A Knowledge Based System for Cost Estimation of ...

Available online at www.sciencedirect.com

ScienceDirect Procedia Engineering 97 (2014) 2313 – 2322

12th GLOBAL CONGRESS ON MANUFACTURING AND MANAGEMENT, GCMM 2014

A Knowledge Based System for Cost Estimation of Deep Drawn Parts Vishal Naranjea*, Shailendra Kumarb, H.M.A Husseinc a*

Assistant Professor, Department of Technology, University of Pune, Ganeshkhind, Pune- 411007 b Associate Professor, Dept. of Mechanical Engineering, SVNIT, Surat-395007 c Assistant Professor, Advanced Manufacturing Institute, King Saud University, Riyadh- 11421, Saudi Arabia

Abstract Cost estimation of deep drawn sheet metal parts requires substantial knowledge and experience. It is a time consuming task and dependent on the expert personnel available. This paper presents the research work involved in the development of a knowledge based system for cost-estimation of deep drawn sheet metal parts. The proposed system is structured in form of two modules viz., module COMPC for estimation of total manufacturing cost of deep drawn parts and module TOOLC for estimation of total tooling cost of deep drawing tool. The system modules are coded in VB.6.0 in windows operating environment. The system has been tested for various types of axisymmetric deep drawn sheet metals parts taken from industries. System is flexible enough as its knowledge base can be extended or modified easily. © Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license ©2014 2014The The Authors. Published by Elsevier Ltd. (http://creativecommons.org/licenses/by-nc-nd/3.0/). Selection and peer-review under responsibility of the Organizing Committee of GCMM 2014. Selection and peer-review under responsibility of the Organizing Committee of GCMM 2014

Keywords: Cost estimation, Deep drawn sheet metal parts, Knowledge based system

1. Introduction Cost estimation of sheet metal parts is a complex and time consuming task that requires a great deal of knowledge and experience. In today’s competitive environment, early and accurate cost estimation has become vital for taking important decisions such as material selection, production processes and mainly morphological

* Corresponding author. Tel.: +91-020-25601270; fax: +91-020-25690055 E-mail address: [email protected]

1877-7058 © 2014 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license

(http://creativecommons.org/licenses/by-nc-nd/3.0/). Selection and peer-review under responsibility of the Organizing Committee of GCMM 2014

