Expert system for automated offering following a feature based approach.

Duerr, Holger ; Teich, Tobias ; Anh, Tran Ngoc 等

1. INTRODUCTION

The following procedures included in the three steps given below are focused in a research project, which lead to this paper:

* Feature based design (FBD) of components and manufacturing elements by means of a feature database--Computer Aided Design (CAD)

* Feature based main planning and specification of the manufacturing equipment--Computer Aided Manufacturing (CAM)

* Offering based on calculated machining times and pre-calculated production costs

This contribution is focused on making the offers in an integrated automated way. Thus, we could demonstrate that offering can be executed automatically and with sufficient accuracy of results. Within this context, the results obtained by main planning of the technological sequence are absolutely sufficient at the offering stage. When determining costs with this methodology, production time is calculated in a feature oriented way, and the characteristics of machine cost per hour are taken into consideration.

2. FEATURE BASED OFFERING

2.1 Overall concept

The general feasibility of the product is verified according to geometric, technological and physical properties. To do this, queries with corresponding information about individual technical elements of the product are sent to the database of the resource model. In this step, it can be determined whether the resources necessary to manufacture the individual geometrical shapes within particular constraints are existent. As a result, we obtain a set of manufacturing steps for the product. These steps are the basis for determining the manufacturing times and costs.

2.2 Feature-based design of the product's requirements

"Features" denotes the following: Features are regarded as objects used to represent work pieces, and these objects have functional, geometric and technological properties. They can also include specific application-oriented data, such as times and costs (Ehrenspiel & Kiewert 2007).

According to our methodology, the following steps must be followed for feature representation:

* Define a work piece coordinate system

* Represent global information

* Decompose work piece geometry by feature classes (base-, transition-, manufacturing- and machining shapes)

* Allocate dimensional relations to features

* Allocate technological requirements

* Generate feature catalogue as a means of communication among designing engineers and production planners.

The feature classes should be defined in a manner as closely related to the product as possible. Differentiating feature classes first of all makes it possible to simplify representation of location. The base shape features represent the work piece's initial shape. Manufacturing- and Machining features are understood as surface modifications to base shape features. Transition shape features are edge modifications of marginal contours. From the perspective of costs, they are irrelevant to the estimating process. For application of the feature description model, STEP (Standard for the exchange of product model data) is used. This ISO-Standard is segmented into different AP (Application Protocol). For our requirements AP 224 is suitable and is used for description of the demanded product. In Chapter 2.3, we introduce modeling of the manufacturing resources and rough process planning as another prerequisite for determining manufacturing times and costs.

2.3 Competency-based modeling of the manufacturing resources and rough process planning

The individual SMEs' competencies include human resources (engineering services, qualifications and levels of experience, know-how etc.), manufacturing resources (machinery, equipment and related items) and technologies ( manufacturing methods, etc.) (Durr et al. 2001; Hanel et al. 2006; Lassig et al. 2006). Competency-based descriptions of resources make it possible to balance product requirements and manufacturing capacities with each other, in order to generate technologically relevant process variants (Teich et al., 2008).

The technologically feasible process variants are generated as a result of feature-based balancing of product requirements (demand vector) with available resources (offer vector). This means, that for each feature of the customer's demand, a machine is assigned automical.

This process information is the basis for the concept used to calculate the manufacturing costs outlined in chapter 2.4.

2.4 Determining manufacturing times and costs

The determination of manufacturing time and costs is divided into two steps. For the accurate time determination, our model relies on the on the REFA time diagram (REFA 1997), which is briefly introduced in figure 1.

[FIGURE 1 OMITTED]

Machining time [t.sub.h] is calculated based on features with sufficient accuracy. Time is computed according to the known equations for each manufacturing technology. The determination of manufacturing costs is the second step in the automated offering model introduced in this paper. Costs can be determined using different calculation methods like global estimating of allowances, direct cost calculation or estimating the machine costs per hour.

