Acessibilidade / Reportar erro

On the Efficiency of Alternatives in Process Plans

Abstract

Process planning is a very important industrial activity, since it determines how a part or a product is manufactured. Process planning decisions include machine selection, tool selection, and cutting conditions determination, and thus it is a complex activity. In the presence of unstable demand, flexibility has become a very important characteristic of today's successful industries, for which Flexible Manufacturing Systems (FMSs) have been proposed as a solution. However, we believe that FMS control software is not flexible enough to adapt to different manufacturing system conditions aiming at increasing the system's efficiency. One means to overcome this limitation is to include pre-planned alternatives in the process plan; then planning decisions are made by the control system in real time to select the most appropriate alternative according to the conditions of the shop floor. Some of the advantages of this approach reported in the literature are the reduction of the number of tool setups, and the selection of a replacement machine for executing an operation. To verify whether the presence of alternatives in process plans actually increases the efficiency of the manufacturing system, an investigation was carried out using simulation and design of experiments techniques for alternative plans on a single machine. The proposed methodology and the results are discussed within this paper.

Process planning; manufacturing features; design of experiments; flexible manufacturing systems; simulation


On the Efficiency of Alternatives in Process Plans

João Carlos Espíndola Ferreira

Universidade Federal de Santa Catarina

Departamento de Engenharia Mecânica

GRIMA/GRUCON, Caixa Postal 476

88040-900 Florianópolis, SC. Brazil

jcf@grucon.ufsc.br

Richard A. Wysk

Pennsylvania State University

Department of Industrial and Manufacturing Engineering

Leonhard Building

University Park, PA – 16802, U.S.A.

rwysk@psu.edu

Process planning is a very important industrial activity, since it determines how a part or a product is manufactured. Process planning decisions include machine selection, tool selection, and cutting conditions determination, and thus it is a complex activity. In the presence of unstable demand, flexibility has become a very important characteristic of today's successful industries, for which Flexible Manufacturing Systems (FMSs) have been proposed as a solution. However, we believe that FMS control software is not flexible enough to adapt to different manufacturing system conditions aiming at increasing the system's efficiency. One means to overcome this limitation is to include pre-planned alternatives in the process plan; then planning decisions are made by the control system in real time to select the most appropriate alternative according to the conditions of the shop floor. Some of the advantages of this approach reported in the literature are the reduction of the number of tool setups, and the selection of a replacement machine for executing an operation. To verify whether the presence of alternatives in process plans actually increases the efficiency of the manufacturing system, an investigation was carried out using simulation and design of experiments techniques for alternative plans on a single machine. The proposed methodology and the results are discussed within this paper.

Keywords: Process planning, manufacturing features, design of experiments, flexible manufacturing systems, simulation

Introduction

Process plans contain the information necessary for the manufacture of a part (or batch), which typically includes the operations, their sequence, necessary machines, cutting tools, and cutting conditions. Traditionally, process plans have not contained alternative operations (these plans without alternatives are called 'Linear Plans'), and the machine operator or control system software follows these instructions as they have been prepared. In a dynamic factory environment, situations may occur which prevent the process plan instructions from being followed strictly, and some of these situations are as follows:

· Tooling may be unavailable and/or a machine may need repair. Under these circumstances, some time needs to be devoted by the process planner or machine operator to realise a 'replacement operation(s)', which is a difficult activity, and that has to be done under pressure. In the case of an FMS, the control software is not usually flexible enough to change the process plan in real time, and consequently such problems will affect production significantly, resulting in longer customer delivery times.

· The operator decides to use a different operation than the one in the plan, since according to his/her experience another operation may be more efficient. However, one cannot be certain that the 'new' operation will be better than the other.

· Sometimes an operation in the linear plan requires a tool that needs to be setup in the machine carousel, but another operation could reduce the manufacturing time of the batch by simply using one or more tools already setup at the machine, thus reducing the tool setup time.

A possible solution to the above problems is the introduction of pre-planned alternatives in the process plans, which would:

(a) Allow the process planner, the machine operator or the control system software of an FMS to search for the best solution to a disruption problem on the shop floor, since alternatives are already available;

(b) Allow the decision to be made taking into consideration the context of the next operation to be performed, such as the use of a cutting tool that was already setup at the machine in order to reduce the manufacturing time of the batch.