doi:10.1016/j.proeng.2014.12.476

2314

Vishal Naranje et al. / Procedia Engineering 97 (2014) 2313 – 2322

characteristics of the product (Duran 2012). Making a wrong decision in cost estimation i.e. underestimates and over estimate of costs slow down the new product development process. The accuracy of cost estimates is therefore very essential for survival of an organization (Bouaziz et al. 2006). Deep drawing process is widely used for manufacturing of complex axisymmetric and irregular hollow shaped sheet metal parts. Deep drawn parts are used in a variety of high-tech products, auto parts and generic household products. Accurate cost estimation of a deep drawn sheet metal part is tedious and highly experience-based task. From a methodological point of view, cost estimation may be based on qualitative or quantitative approaches. There are various approaches used by researchers such as expert judgment, heuristic rules, statistical methods (parametric and neural network), analogous, generative/analytical etc. In recent years, knowledge based system (KBS) approach of artificial Intelligence (AI) has been used to solve the cost estimation problems. Bidanda et al. (1988) developed an intelligent system for castability analysis and product cost estimation for a permanent-mould casting company. Welsh (1988) proposed a customer order relaying assistant (CORA) expert system that accesses relay types and style numbers and uses hierarchically organized inheritance functions to provide bills of material and quotations. An expert system methodology for material selection to satisfy required design and production criteria was developed by Shetty and Napolitano (1990). Poli et al. (1990) described how costs can be trimmed by using a systematic method to rate design based on certain casting parameters like design features, casting size etc. Bock (1991) developed an expert system for materials/process combination selection. This system has four basic modules including knowledge base of materials, knowledge acquisition mechanism, design analysis, and cost estimation. Abdallah and Knight (1994) developed an expert system for the concurrent product and process design of mechanical parts. Luong and Spedding (1995) described generic KBS for process planning and cost estimation in the hole making process. Wei and Egbelu (2000) proposed a framework to estimate the lowest product manufacturing cost from the AND/OR tree representation of an alternative process. Ben-Arieh (2000) proposed a hybrid cost estimation system for rational parts that uses a combination of the variant approach and explicit cost calculation. Shehab and Abdallah (2002) developed a system that has the capability of selecting material, as well as machining process and parameters based on a set of design and production parameters. This system also estimates the product cost throughout the entire product development cycle including assembly cost. Jung (2002) proposed a feature-based cost estimation system for machined parts. Seo and Ahn (2006) presented a learning algorithm based estimation method for the maintenance cost as life cycle cost of products. Verlinden et al. (2008) developed a system for cost estimation of sheet metal operations by analyzing geometric data from CAD file and using regression techniques and neural networks. Che (2010) used a Particle swarm optimization (PSO) based approach in training artificial neural network for cost estimation of plastic injection molding. Deng and Yeh (2011) applied least squares support vector machine method for solving the problem of estimation of manufacturing cost of airframe structural projects. Duran and Rodriguez (2012) presented a manufacturing cost estimation system for piping elements using artificial neural networks. Wang et al. (2013) integrated particle swarm optimization and back propagation neural network algorithm for cost estimation of plastic injection molding parts. From the critical review of published literature, it is found that no research efforts have been applied to develop a computer-aided system for cost estimation of deep drawn sheet metal parts. Therefore, to reduce complexity and dependency on domain experts, there is stern need for development of a KBS for cost estimation of deep drawn parts to assist die designers of sheet metal industries. Such a system will certainly help to improve the productivity and to reduce dependency on domain experts. In the present research work a KBS has been developed to impart expert advices on cost estimation of deep drawn parts. Production rule based approach has been used for development of the proposed system. 2. Proposed System for cost estimation of deep drawn parts The proposed KBS labeled as AUTOCDP (Automatic cost estimation of deep drawn parts) is developed for automatic cost estimation of tooling and manufacturing cost of deep drawn parts. Development procedure of construction of system modules of proposed system includes knowledge acquisition, knowledge representation in the form of production rules, verification and sequencing of production rules, selection of operating system and identification of suitable hardware, selection of programming language, construction of knowledge base, choice of search strategy and preparation of user interface. Knowledge required for cost estimation of deep drawn parts is

Vishal Naranje et al. / Procedia Engineering 97 (2014) 2313 – 2322

2315

essentially collected by online and off-line consultation with experienced die designers, process planners, and also referring research articles, catalogs and manuals of various sheet metal industries. The acquired knowledge is represented in the form of production rules of ‘IF-Then’ clause. The production rules framed for each module are cross-checked from a team of die design experts by presenting them IF condition of the production rule of IF-THEN variety. These rules are coded in Visual Basic 6.0. The proposed KBS is implemented on a personal computer with hardware configuration as core2 Duo processor with 4 GB RAM. To construct the knowledge base of proposed system, knowledge acquired from various sources of knowledge acquisition is stored in form of production rules of IF-THEN variety. The rules and knowledge base are linked together by an inference mechanism. Inference engine searches the knowledge base as to what rule(s) to fire and then executes the action part of the fired rule. The whole system comprises of more than 300 production rules. A sample of production rules incorporated in the proposed system is given in Table 1. Production rules in each module are arranged in a structured manner. This arrangement allows insertion of new production rules even by relatively less trained knowledge engineer. In the proposed system forward chaining search strategy is used for searching the solution for cost estimation of deep drawn parts. The graphical user interface (GUI) is created under WINDOWS environment using Visual Basic 6.0. Table 1. A sample of production rules incorporated in the proposed system Sr. No.

