500 and 1000 rpm) under dry sliding conditions were investigated. The potential of using back-propagation neural network with 4-10-1 architecture was explored to predict the wear rate of the heat-treated composites. The results show that the performance of Levenberg-Marquardt training algorithm is superior to all other algorithms used. The well-trained ANN system satisfactorily predicted the experimental results and can be handy for an optimum design and also an alternative technique to evaluate wear rate.
A Physically-Based Abrasive Wear Model
A simple physically-based model for the abrasive wear of composite materials is presented based on the mechanics and mechanisms associated with sliding wear in soft (ductile)-matrix composites containing hard (brittle) reinforcement particles. The model is based on the assumption that any portion of the reinforcement that is removed as wear debris cannot contribute to the wear resistance of the matrix material. The size of this non-contributing portion (NCP) of reinforcement is estimated by modeling three primary wear mechanisms, specifically, plowing, cracking at the matrix/reinforcement interface or in the reinforcement, and particle removal. Critical variables describing the role of the reinforcement, such as relative size, fracture toughness and the nature of the matrix/reinforcement interface, are characterized by a single contribution coefficient, C. Predictions are compared with the results of experimental two-body (pin-on-drum) abrasive wear tests performed on a model aluminum particulate-reinforced epoxy-matrix composite material.
Wear analysis of polyamide based on a statistical approach
International Journal of Materials and Product Technology, 2017
Polymers are increasingly used for numerous tribological applications. Abrasive wear behaviours of cast polyamides of Kestoil (KT) and Kestamid (KS) were investigated using a Taguchi approach with combination effect of load, speed, distance and grit size. An orthogonal array and analysis of variance (ANOVA) were applied to investigate the influence of process parameters on weight loss. The results indicated that SNR decreased with increasing load and sliding distance while increased with increasing grit size considerably and sliding speed a slightly, but KT showed lower weight loss than that of KS under same conditions. ANOVA indicated that abrasive size and sliding distance had great effects on the weight losses, which were at 49.78%, 13.27% for KT samples and 29.22% and 11.71% for KS samples, respectively. Furthermore, a confidence interval (CI) for predicted mean on confirmation run for KT and KS samples was found about ±5.02, ±3.70 at 95% confidence level, respectively.
Statistical Analysis for the Abrasive Wear Behavior of Al 6061
Journal of Minerals and Materials Characterization and Engineering, 2014
In the present study, a mathematical model has been developed to predict the abrasive wear behavior of Al 6061. The experiments have been conducted using central composite design in the design of experiments (DOE) on pin-on-disc type wear testing machine, against abrasive media. A second order polynomial model has been developed for the prediction of wear loss. The model was developed by response surface method (RSM). Analysis of variance technique at the 95% confidence level was applied to check the validity of the model. The effect of volume percentage of reinforcement, applied load and sliding velocity on abrasive wear behavior was analyzed in detail. To judge the efficiency and ability of the model, the comparison of predicted and experimental response values outside the design conditions was carried out. The result shows, good correspondence, implying that, empirical models derived from response surface approach can be used to describe the tribological behavior of the above composite.
Modelling and analysis of abrasive wear performance of composites using Taguchi approach
Short lignocellulosic fibres are extensively used these days as reinforcing materials in many thermoset and thermoplastic matrices due to their low cost, lower density than inorganic fibres, environmentally-friendliness, and the relative ease of obtaining them. Such fibres would not contribute to the wear and tear of polymer processing equipment and may not suffer from size reduction during processing, both of which occur when inorganic fibres or fillers are used. These fibres can also be easily moulded to wide variety of shapes during composite preparation. However, modelling and analysis of behaviour of composites reinforced with short fibre drawn from agricultural resources has been studied to a limited extent. Particularly, the optimum size of short fibre just capable of transferring the load and flexibility during preparation has not been studied through a simple systematic modelling approach due to the complexity involved in its modelling aspect. To this end, an attempt has been made in this work to study the abrasive behaviour of untreated sugarcane fibre reinforced composites in a simplified manner and develop empirical model. The effect of various test parameters and their interactions have been studied using Taguchi method to find out optimal parameter setting for minimum wear (weight loss). It has been observed that fibre length plays a major role in wear phenomenon. The length of the fibre has been optimized using a popular evolutionary technique known as particle swarm optimization (PSO) and neural network. The study recommends that fibre length should be 7-8 mm for minimum wear of the composites.