Previous work on the use of alternatives in process plans reported that alternatives lead to benefits to the manufacturing system (Wilhelm and Shin, 1985; Kruth and Detand, 1992; Xirouchakis, et al., 1998). According to these papers, alternatives can be used successfully:

(a) to solve in a short amount of time the problem of disruptions in the shop floor, such as machine breakdowns, machine overloads and rush orders;

(b) to reduce in-process inventory;

(c) to increase equipment utilisation.

In a previous paper, the authors showed that even in the case of ideal conditions, such as absence of machine breakdown and overload and infinite tool life (Ferreira and Wysk, 2001), the presence of alternatives leads to benefits such as the reduction of the total manufacturing time of the parts.

In order to show further the need for including pre-planned alternatives in process plans, a procedure combining simulation and experimental design was developed, in which different batches of parts are produced on a machining centre It is assumed that there are no machine breakdown or tool failure. Of course in the case of more realistic conditions (i.e. more machines, parts, and tool regrinds and replacements), the influence of alternative process plans will be even more significant.

An important factor in this scenario is the influence of the cutting tools in the manufacture of different part batches. In this work it is assumed that the cutting tools needed to perform the operations may be located either at the machine tool carousel or at the tool room. This is done because a cutting tool initially located at the tool room may give rise to a greater throughput than a tool already setup at the machine.

The influence of factors related to the characteristics of the parts being manufactured on throughput is investigated. These factors include part variety and complexity. These factors were included in the procedure through the combination of manufacturing features.

No previous work has been found in the literature that applies design of experiments and simulation methods to investigate the scenario proposed in this paper.

A detailed description of this procedure and the results of this investigation are presented in the following sections.

The Proposed Simulation Experiment

To illustrate the effects of using alternative process plans, an experiment was designed in which various part and machine specifics could be analysed. A standard set of features was created and process plans for each feature were devised. These process plans contained at least two alternatives to produce each feature. Parts were then randomly generated with some of these features. The purpose was to generate some 'random' parts that could be used to test the effects of using alternative process plans. Rather than test alternative plans for various types of manufacturing systems, we decided to limit our investigation to a single machine, feeling that if significant gains are made at a single machine then it follows that even greater effects will be realised for systems. The procedure used for this research is as follows:

(a) Propose a single flexible machine, which could be used to simulate various process planning and control alternatives.

(b) Choose an adequate response variable, which in this paper is considered throughput (i.e. the number of parts produced in 100 hours).

(c) Identify some control factors that could be used to construct an experimental design, and ascertain their corresponding levels, which are shown in Fig. 1. A detailed description of each of these factors is given later in this paper.

(d) Conduct the experiments by maintaining each factor at a certain level, and then calculate the throughput. Control factors are then changed one at a time and the throughput is again determined via simulation. Perform these simulations for all factor level combinations (full-factorial design). It can be noticed in Fig. 1 that there are three factors each with two levels, and four factors each with three levels, which results in a total of 23x34 = 648 trials. Also, in order to take into account different possible results under the same levels, five replications are performed in each trial.

(e) Perform an analysis of the results obtained, which consists of verifying (i) how each factor influences the throughput; (ii) how the types of process plans (i.e. linear or with alternatives) influence the throughput for different levels of the other control factors.


A description of each of these steps is given below.

The Physical System

The system considered consists of one CNC machining centre with a certain tool capacity in its carousel. Three different orders for the manufacture of one batch of parts each are sent to the machine. It is assumed that when the tools for the manufacture of the first batch are setup in the machine, there is still no information about the necessary tools for the manufacture of the second batch (and the same situation occurs for the third batch). It is also assumed that the manufacturing conditions are ideal, i.e. no machine breakdown, and infinite tool life.