IF (condition)

THEN(action)

1

Operation = Deep drawing, and Material = Mild steel

Set density = 0.007359 Kg/mm3, and Rate (per Kg)= Rs. 50, and Scarp value (per Kg) = Rs. 20

2

Operation = Deep drawing, and Material = deep drawing quality steel

Set density = 0.0083 Kg/mm 3, and Rate (per Kg) = Rs. 65, and Scarp value (per Kg) = Rs. 25

3

Operation = Deep drawing, and Material = Cast Iron

Set density = 0.009867 Kg/mm3, and Rate (per Kg) = Rs.70, and Scarp value (per Kg) = Rs.30

4

Machine = Lathe

Set machining cost per hour = Rs.290

5

Machine = Milling

Set machining cost per hour = Rs.290

6

Machine = Surface grinding

Set machining cost per hour = Rs. 290

7

Material = P 20

Set density = 0.005687 Kg/mm3, and

8

Operation = Deep drawing, and Material = Aluminum

Set tensile strength = 315MPa, and

9

Operation = Deep drawing, and Material=Aluminum Alloy

Set tensile strength = 414 Mpa, and Density = 0.003 Kg/mm3, and Rate (per Kg) = Rs. 450

10

Required drawing force ≤ 50 KN, and Press tonnage < 75 ton

Select deep drawing press = 75 ton, and Machine operating cost per stroke = Rs. 0.3

12

Operation = Deep drawing, and and 75 ton < Press tonnage < 100 ton

Select deep drawing press = 100 ton, and Machine operating cost per stroke = Rs.0.4

13

Operation = Deep drawing, and 100 ton < Press tonnage < 150 ton

Select deep drawing press = 150 ton, and Machine operating cost per stroke = Rs. 0.5

14

Operation = Deep drawing, and 50 ton < Press tonnage < 75ton

Select deep drawing press = 75 ton, and Machine operating cost per stroke = Rs. 0.3

Rate (per Kg) = Rs. 200 Density = 0.0026 Kg/mm3, and Rate (per Kg) = Rs. 350

2316

Vishal Naranje et al. / Procedia Engineering 97 (2014) 2313 – 2322

START Enter part and machine details

Data File COST.DAT

Module COMPC (Module for calculation of raw material and manufacturing cost of deep drawn part)

List of file extracted BLDIA.DAT (Data file of Blank diameter) DRWSEQ. DAT (data file Data file of draw sequence) SLWS.DAT (Data files of strip size)

Read Data File BLDIA.DAT

Read Data File DRWSEQ.DAT

Read Data File SLWS.DAT

Machines, labour and materials database

Display raw material manufacturing cost of deep drawn parts

Data File COMPC.DAT

Module TOOLC (Module for calculation of material and manufacturing cost of tooling) List of file extracted DBLCK.DAT (Data file of die block) STRP.DAT (Data file of stripper and stripper plate BHOLD.DAT (Data file of Blank holder) DIMDS.DAT (Data file of dimensions of die set) Read Data File DBLCK.DAT

Read Data File STRP.DAT

Read Data File BHOLD.DAT

Read Data File DIMDS.DAT

Display tooling material and manufacturing cost STOP Fig. 1 Execution of the Proposed System AUTOCDP

Data File TOOLC.DAT

Vishal Naranje et al. / Procedia Engineering 97 (2014) 2313 – 2322

2317

The proposed system labeled as AUTOCDP is structured into two modules. Execution of the system AUTOCDP is depicted in Fig.1. The system works in conjunction with KBS already developed by authors for process planning of deep drawn parts (Naranje and Kumar 2013a), strip layout design (Naranje and Kumar 2013b) and selection of die components of deep drawing die (Naranje and Kumar 2012). The outputs of various modules developed for process planning system and selection of die components are stored in different output data files. These data files are recalled automatically during execution of the proposed system. Initially system invites the user to enter part data information through GUI. The system stores these part data automatically in a data file labeled as COST.DAT. This data file is further utilized in estimation of manufacturing cost of deep drawn parts as well as tooling cost. The first module namely COMPC of proposed system AUTOCDP is developed to calculate the cost of raw material and manufacturing including various overheads for production of deep drawn parts. Inputs to this module are machine type, machining rate, part cost rate, labour cost etc. which are taken automatically from COST.DAT data file. Other inputs such as blank volume, number of draws, and strip size are taken automatically from data files generated during the execution of already developed modules BLDIA, DRWSEQ and SLWS for blank size calculation, determination of number of draws required, and strip size respectively (Naranje and Kumar 2013a, 2013b). Outputs generated by the module COMPC in form of cost estimation of raw material and manufacturing are saved automatically in a data file labeled as COMPC.DAT. The second module labeled as TOOLC is developed to impart advices to the die designer and process planner for estimation of tooling cost. The Inputs are automatically taken from data files DBLCK.DAT, STRP.DAT, BHOD.DAT, DIMDS.DAT generated during execution of modules DBLCK, STRP, BHOLD and DIMDS already developed for selection of die components (Naranje and Kumar 2012). Inputs are also taken from the data base of machines, materials and labour. Outputs of the module TOOLC in form of cost estimation of tooling material and manufacturing are automatically saved in a data file labeled as TOOLC.DAT. 3. Validation of Proposed System The proposed system AUTOCDP is capable to estimate manufacturing cost, tooling cost and other costs involved in the production of deep drawn parts. The system modules have been validated by considering the problem of costing for wide variety of stepped and tapered deep drawn parts. Input data entry and expert advices generated during the execution of system modules for one industrial sheet metal deep drawn component (Fig.2) are depicted in Figures 3 and 4. Recommendations imparted by the proposed system for example part are found to be reasonable and very similar to those actually used in sheet metal industries.

