Indian Journal of Fibre & Textile Research Vol.38, Septeber 2013, pp 259-264 An iproveent in calculation ethod for apparel assebly line balancing F Khosravi 1, a, A H Sadeghi 1 & F Jolai 2 1 Textile Departent, Airkabir University of Technology, Tehran, Iran 2 Departent of Industrial Engineering, College of Engineering, University of Tehran, Tehran, Iran Received 1 January 2012; revised received and accepted 13 Septeber 2012 A calculation ethod has been developed to deterine the nuber of workstations and cycle tie, considering one achine type at each workstation, norally preferred in apparel industries. The findings are copared with the earlier findings with respect to criterions such as the nuber of workers, the level of workers' skill, the nuber of used achine and the nuber of worker's oveent between achines. The results show iproveent in line balancing when taking into consideration the achine type feature. Hence, a odel is proposed for apparel assebly line balancing to take into consideration the achine type assigned to each workstation so that it has a specific advantage for apparel assebly lines. Keywords: Assebly line balancing, Apparel industry, Calculation ethod, Sewing line 1 Introduction Sewing line involves a set of workstations wherein a specific task is processed according to a specific sequence. Usually, one or ore tasks are assigned to a workstation, and several consecutive workstations for as a sewing line. Therefore, the ai of assebly line balancing in sewing lines is to assign tasks to workstations so that the achines of the workstation can perfor the assigned tasks with a balanced loading 1. In apparel industry, the ost iportant and critical section is sewing. This section usually involves a great nuber of operations each having a short tie to perfor. Contrary to ost equipents in other assebly lines, the sewing achines are stationary and heavy. In any sewing assebly lines one operator is assigned to each workstation for perforing several different operations. If these operations require to be perfored by different achines, then the assigned operator has to ove aong the achines. Since this takes a lot of tie, it is better that all operations assigned to a workstation are perfored by one achine type. In the other words, it is reasonable to consider each achine as a workstation in such assebly lines. Most of the general assebly line balancing odels have not defined a liitation on the type of a Corresponding author. E-ail: fereshteh87@aut.ac.ir tools or achines in each workstation which hinders the application of these odels in apparel assebly line balancing. onetheless efforts have been ade to provide solutions for this kind of lines by different researchers. Recently, assebly line balancing with general resource-constrained has been considered by Coroinas et al. 2. Also, Chen et al. 1 have presented a grouping genetic algorith for assebly line balancing proble of sewing lines in which each task ust be assigned to one of the specified set of alternative achines and one achine ust be assigned to each workstation. Shuon et al. 3 also have presented a layout odel for assebly line balancing and copared the new concept and traditional syste in a sewing line. Use of calculating ethods in the area of assebly line balancing is so sufficient because of its siplicity, availability and practicality. akajia et al. 4 have proposed a siple calculating ethod to deterine the nuber of workstations and cycle tie. This calculating ethod is siple and intelligible, but has not been attended to the said feature of apparel assebly lines, i.e. achine type. This also has soe disadvantages such as excess tie required for oving operator between achines, necessity of ulti-skilled operators, requiring ore nuber of achines and thereupon further space. However, besides these disadvantages, developing a new ethod which could consider the achinery feature