The Throughput and Its Model

The model for determining the throughput for a batch in 100 hours is as follows:

where:

TPb = throughput (number of manufactured parts in 100 hours) of batch 'b'

nb = number of parts in the batch (i.e. batch quantity)

Thsu = time required for the hard setup of a new batch, which includes activities such as fixture qualification and the download of the NC program (in minutes)

ttsui = setup time for each cutting tool 'i', which may include for example its installation into a collet and into a machine tool changer (in minutes)

ntk = number of tools in alternative 'k' necessary to manufacture the batch

nsk = number of necessary tools in alternative 'k' that are already setup at the machine

tremi = time to take cutting tool 'i' out of the carousel (in minutes)

nrk = number of necessary tool replacements in alternative 'k' for each batch

tmi = total machining time with cutting tool 'i' (in minutes)

ttc = automatic tool change time (in minutes)

tLU = sum of the loading and unloading times of each part (in minutes)

For the purpose of this paper, the constant term 'Thsu' was considered equal to zero, and thus a new term (called 'TP0b') becomes the response variable, and it is also called throughput.

Identification of the Factors

The factors that are considered as influencing the throughput of the manufacturing system are shown in Fig. 1, together with their levels.

The parts to be manufactured are considered as being composed of features, which play a very important role in process planning, and consequently in the proposed procedure. Before explaining the specific factors, a description of how the features are considered in this method is given.

The following features are considered in this paper: holes, slots, shoulders, and rectangular pockets. The geometry information of the features is assigned initially to them, and a total of 27 features were considered in this implementation, as shown in Table 1.

With regard to the manufacturing information related to each feature, such as the possible machining operations and corresponding cutting tools and cutting conditions, this is also assigned a priori to each of the features. For the manufacture of each of these features, at least two alternatives are created. Examples of these alternatives are given in Figs. 2 to 5. The alternatives are represented as AND/OR graphs (Wysk et al., 1995). This representation may be applied both to features and to parts.





An AND ('&') node means that all paths originating from it must be traversed, but the sequence could be arbitrary. On the other hand, an OR ('|') node means that only one path stemming from this node must be chosen. The detailed information shown in each node in the graph (i.e. operation, tool type, tool diameter and machining time) is important for a better understanding of the contents of the alternatives.

A description of the alternatives for the manufacture of each of the above features is given below.

(a) Holes: Initially the hole is twist drilled, and since it is assumed that the holes have a high accuracy, a boring tool or a reamer is needed to finish the hole.

(b) Shoulders: End milling tools with diameters larger than the shoulder width were considered for both alternatives, with the first alternative corresponding to the larger milling tool. This was done so that the shoulder is machined in only one radial pass; however, milling tools with diameters smaller than the shoulder width could also have been considered.

(c) Slots: The first alternative consists of a milling tool with a diameter equal to the slot width; but if this tool is not available, an end milling tool with a diameter a little smaller than the slot width is selected. The second alternative consists of a milling tool whose diameter is smaller than the previous end milling tool.

(d) Rectangular Pockets: The rectangular pockets are machined initially using an end milling tool that plunges in the axial direction, and then removes material in the radial direction. If there is any material left, another milling tool with a diameter equal to the fillet diameter is used to machine the remaining material. Three possible alternatives are considered for the manufacture of a pocket:

1) Initially a large tool, with a diameter equal to the pocket width, machines as much material as it can, then a milling tool with diameter equal to the fillet diameter is used to machine the remaining material. If a milling tool with diameter equal to the pocket width is not available, a milling tool with a smaller diameter is used.

2) Initially a tool with a diameter smaller than that of the first alternative, but larger than the fillet diameter, is used for the machining of the pocket, then a milling tool with a diameter equal to the fillet diameter is used to machine the remaining material.

3) A milling tool with a diameter equal to the fillet diameter is used to machine the pocket completely.

A detailed description of the equations used for the calculation of the machining time for the different features is given in (Ferreira and Wysk, 2001).

In this implementation all the cutting tools used for machining the parts are carbide tools, and their cutting conditions are obtained from cutting tool manufacturer's catalogues and machining handbooks (Metcut, 1980).

The features are then used to form the parts that will be machined in the manufacturing system (see Fig. 6). The means by which the parts are formed depends on three of the factors shown in Fig. 1: number of feature types per part, part variety, and number of feature duplications. A discussion of the factors used in the experiment is given in the following section.