Fig. 2 Industrial sheet metal component (Material: Aluminium, Sheet thickness: 2 mm, All dimensions are in mm)

2318

Vishal Naranje et al. / Procedia Engineering 97 (2014) 2313 – 2322 Table 2: Typical input and output during execution of proposed system for example part (Fig. 2) S.No. 1 2 3

Input Select shape of component Enter tool name Enter material of component

Example Data entry Stepped component Draw tool 2 Aluminum

4

Enter thickness of component

2

5

Enter the part feature dimensions of component

6

Enter the no of draws required

D1=100 mm, D2=150 mm, D3=50 mm, S = 80 mm 3

7

Select die component through user interface Select material of top plate Enter length, width and height of top plate

Top plate

10 11

Select other die component Enter length, width and height of bottom plate through user interface

Bottom plate 200 mm,100 mm , 30mm

12

Select the machines required to manufacture of top plate

Milling =1 hr, Surface grinding = 1.5hr Jig boring = 0.5 hr

13

14 15

Similarly select the die components through user interface Punch Die Stripper plate Punch holder Guide pillar Guide bush Back plate Blank holder Die plate Enter labour cost per hr Enter inspection per hr

Rs. 40 Rs. 45

Rs. 4600/Rs. 8000/-

16 17

Enter total designing cost per hr Enter total overhead per hr

Rs. 75 Rs. 50

Rs. 10000/Rs. 70000/Total manufacturing cost of die components = Rs. /125000/-

8 9

Mild steel 200 mm, 100 mm, 25 mm

output Tensile Strength = 315 MPa Density = 0.0026 kg/mm3 Rate = 350 Rs. Blank volume = 86350 mm3 and blank mass = 0.24 kg Component cost =Rs. 7.8 Scrap cost= Rs.1.1 Total cost of material for component = Rs. 6.70 Press tonnage = 243 ton Blank diameter =173.20 mm Rate per draw = Rs.1 Component manufacturing cost = Rs. 2.25 Final component cost = Rs. 8.93

Volume = 500000 mm3, Mass = 3.9 kg Raw material cost of top bottom plate = Rs. 195 Bottom plate selected Volume = 600000 mm3 Mass= 4.71 kg Raw material cost of bottom plate = Rs.235 Total manufacturing cost of top plate = Rs. 1050/-

Total cost of material for all die components = Rs. 36500/-

Total tooling cost = Rs. 2,54,167

Vishal Naranje et al. / Procedia Engineering 97 (2014) 2313 – 2322

Fig. 3 (a) Output of module COMPC

Fig. 3 (b) Output of module COMPC

2319

2320

Vishal Naranje et al. / Procedia Engineering 97 (2014) 2313 – 2322

Fig. 4 (a) Output of module TOOLC.

Fig. 4 (b) Output of module TOOLC.

