Advanced Materials Research ISSN: 1662-8985, Vols. 383-390, pp 6747-6754 doi:10.4028/www.scientific.net/AMR.383-390.6747 © 2012 Trans Tech Publications, Switzerland

Online: 2011-11-22

Adaptive Aluminum Extrusion Die Design Using Case-Based Reasoning and Artificial Neural Networks Suthep Butdee Integrated Manufacturing System Research Center (IMSRC) Department of Production Engineering, Faculty of Engineering King Mongkut’s University of Technology North Bangkok (KMUTNB) Bangkok, Thailand e-mail: [email protected] Keywords-Adaptive Aluminium Extrusion Die Deisgn, CBR, ANN, Maximum Production Yield.

Abstract—Aluminum extrusion die design involves with two critical parts; die features and its parameters. Presently, die design process is performed by adaptation approach. The previous dies together with their parameters are collected and stored in a database under the well-memory organization. Case-Based Reasoning (CBR) has been applied and enhanced the design productivity. However, the CBR method has an excellent ability only that an exact or similar design features are existed. Reality, aluminum die design requires regularly changed according to the profile changes. Therefore, it needs to predict optimum parameters to assist in the process of aluminum profile extrusion. This paper presents the redesign process using adaptive method. In this case, CBR & ANN method are combined and development. The CBR uses for die feature adaptation; whereas the ANN is used for parameter adaptation and prediction to a new profile and die design. The actual production yield is given and the ANN will find the best size of billet length in order to receive the maximum yield.

Introduction Aluminum extrusion process can be defined as a metal forming process in which a specific section is produced in length by forcing aluminum under pressure through one or more die orifice. Die shapes are designed according to the profiles requested by customers. Such profiles consist of various cross sectional shapes; solid, hollow or combination and made by specific designed dies. The die shapes are made depending on the profile shapes. They are solid die, hollow die and semihollow die. The typical die includes the die backer enclosed in die ring, bolster and sub-bolster. They are ten assembled in a tool carrier [1]. Die design involves with two critical parts; die features and its parameters. Die design can be created from scratch or adapt to meet with the new die requirement. The previous existing dies together with their parameters are collected and stored in a database under the well-memory organization. CBR method is applied to such concept. However, the CBR method has an excellent ability only that an exact or similar design features are existed. Reality, aluminum die design requires regularly changed according to the profile changes. Therefore, the design process should have mechanism to predict optimum parameters to assist in the process of aluminum profile extrusion. This paper presents the die-redesign process using adaptive method. In this case, CBR & ANN method are combined and development. The CBR uses for die feature adaptation; whereas the ANN is used for parameter adaptation and prediction to a new profile and die design. The actual production yield is given and the ANN will find the best size of billet length in order to receive the maximum yield. Reviews of Related works The aluminum extrusion process is the process of forcing it to flow through a shaped opening in a die. Extruded material emerges as an elongated piece with the same profile as the die opening [3]. The process involves with several significant components such as press machine, die, billet, oven, stretching equipment, sawing machine, etc. In addition, the methods of producing the final profile are also important. They influence quality, efficiency and productivity particularly the information All rights reserved. No part of contents of this paper may be reproduced or transmitted in any form or by any means without the written permission of Trans Tech Publications, www.ttp.net. (ID: 130.203.136.75, Pennsylvania State University, University Park, USA-17/05/16,08:38:33)

6748

Manufacturing Science and Technology, ICMST2011

of making die which relate to use such die in the shop floor. Mostly, they are occurred by operators’ experience. However, modern computer technology is able to assist experienced knowledge management. Data, information and knowledge can be managed by computational method. Case-based reasoning is well known and widely used in various domains of applications. However, the effective die design must be used accurately on the shop floor. Therefore, successfully production control parameters are important and also needed experienced skills. In addition, new dies are often requested by customers and needed to compute and predict based on the old experiences. ANN method is then adopted to join with the CBR. This section presents the previous existing related works of the CBR and ANN both in general and applications to aluminum die design in the aspect of adaptive design concept. Firstly, the die features are explained. There are three main die types as it prior mentioned; solid, hollow and semi-hollow. They are used by different functions. The solid die is used to produce solid profile. It is not too complicated in design and operations, whereas the hollow or semi-hollow die is quite complicated in order to produce the hollow aluminum profiles. The solid profile is used in the purposes of strong and heavy weight and increased surface areas which benefit to well-conduct of electricity current flow. On the other hand, the hollow profile is used for light weight, big sizes, and sometimes for a home living decoration. Presently, Aluminum factory is mainly manufacturing productivity. A press machine should produce a profile smoothly and good quality with maximum production yield. Figure 1 shows the aluminum die components. The solid die is shown on the left. There are four components; feeder, die, backer and bolster. The hollow die is shown on the right. They compose of die mandrel, die cap and bolster. It can be found that both of the two die types are based on geometry shapes and can be defined by features.