260 IDIA J. FIBRE TEXT. RES., SEPTEMBER 2013 in balancing apparel assebly line is strongly required. Therefore, the present study was aied at developing a ethod for deterining the nuber of workstations and cycle tie, considering one achine type at each workstation. These calculating ethods are then applied to a sewing process in practical clothes (jeans) aking line, to exaine the utility and usefulness of these ethods. 2 Materials and Methods 2.1 Proposed Calculation Methods Based on the objective used for perforance evaluation, the assebly line balancing proble is classified into three groups, naely (i) deterining the iniu nuber of workstations for a given cycle tie, (ii) iniizing the cycle tie for a given nuber of workstations and (iii) axiizing the line efficiency and deciding the nuber of workstations and cycle tie. The study includes ethods for calculating the optiu nuber of workstations and also the optiu value of cycle tie. 2.1.1 Calculating the Optial uber of Workstations Optial nuber of workstations is deterined so that in addition to satisfying the constraints such as technical sequence of tasks and cycle tie, the restriction of allocating only one type of achine to each workstation is also fulfilled. Consideration of this restriction can solve different probles. For exaple, probles of allocating specific tasks to specific workstations, task grouping according to task skill levels or allocating specific tasks to only a left-of-line or right-of-line station in a two sided assebly lines 5. Assuing that each task can be assigned to each workstation, the iniu nuber of workstations ( in ) is obtained by following equation: in n = 1 ti i 1 (1) = + τ where t i is the processing tie for task i ; and τ, the tie available in each workstation called cycle tie. But the above assuption is possible only when in each workstation the required achines perfor all tasks assigned to that workstation. This requires, on the one hand, uch traffic of the operator between achines in a workstation and on the other hand, availability of ultiple achines in each workstation. In addition, the operator ust be able to handle all these achines. ow, relaxing the above assuption i.e. considering the achine type, the iniu nuber of workstations required for each type of achine can be calculated separately. For this purpose, one should divide total tie of tasks (require by the siilar achine to perfor) by cycle tie and then add one to integer part of this product. Hence, first we calculate the total tie of tasks (T j ) using the following equation: K T j = tij j = 1,2,..., (2) i = 1 where t ij is the processing tie for task i by achine j; j = 1,2,..., ; and K, the nuber of tasks. Then the iniu nuber of workstations for each achine is obtained by dividing the products of Eq. (2) by cycle tie and adding one to the integer part of the product, which results in following equation: Tj j = 1, 2,..., (3) j = + 1 τ ow, the iniu nuber of workstations is achieved by the su of j s, as shown below: in = j (4) j= 1 Maxiu nuber of workstations is equal to the nuber of tasks, theoretically, according to following equation: ax = K...(5) The practical iniu nuber of workstations is given by the following equation 4 : τ feas = ( i) ti >...(6) 2 where ( i ) is the nuber of i. The optiu nuber 4 of workstations (*) is in the range of * ax ( in, feas ) ax...(7)
KHOSRAVI et al.: IMPROVEMET I CALCULATIO METHOD FOR APPAREL ASSEMBLY LIE BALACIG 261 So, the best result is obtained if ( in ) * ax, feas =...(8) 2.1.2 Calculating Optial Cycle Tie Miniizing the cycle tie axiizes the production rate or equivalently iniizes the total idle tie for the given nuber of workstations. Assebly of each product unit requires perforing a nuber of tasks with fixed operation ties. Precedence constraints deterine the sequence of tasks processing. In our proble, the achine type on which each task ust be copleted is given. The production period and the quantity of production for each order ust also be specified. Knowing this condition, we balance the line with the ai of iniizing the idle tie with the given nuber of workstations. Therefore, we should iniize the cycle tie for which the axiu total tie of tasks in the workstation is a lower bound. For iniizing the total idle tie ( ), the total sewing tie ust be equal to the product of workstation nuber and the cycle tie ( = 0). But as this case is rare, we try to iniize. So, firstly the rough cycle tie was calculated using the following relationship 4 : τ theo = H Q...(9) where H is the production period; and Q, the quantity of production for each order. Secondly, using the iniu nuber of workstations obtained fro Eqs (3) and (4), the perfectly balanced tie is given by the following equation: T τ = j j= 1...(10) j by akajia et al. 4 and it is regarded as an extension of their ethod. 3.1 Calculation of Workstation and Cycle Tie A case study considering a jean anufacturer is used to deonstrate the ethods to calculate the nuber of workstations and cycle tie. The real sewing syste has been used by akajia et al. 4. This sewing line produces the workan jeans using conditions as shown in Table 1 and Fig. 1. In Fig. 1, the nubers show the process sequence and are equivalent to the nubers shown in Table 1. 