Number of Feature Types per Part

The factor called 'number of features types per part' ('nfpp') was used in the experimentation. This factor corresponds to the number of different features in each part. In this paper, it is assumed that the number of different features in each of the three parts is the same. Also, each instantiated feature is considered as being of a different type than any other, even if these features are for instance slots. Three levels corresponding to the number of feature types per part were used, as shown in Table 2.

A random number generator was used to obtain the number of feature types per part, based on the factor's level. For example, if the level is medium, a possible number of feature types is 5, which means that each part will have 5 feature types each. The decision about what features to be assigned to a part depends on the factor 'Part Variety'.

Part Variety

Again we used three levels for part variety: low, medium and high. In the case of a high part variety, all features in a part must be different than the features in the other two parts. In the case of a medium level, different parts would share some of the features, and for a low level an even higher number of the same features would belong to the three parts. In order to do that, the following procedure was devised:

(a) Create a list, called 'assignment list', which contains all the features that will be assigned to the three parts. The number of features in this list is calculated as shown in Table 3.

Where

nparts = total number of batches of parts to be manufactured (= 3)

nfpp = number of feature types per part

nal = number of features in the assignment list

The notation |u|– means that 'u' (u Î Â) is rounded down to the nearest integer.

(b) The assignment list is filled with feature codes (shown in Table 1), and this is performed using a random number generator. For instance, in Example 1 shown in Table 3, where nfpp = 3, the following assignment lists could have been generated for the low and high levels respectively:

Low ® 4, 18, 11, 16 (nal = 4)

High ® 16, 27, 12, 26, 7, 24, 2, 19, 1 (nal = 9)

(c) Finally, each feature in the list is assigned to each part, without duplication. This procedure follows the steps below:

I. Divide the number of features in the assignment list (nal) by the number of batches to be manufactured (= 3). Round this number down to the nearest integer, which is called 'naf'. For instance, in example 1 in Table 3, the value of naf for the medium level is calculated as follows:

II. Starting from the first element in the list, assign naf features from the list to Part 1; then the following naf features to Part 2; and finally the next naf features to Part 3. This is shown in Fig. 7(a).

III. The remaining features (if any) are assigned one by one to Part 1, Part 2 and Part 3 respectively, until the assignment list is exhausted (see Fig. 7(b)).

IV. If the number of assigned features in each part has not reached nfpp, return to the first feature in the list, and keep assigning features one by one to the parts, without feature duplication on each part, until the total number of assigned features in each part is equal to nfpp.



Some results of the application of this procedure for different situations are given in Fig. 8.


Number of Feature Duplications

The factor referred to as 'number of feature duplications' corresponds to the number of copies of each feature in each part. This factor was also examined in the experiment. For each type of feature in each part, a random number generator is used to obtain the number of duplications of each feature. The levels of this factor are shown in Table 4. It should be noticed that for each level the number of duplications may be different for each feature. For example, in the case of a low level for this factor, assuming that a part is composed of hole 3, slot 2 and pocket 4 (see Table 1), the duplication number for each of these features on this level may be 4, 1 and 3, respectively.

Batch Quantity

Three different batch quantities were also investigated: low (1-10), medium (11-100) and high (101-500). In the case of a low batch quantity, for each of the five replications a random number generator is used to generate the batch quantity, which lies between 1 and 10. The same method is applied to the medium and high levels.

Batch Sequence

Batch sequence was investigated using two levels: fixed or random. A random batch sequence consists of the part batches being manufactured in a different sequence in each replication (e.g. first replication ® 2-1-3; second replication ® 3-2-1; third replication ® 1-2-3, etc.). A random number generator is used for obtaining each of the random sequences. In the case of a fixed sequence, this sequence is also obtained through a random number generator, but this sequence is generated only once for the five replications.

Carousel Capacity

Another factor used in the experimental model was that of carousel capacity. This was considered important because it significantly affects the number of tools that can be used for the manufacture of a batch. Two ways in which these limitations may occur are:

(a) a certain batch may require more tools than the carousel capacity. In this case, the manufacture would not be feasible (at least with only one complete part program download).