Figure1: Aluminum Die Components; Solid Die on the Left and Hollow Die on the Right [2] Feature-based technology on CAD is commonly used to enhance design productivity. Traditionally, CAD tooling is offered for three methods; Solid modeling, Wire-frame modeling and Surface modeling. Solid modeling is one of the effective modeling which is the most similar object to the real part [3, 4]. However, the solid model is still needed to develop in the aspects of fast design respond and design based on the previous parts or features. Therefore, the feature based technology is emerged [5]. The principles of feature can be several characteristics. They are form features, function features, tolerance features, assembly features, material features and property features. The original concept of feature base is derived variant approach of process planning which is believed that the similar parts are should have the similar process plans. The variant approach contains four steps. First is collected the existing process plans and stored in the library. Second, the stored plans are classified, coded and indicated. Third, the new part is input and matched to the system library with the same or similar parts. The matched parts are retrieved and shown. The information is shown which are not only the parts, but also their process plans. Fourth, one or many of the shown process plans are selected and modified to fit with the new part. This concept can save a lot of time and can be applied to many fields of industrial manufacturing; design, planning, production, etc. Feature-based design tools have been applied to many applications. Thomson [6] expresses the applications of the feature concept to reverse engineering to machine maintenance. This method is used to assist CMM in order to get machine part features from the case library. Gayreti and Abdalla [7] present a feature based prototype system for the evaluation and optimization of manufacturing processes. A feature is defined as generic shape carrying product information, which aid design or communication between design and manufacturing. Feature based technology has been applied to combine with case-based reasoning method. The basic idea of casebased reasoning is that new problems can be tackled by adapting solutions that were used to solve

Advanced Materials Research Vols. 383-390

6749

previous problems [8]. It is shown by many papers that are more efficient to solve problems by using previous similar solutions than to generate the entire solutions from scratch. The example applications are in the areas of architecture design [9], in chemical process engineering [10], and in mechanical design [11] as well as design for mass customization [12] etc. CBR consists of four main elements; case problem identification, case retrieval, case adaptation, case storage. All of the cases are linked to data case library that is accepted as knowledge-based library. In addition, there are another two components that assist the case retrieval and the case adaptation. They are case indexing method and case modification rule respectively. As it is found from the previous work, CBR is effectively performed together with feature CAD approach but it is weak on parameter management. ANN method is needed to collaborate. Artificial neural network is a mathematical model for parallel computing mechanisms as same as biological brain. They consist of nodes, linked by weighted connections. Neural networks are constructed by hierarchical layers, which are input, hidden, and output layer respectively. Neural networks learn relationships between input and output by iteratively changing interconnecting weight values until the outputs over the problem domain represent the desired relationship. The mathematical model of the biological neuron, there are three basic components as presented in Figure 2 First, the synapses of the neuron are modeled as weights. The value of weight can be presented the strength o the connection between an input and a neuron. Negative weight values reflect inhibitory connections, while positive values designate excitatory connections. Second component is the actual activity within the neuron cell. This activity is referred to as linear combination. Finally, an activation function controls the amplitude of the output of the neuron. An acceptable range of output is usually between 0 and 1, or -1 and 1. Fixed input x p = ±1 x0 x1

x2

wk 0

wk 0 = bk (bias )

wk 1 wk 2

Activation function

νk ∑ Summing junction

xp

wkp

ϕ (i)

Output yk

θk Threshold

Input signals

Figure 2: Perceptron Neuron Model Each neuron calculates three functions. The first is propagation function as shown in equation 1, νk =

∑w

kj

x j + bk

(1)

Where wkj is the weight of the connection between neuron k and j, yk is the output from neuron k, and bk is the bias. The second is an activation function. The output of a neuron in a neural network is between certain values (usually 0 and 1, or -1 and 1). In general, there are three types of activation functions, denoted by ϕ(•). Firstly, there is the threshold function which takes on a value of 0 if the summed input is less than a certain threshold value (v), and the value 1 if the summed input is greater than or equal to the threshold value.

ϕ (ν)=

{

1 if ν ≥0 0 if ν