Vishal Naranje et al. / Procedia Engineering 97 (2014) 2313 – 2322

2321

Fig. 4 (c) Output of module TOOLC

4. Conclusion: A knowledge based system has been developed for automatic cost estimation of deep drawn sheet metal parts. This system is based on production rule based approach of AI and coded using Visual Basic. The system is capable to assist semi-skilled engineers or technicians to estimate the total cost for production of deep drawn parts accurately and consistently within a short time. Further it reduces quotation time and increase the confidence while placing an order. System is successfully validated for wide variety of deep drawn parts of various small scale and medium scale sheet metal industries. In future similar type of approach can be used for cost estimation of other sheet metal processes like blanking, piercing, stamping etc. Acknowledgements This work is supported by NSTIP strategic technologies programs, Grant number (12-INF2816-02) in the Kingdom of Saudi Arabia.

References [1] H. S. Abdalla, J. Knight, An expert system for concurrent product and process design of mechanical parts, Journal of Engineering Manufacture. 208 (1994) 167–172. [2] D. Ben-Arieh, Cost estimation system for machined parts, International Journal of Production Research. 38 (2000) 4481–4494. [3] B. Bidanda, M. Kadidal, R. E. Billo, Development of an intelligent castability and cost estimation system, International Journal of Production Research. 36 (1998) 547- 568.

2322

Vishal Naranje et al. / Procedia Engineering 97 (2014) 2313 – 2322

[4] L. Bock, Material Process selection methodology, Cost Engineering. 33(1991) 9-14. [5] Z. Bouaziz, J. Ben Younes, A. Zgha, Cost estimation system of dies manufacturing based on the complex machining features, International Journal of Advanced Manufacturing Technology. 28 (2006) 262–271. [6] Z. H. Che, PSO-based back-propagation artificial neural network for product and mold cost estimation of plastic injection molding. Computers and Industrial Engineering, 58 (2010) 625–637. [7] S. Deng, T. Yeh, Using least squares support vector machines for the airframe structures manufacturing cost estimation, Journal of Production Economics. 131(2011), 701-708. [8] O. Duran, J. Maciel, N. Rodriguez, Comparisons between two types of neural networks for manufacturing cost estimation of piping elements, Expert Systems with Applications, 39(2012), 7788-7795. [9] J. Jung, Manufacturing cost estimation for machined parts based on manufacturing features, Journal of Intelligent Manufacturing, 13(2002) 227-238. [10] L.H.S. Luong, T. Spedding, An integrated system for process planning and cost estimation in hole making, International Journal of Advanced Manufacturing Technology. 10(1995), 411–415. [11] V. Naranje, S. Kumar, A knowledge based system for selection of components of deep drawing die, American Journal of Intelligent Systems. 2(2) (2012), 1-11. [12] V. Naranje, S. Kumar, A Knowledge based system for process planning of axisymmetric deep drawn parts, Key Engineering Materials. 549 (2013a), 239-246. [13] V. Naranje, S. Kumar, A knowledge-based system for strip layout design for progressive deep drawing die, International Journal of Computer Applications in Technology, 48 (3) (2013b), 222-234. [14] C. Poli, E. Sunderland, L. Fredette, Trimming the cost of die castings, Machine Design. 5 (1990), 99-102. [15] K. Seo, B. Ahn, A learning algorithm based estimation method for maintenance cost of product concepts, Computers and Industrial Engineering 50 (2006), 66–75. [16] E. Shehab, H. Abdalla, An intelligent knowledge-based system for product cost modeling, International Journal of Advanced Manufacturing Technology. 19 (2002), 49–65. [17] D. Shetty, F. Napolitano, Expert system methodology for optimum metal selection, In: Intelligent Processing of Materials (New York: American Society of Mechanical Engineers, Materials Division), 1990, pp.113-126. [18] B. Verlinden, J. Duflou, P. Collin, D. Cattrysse, Cost estimation for sheet metal parts using multiple regression and artificial neural networks: A case study, International Journal of Production Economics. 111(2008) 484-492. [19] H. S. Wang, Y. N. Wang, Y. C. Wang, Cost estimation of plastic injection molding parts through integration of PSO and BP neural network, Expert Systems with Applications. 40 (2013) 418-428. [20] Y. Wei, P. Egbelu, A framework for estimating manufacturing cost from geometric design data, International Journal of Computer Integrated Manufacturing. 36 (2000), 50–63. [21] W. L Welsh, System does an expert job on a PC, Manufacturing Systems. 6 (1988) 26- 29.