(b) Even when the tools necessary to machine each batch do not exceed the carousel capacity, the tools necessary to machine two successive batches may exceed the carousel capacity, and in this case it would be necessary to replace some of the tools that are in the carousel. The time for tool replacement is included in Eq. (1).

In this paper this factor is considered as having two levels, which are: 12 tools and 24 tools.

Process Plan

The process plan factor has two levels: linear or AND/OR plan. These different levels are described below.

(a) Linear Plan: in this level, for each of the features that are assigned to a part, the linear plan is composed of the first alternatives for the manufacture of each feature.

(b) AND/OR Plan: in this level, all the alternatives of each feature are taken into consideration, and they are combined with all the alternatives of every other feature. The throughput is calculated for all possible combinations of alternatives, and the minimum throughput and its corresponding cutting tools are obtained. The maximum number of possible process plans corresponds to the following levels:

Part Variety = High

Number of feature types per part = High, and within this level the largest number is 9.

Rectangular pockets have the largest number of alternatives (i.e. three), and so the largest amount of possible process plans will occur when all six pockets are in one single part, and of course the other three features would have two alternatives at most. In this case, the total number of alternative paths to be analysed would be: 36x23 = 5832.

The Experiments

In order to compare the influence of the types of process plans on the throughput, the experiments were performed in the following manner: for the same levels of the factors part variety, batch quantity, number of feature types per part and number of feature duplications, the other three factors, i.e. process plan type, batch sequence and carousel capacity had their levels altered. By doing this, the parts and the batch quantity are kept the same, which reduces the possible influence of variations on the results.

Some of the data used in the calculations are shown below:

  • Automatic tool change time (ttc) = 0.2 min

  • Setup time for each tool (ttsu) = 10.0 min

  • Time to take a cutting tool out of the carousel (trem) = 2.0 min

  • Loading time (tL) = Unloading time (tU) = 1.0 min ® tLU = 2.0 min

  • Material of each part: 4140 steel (HB 230)

Examples of results obtained for some trials under different levels are shown in Table 5.

Analysis of the Results

In planning the experiment, several factors were used, which we knew would affect throughput. For instance, as the number of features types on a part increases, the related processing time will also increase and the throughput will decrease. This is an expected result that was verified. This factor was considered important because it may also affect whether an alternative plan becomes advantageous. Feature duplication should increase the machining time, leading to a lower throughput. As batch quantity increases the contribution of the setup time should be lower, and the related throughput increased. With this in mind, the experiment was run. The amount of data obtained was significantly large, and the analysis of these data is subdivided into two parts:

(a) Analysis of the influence of all control factors on the throughput;

(b) Analysis of the influence of the specific types of process plans on the throughput.

A description of these analyses is given in the following sections.

Analysis of the Influence of All Factors on the Throughput

In order to verify the influence of the factors on the throughput, the statistical technique called ANOVA (Analysis of Variance) was applied (Ross, 1995). A one-way ANOVA was performed using the commercial software Minitab® (Minitab, 1996).

The one-way ANOVA was performed for all control factors, in order to verify their effect on the throughput, and the results are shown in Table 6. The term 'F', also called 'F statistic', is the ratio of total variance to the error variance, whereas the term 'P' is the percent contribution, which can be directly related to the confidence interval on which the values of a statistical parameter are likely to fall.

Among these factors, the null hypothesis (same means) was rejected for the factors no. of feature types per part, no. of feature duplications, batch quantity, and part variety, which means that with 95% confidence they have a direct effect on the absolute value of throughput. These four factors are responsible for what is to be manufactured (i.e. the parts and their batches), and thus their influence on the throughput was expected to be very high. Henceforth in this paper these four factors will be called 'W-factors' ('W' for 'what').

On the other hand, for the factors process plan, carousel capacity and batch sequence, the null hypothesis was not rejected, and thus with 95% confidence it is inferred that the type of process plan, the batch sequence and the carousel capacity do not have a significant effect on the absolute value of throughput.