3.1.1 Optial uber of Workstations Firstly, theoretical axiu nuber of workstations in this exaple is equal to 15. ax = 15 Secondly, Table 2 shows the theoretical iniu nuber of workstations, as calculated fro Eqs (2)-(4), for each achine type and total iniu of workstations. Also, the practical iniu nuber of workstations is calculated using the following equation: { ( ) i } feas = i t > 28.5 = 5 The range of the optiu nuber of workstations is * 9 15. Consequently, the nuber of workstations is iniu i.e. 9. In Fig. 2, technical sequence of process is shown. ow, the ain proble is how to deterine the cycle If τ theo τ, the calculation is repeated to iniize idle tie. 3 Results and Discussion This study proposes a calculating ethod to deterine the nuber of workstations and cycle tie, considering one achine type in each workstation. With restricting each workstation to one achine type, the proposed calculating ethod is found to ore practical and ore coincide with real requireent of apparel industry. The proposed calculating ethod is based on the ethod developed Fig. 1 Sewn points of workan jeans 4
262 IDIA J. FIBRE TEXT. RES., SEPTEMBER 2013 Table 1 Sewing conditions 4 Sequence Work nae Sea type Sewn length Processing tie, s 1 Bar-tacks Lock stitch 12 15 2 Butted seas on the groin Double chain stitch (3 needles, 4 yarns) 450 31.5 3 Butted seas under the groin Double chain stitch & sea 1680 23 4 Stitch under the groin Double chain stitch 1680 16.3 5 Bar-tacks Lock stitch 12 10.5 6 Side sea (right and left) Double chain stitch (3 needles, 4 yarns) 1090 2 40.6 7 Labelling Lock stitch 60 14.9 8 Belt sea Double chain stitch 820 36.6 9 Botto sea (left) Double chain stitch 460 26.4 10 Botto sea (right) Double chain stitch 460 16.9 11 Bar-tacks Lock stitch 12 29 12 Belt stitch Lock stitch 790 23.4 13 Button calking - - 7.8 14 Linking of eyelet hole Double chain stitch 54 6 15 Band looping Lock stitch 12 57 Machine nuber (j) Table 2 Processing tie and iniu nuber of achines required for doing each task Machine type Allocated tasks to the achine Processing tie, s Total tie, s uber of required achines a ( j ) 1 Bar-tack achine 1, 5, 11 15, 10.5, 29 54.5 1 2 Double chain stitch 2, 3, 4, 6, 8, 9, 10 31.5, 23, 16.3, 40.6, 191.3 4 achine 36.6, 26.4, 16.9 3 Lock stitch achine 7, 12, 15 14.9, 23.4, 57 95.3 2 4 Button sewing achine 13 7.8 7.8 1 5 Eyelet sewing achine 14 6 6 1 a Equal to the iniu nuber of workstations required for each achine. in 5 = j = 9 j = 1 Fig. 2 Technical sequence of process and representation of start and end tie. If the axiu station tie (P ax ) is enough to overcoe the required production quantity, this tie is set as the cycle tie. In this exaple, the axiu tie is equal to 57s, and the required production (505 clothes/ day) is obtained by Eq (9), as shown below: According to Table 2 and the technical sequence in Fig. 2, first allocating the tasks to the workstations is done for =9. But, with the three constraints that are cycle tie, technical sequence and achine type, based on the concept of Jackson s enueration ethod 6 and considering the achine type. It is not possible to allocate the tasks to 9 workstations. There are two following feasible cobinations with 10 workstations: Both of these cobinations have the idle tie equal to 215.1 s. If the process 1 is selected, naely
KHOSRAVI et al.: IMPROVEMET I CALCULATIO METHOD FOR APPAREL ASSEMBLY LIE BALACIG 263 (1, 5, 11)-(2, 3)-(4, 6)-(9, 10)-(7)-(8)-(12)-(15)-(13)- (14), the relation aong tie, the nuber of operator, the nuber of achine and achine type factors is shown in Fig. 3. Figure 3 shows that station tie does not exceed the cycle tie τ=57 s. Hence, the unbalanced tie is the sae idle tie and the nuber of operators is reduced to 10 fro 11 without any changes in the production quantity. Process efficiency is equal to 354.9 E = = 0.62 62% 10 57 ( ) 3.1.2 Optial Cycle Tie ow, suppose the production is perfored under the sae conditions, with this difference that 11 workers are in the production line and 700 clothes/ day are produced (if a workday includes 8 h). In these conditions, the theoretical cycle tie τ theo for 11 workers/ day is obtained by Eq. 9 as shown below: 8 3600 τ theo = = 41.1s 700 The iniu workstations nuber is obtained by Eqs (2)-(4), as 5 types of achines consisting of 1 Fig. 3 