To achieve sufficient accuracy when calculating the production cost, we chose the paradigm based on machine cost per hour. The function of machine cost per hour includes the addition of overhead costs proportional to the used machine time. Machine costs per hour [k.sub.hM] are calculated using the ratio of the total of machine-related manufacturing overhead cost types per annum [K.sub.Mach]. to possible annual machine time [T.sub.N] at the planned degree of utilization:

[k.sub.hM] = [K.sub.Mach.]/[T.sub.N] ([member of]/h]) (1)

The Machine-dependent manufacturing overhead cost (Mach.-depend. FGK) of a batch corresponds to machine costs per hour [k.sub.hM] multiplied by manufacturing time [T.sub.Fz].

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (2)

The machine dependent production overhead costs are calculated automatically based on the feature description applied to each, demand and supply. The cost of production can then be determined by the scheme shown in table 1, including direct costs for production material, labor and special costs of production as well as overhead costs for material and production.

3. CONCLUSION

The project focuses on integrated and automated estimating based on feature modeling and competency-based resource description in virtual production networks. The authors were able to verify that it is possible to draft estimates using a method that is both automated and sufficiently accurate. Generating a rough plan of the technological sequence is entirely sufficient at the estimate stage. Cost calculation is centered on feature-oriented calculation of production time and uses characteristics of machine costs per hour. Implementation of the entire concept on a computer was demonstrated with selected manufacturing examples.

4. REFERENCES

Durr, H.; Mehnert, J. & Teich, T. (2001) Integration of decentral Developing and Process Planning Competence into the model of Extended Value Chain Management, Proceedings of DAAAM, International Symposium: Intelligent Manufacturing & Automation, 24.-27. October, S. 125, Jena, ISBN-Nr.: 3-901509-19-4

Ehrlenspiel, K.; Kiewert, A. & Lindemann, U. (2007) Kostengunstig Entwickeln und Konstruiere, Springer, Heidelberg

Hanel, T.; Shatyan, A. & Durr, H.(2006) Development and evaluation of competence profiles in non-hierarchical production networks, Proceedings of FAMP06, Flexible Automation & Intelligent Manufacturing, Limerick, June 26th-28th, Ireland, ISBN--Nr.: 1874653933

Lassig, J.; Heinrich, S. & Durr, H. (2006) Supply Chain Executive Monitor for Controlling and Failure Management in Supply Chains, DAAAM International Scientific Book 2006, S. 395-408, Published by DAAAM International, ISBN 3-901509-47-X, ISSN 1726-9687, Vienna, Austria

Mehnert, J. & Durr, H. (2004) Generation and Evaluation of technological Search Patterns for the target-exact Addressing of potential Suppliers Proceedings of FAM'04 Flexible Automation & Intelligent Manufacturing, Toronto, Canada, July 2004, ISBN--Nr.: 0-662-37218-2

REFA (1997) Worterbuch der Arbeitswissenschaft, Verband fur Arbeitsstudien und Betriebsorganisation (Hrsg.), Carl Hanser Verlag, Munchen

Teich, T.; Zimmermann, M.; Unger, K. & Militzer, J. (2008) Functional Characterization of Resources for Automatic Request Handling, Proceedings of the Fifteenth International Working Seminar on Production Economics, Innsbruck (Austria), March 3th-7th, vol. 4, pp. 213-225

Table 1: Basic methodology of cost calculation for estimating

 Selling price

 Prime costs

 Cost of production (model for
 feature-oriented estimating)

 Material costs Manufacturing costs

Production material Direct labor costs

Material overhead cost Machine-dependent production
(allowance proportional to overhead costs (feature-oriented)
production material)
 Residual production overhead costs
 (allowance proportional to direct
 labor costs)

 Special direct costs of production

Administration- and sales overhead costs (allowance
proportional to production costs)

Special direct selling costs

Profit margin (as a percentage of prime costs)