A main effects plot for all the factors is shown in Fig. 9, which illustrates visually the degree of influence of these factors on the throughput. Although the process plan factor was considered as not having an effect on throughput, notice in Fig. 9 that there is a small increase in throughput when using an AND/OR plan compared with a linear plan.


Analysis of the Influence of Process Plans on the Throughput

After observing the influence of all the factors on the throughput, it is necessary to perform a specific analysis of the influence of the factors considering the presence or absence of alternatives in the process plans.

Although it was concluded in the previous section that the type of process plan did not influence significantly the throughput compared with the W-factors, when observing the data, the process plan factor still had some effect on the throughput. An example of the influence of alternatives in process plans is given by the calculation of the largest percentage increase in throughput, which was equal to 20.0%, and that took place for the following factor levels:

  • No. of Feature Types per Part = High

  • No. of Feature Duplications = Low

  • Batch Quantity = Low

  • Part Variety = High

  • Batch Sequence = Fixed

  • Carousel Capacity = 12 tools

This corresponds to an increase of 47 units over the 235 units produced with a linear process plan.

In almost all the situations, the presence of alternatives increased the throughput, and this is given by the positive inclination of the lines shown in Fig. 10.


An interesting situation occurred when the control factors were set as follows:

  • No. of Feature Types per Part = Low

  • No. of Feature Duplications = Medium

  • Batch Quantity = Low

  • Part Variety = Low

  • Both sequence types, but for the random sequence the performance of the linear plan was even better than the AND/OR plan, compared with the fixed sequence.

In that situation, which is shown in Table 7, the throughput was higher for the linear plan compared with the one for the plan with alternatives. This occurred because, since the cutting tools are setup considering one batch at a time, the first batch in the case of the linear plan required a larger amount of tools than the plan with alternatives. However, for the second batch, the linear plan ends up using a tool that was already setup at the machine, whereas the best plan among the alternatives needs a tool that was not already setup. And the setup time for this tool ends up contributing to the decrease in the throughput of those batches. This situation is shown graphically in Fig. 11.


Although this situation occurred for the above factor levels, it could have occurred for other levels, provided the batch quantity factor remained in the low level, since for the other levels of the batch quantity (i.e. medium and high), the portion corresponding to machining would be greater than the portion related to the tool setup.

A more comprehensive analysis of the influence of the process plan types on the throughput was also performed, where for each of the factors the percentage increase in average throughput for the factor levels was calculated. This was done by calculating the average throughput for level 'i' of one of the four factors 'j', considering the plans with alternatives (referred to as 'mija'), and then calculating the average throughput for the same level 'i' of the factor 'j', but for the linear plans (referred to as 'mijl'). The percentage increase in throughput 'h' is given by:

The values of 'h' were calculated for the different factors, and they are shown in Table 8, and their graphical representation is given in Figs. 12 and 13. The range, which corresponds to the maximum percentage increase in throughput minus the minimum percentage value, is included in Table 8, and the factors have been placed in the Table in descending order of range.



By observing Fig. 12, it can be concluded that, in average, the presence of alternatives is always beneficial, i.e. for all factor levels, there was an increase in throughput, and the minimum average increase was 2.4%.

It can also be noticed that the number of feature types per part is the factor that most affects the average increase in throughput when using AND/OR plans, as it is the factor that influenced most the absolute value of throughput. On the other hand, the number of feature duplications, which was the second most influencing factor on the throughput, is the factor among the W-factors that least influences the average increase in throughput when using planning alternatives. But still on its worst level there is some average increase in throughput.

The larger the number of feature types per part, the larger the average increase in throughput in the presence of alternatives, which means that the more complex a part is, the greater is the manufacture gain when alternatives are present. It should also be noticed that even for simple parts (low level of this factor), a 2.4% average increase was obtained, and thus AND/OR plans are also recommended for these conditions.

The same reasoning applies to part variety, which is the factor that measures the degree of differences between two subsequent batches. For a high level of this factor, the greater is the average throughput when using alternatives compared with linear plans. Of course, in the presence of uncertain demand, which is one of the characteristics of today's production, part variety is usually high, and thus a move towards including alternatives in process plans is recommended.