Process foration with 10 workstations and 5 achine type achine 1, 5 achines 2, 3 achines 3, 1 achine 4 and 1 achine 5 are required. So, the nuber of required achine is equal to 12 ( in 5 = j = 12 ). A cobination of eleental work is fored with the workstations nuber whose specifications are given in Table 3. After process foration, the cycle tie is again 41.1 s, and also the production quantity will be the desired aount. The efficiency of this process foration is: 354.9 E = = 0.72 72% 12 41.1 ( ) 3.2 Coputational Results and Coparisons Table 4 shows a coparison between the result of akajia's odel and our proposed ethod. According to the table, the proposed ethod, in coparison with the priary ethod, needs 2 less achines and 3 ore workers. In the proposed ethod all workers are one-skilled because of one achine type in each workstation. But, in the priary ethod, 3 one-skilled workers, 3 two-skilled workers and 1 of three-skilled worker are required. On the other hand, as entioned in the introduction, in case of existing several types of achines in workstations, it is needed to consider the additional tie for oving operators between achines while it is not considered in priary ethod. But in the proposed ethod, this requireent is ipleented by considering one type of achine in each workstation. If the entioned tie for each oveent is an average of 20 s, the cycle tie of akajia's ethod will be equivalent to 77.2 s and it eans 9 workers are required to produce production quantity siilar to its proposed ethod. To suarize, the proposed ethod will need 10 workers and 10 achines, whereas akajia's ethod will need 9 workers and 12 achines. The j= 1 Table 3 Cobination with 12 workstations and iniu cycle tie Workstation nuber 1 2 3 4 5 6 7 8 9 10 11 12 Machine type 1 1 2 2 2 3 2 2 3 3 4 5 Tasks 1,5 11 2 3,4 6 7 8, 9, 10 8, 9, 10 12, 15 12, 15 13 14 Total task ties 25.5 29 31.5 39.3 40.6 14.9 39.95 39.95 40.5 40.5 7.8 6 The fixed cycle tie 41.1 Unbalanced tie 15.6 12.1 9.6 1.8 0.5 26.2 1.15 1.15 0.6 0.6 33.3 35.1 137.7
264 IDIA J. FIBRE TEXT. RES., SEPTEMBER 2013 Table 4 Coparison between proposed ethod and akajia's ethod for calculating nuber of workstations Coparison criteria akajia's ethod Proposed ethod uber of workers 1 10 3 3 Total 7 uber of achines 12 10 uber of worker's oveent between achines: sae coparison can be done when applying the calculating ethod of cycle tie between two entioned ethods. The result of this coparison is shown in Table 5. According to this table, the proposed ethod in coparison to priary ethod needs 1 less achine and 3 ore workers, but all of the are one-skilled, whereas in the priary ethod, 1 four-skilled worker, 1 two-skilled worker and 7 oneskilled workers are required. Here also, considering an approxiately 20 s for necessary oveent between achines, the cycle tie based on akajia's ethod will be equivalent to 77.2 s and it eans 13 workers are required to provide the production quantity siilar to its proposed ethod. To suarize, the proposed ethod will need 12 workers and 12 achines whereas akajia's ethod will need 13 workers and 13 achines. 4 Conclusion In this study, akajia's calculation ethod, which is an accurate, siple and useable ethod for 2 Table 5 Coparison between proposed ethod and akajia's ethod for calculating cycle tie Coparison criteria akajia's ethod Proposed ethod uber of workers 7 12 Total 9 12 uber of achines 13 12 uber of worker's oveent between achines balancing apparel assebly lines, is iproved by considering the achine type. By this iproveent, the proposed odel will be atched with real ters in apparel industries, vain transfer of operators between achines at the workstation will be reoved and we will not need expert and ulti-skilled workers. References 1 Chen J C, Hsaio M H, Chen C C & Sun C J, A Grouping Genetic Algorith for the Assebly Line Balancing Proble of Sewing Lines in Garent Industry (Departent of Industrial and Systes Engineering, Chung Yuan Christian University, Chung Li, Taiwan), 2009. 2 Coroinas A, Ferrer L & Pastor R, Int J Prod Res (in press). 3 Shuon Md. R H, Arif-UZ-Zaan K & Rahan A, Productivity iproveent through line balancing in apparel industries, paper presented at the International Conference on International Engineering and Operations Manageent, Dhaka, Bangladesh, 9-10 January 2010. 4 akajia M, Uchiyaa S & Miura Y, J Text Mach Soc Jpn, Transactions, 33 (4) (1980) 22. 5 Johnson R V, Manage Sci, 29 (11) (1983) 1309. 6 Jackson J R, Manage Sci, 2 (3) (1956) 261. 3