For a high batch quantity the average increase in throughput was equal to 2.9% when using alternatives, which may not be a significant increase in throughput. But the lower the batch quantity, the greater the average increase in throughput. Since a large majority of today's manufactured products are produced in small batches (nb < 50) (Chang and Wysk, 1985), the presence of alternatives is also recommended under these circumstances.

The larger the number of feature duplications, the lower the average increase in the throughput, but it still is relatively high (i.e. 2.9%).

It can be noticed in Fig. 13 that the variation caused by the batch sequence and the carousel capacity was small, which confirms their small influence on the average increase in throughput when using planning alternatives. But in this figure it can be noticed that for a lower carousel capacity (i.e. 12 tools), which constrains significantly the selection of operations and cutting tools, the presence of alternatives increased the throughput in 0.4% compared with the carousel with a capacity for 24 tools.

Conclusions

In this paper a procedure was presented which coupled simulation and design of experiments techniques in order to investigate the effectiveness of using alternative process plans. The intent was to verify whether the inclusion of pre-planned alternatives in process plans would increase the efficiency of a manufacturing system. The results show that the efficiency is increased with the alternatives, and thus their inclusion is recommended.

The results presented in this paper were obtained for ideal conditions, and for a simple manufacturing system. Even so, the presence of alternatives was shown to increase the manufacturing efficiency. Under more realistic conditions, and for more complex manufacturing systems, it is thought that the presence of alternatives will increase the manufacturing efficiency even more, since they will be used not only for tool setup considerations and uncertain demand, but also for situations which cause shop floor disruptions, such as machine or tool breakdown. In the future we intend to investigate the efficiency of alternatives in process plans under these conditions.

It is considered that, for situations that are more complex (i.e. two or more machines, and the likelihood of machine breakdown and tool failure, the proposed model will have to be enhanced to cater for these new conditions.

Another conclusion is that, based on the observed data, the presence of alternatives increases the efficiency of a manufacturing system under conditions such as more complex parts and low batches, which represent a significant amount of today's manufacturing reality.

Acknowledgement

The first author would like to thank CNPq for the financial support to this project.

Article received: July, 2000. Technical Editor: Átila P. S. Freire

  • Chang, T.C. and Wysk, R.A. 1985, "An Introduction to Process Planning Systems", Prentice-Hall.
  • Ferreira, J.C.E. and Wysk, R.A., 2001, "An Investigation of the Influence of Alternative Process Plans on Equipment Control", Journal of Manufacturing Systems, 19(6), 393-406
  • Kruth, J.P. and Detand, J., 1992, "A CAPP System for Nonlinear Process Plans", Annals of the CIRP, 41(1), 489-492.
  • Metcut Research Associates, 1980, "Machining Data Handbook", Vol. 1, 3rd Edition.
  • Minitab Inc., 1996, "Minitab Reference Manual", Release 11, Windows.
  • Ross, P.J. 1995, "Taguchi Techniques for Quality Engineering", 2nd ed., McGraw-Hill.
  • Wilhelm, W.E. and Shin, H.-M., 1985, "Effectiveness of Alternate Operations in a Flexible Manufacturing System", International Journal of Production Research, 23(1), 65-79.
  • Wysk, R.A., Peters, B.A. and Smith, J.S., 1995, "A Formal Process Planning Schema for Shop Floor Control", Engineering Design and Automation, 1(1), 3-19.
  • Xirouchakis, P., Kiritsis, D. and Persson, J.-G., 1998, "A Petrinet Technique for Process Planning Cost Estimation", Annals of the CIRP, 47(1), 427-430.

Publication Dates

  • Publication in this collection
    19 Aug 2002
  • Date of issue
    2001

History

  • Received
    July 2000
The Brazilian Society of Mechanical Sciences Av. Rio Branco, 124 - 14. Andar, 20040-001 Rio de Janeiro RJ - Brazil, Tel. : (55 21) 2221-0438, Fax.: (55 21) 2509-7128 - Rio de Janeiro - RJ - Brazil
E-mail: abcm@domain.com.br