Proceedings of the 41st International Conference on Computers & Industrial Engineering

Transcription

1 EVALUATING THE IMPACT OF INTRODUCING RFID TECHNOLOGY IN MANUFACTURING SYSTEMS USING SIMULATION Aly M. Owida 1, Khaled S. El-Kilany 2, and Aziz E. El-Sayed 3 Department of Industrial and Management Engineering, College of Engineering and Technology, Arab Academy for Science, Technology, and Maritime Transport (AASTMT), Alexandria, Egypt. 1 Aly.Owida@staff.aast.edu, 2 kkilany@aast.edu, 3 azizezzat@aast.edu Abstract: Radio frequency identification (RFID) technology has significant impact on product tracking and identification in manufacturing systems. Most of the business cases that implement the RFID technology in their operations have achieved various benefits. RFID technology can reduce the operating errors that affect the efficiency of the operations which results in improving different performance measures such as cycle time, throughput, work-in-process, resources utilization, and average waiting time in queues. In addition, several benefits such as improved items monitoring, lower lead times, and better inventory control can be achieved by introducing RFID technology. Recent developments in RFID technology and other supporting technologies have created opportunities for real-time traceability and better visibility in shop floor operations. This work investigates the effectiveness of introducing RFID technology in tracking and identification processes for products flow in a job shop manufacturing facility. A leading furniture manufacturer in Egypt has been selected as a case study. The manufacturer produces a large number of customized furniture products. Errors in tracking and identification usually occur due to the large number of products present on the shop floor. Introduction of radio frequency identification technology at different stages of manufacturing is proposed to overcome these errors. Different simulation models have been developed for the post-assembly processes in the facility. These models have been developed with an intent to capture all the features that characterize a real furniture manufacturing facility. Simulation is used to assess the impact of introducing the RFID technology on a number of performance measures. Analysis and comparison of simulation results for the base and proposed models show that RFID implementation can improve the overall performance of the facility. Keywords: Radio frequency identification, job shop, modeling and simulation 1. INTRODUCTION The competition to achieve high customer service levels at minimal cost has placed a strong emphasis on the control of information and material flows in today s manufacturing and retail environments. Most companies have made substantial investments in innovative systems enabling them to improve the level of automation of their supply chain processes [1]. Radio Frequency Identification (RFID) technology is one of the emerging technologies that are being used by a number of organizations such as manufacturers, retailers, logistics providers, hospitals, and libraries [2, 3]. The swift development of information technology (IT), such as RFID, is one of the decisive factors to improve competitive advantage of enterprises [4]. Personalized products or tailored-made solutions are taking over large shares of the marketplace from mass produced goods and standardized solutions. Therefore, products tracking and identification becomes a very important issue in manufacturing and logistics. RFID has emerged as part of a new form of interorganizational system that aims to improve the efficiency of the tracking and identification processes [5]. 211

2 RFID technology was applied in various areas, accompanied with the simulation, such as supply chain [6, 7], manufacturer-retailer supply chain [8, 9], inventory management in supply chain [10, 11], logistics and reverse logistics [12], inventory management of time-sensitive materials on shop floor [13], and hospitals supply chain and asset management [14, 15]. Most of these papers are trying to assess the benefits of adopting the RFID technology whatever the application area. This work presents the development of a simulation model for a furniture manufacturer using the ExtendSim OR v7 simulation environment. The model of the shop floor has been developed with the purpose of containing all of the features that makes a real furniture manufacturing facility. A number of experiments have been designed and tested using the developed simulation models to assess the impact of introducing the radio frequency identification technology on the performance of this manufacturing facility. The next section presents a brief description of the problem addressed in this work. In section 3, model development is explained. The results and analysis are presented in section 4. Finally, in the last section, conclusions drawn from this work are discussed. 2. PROBLEM DESCRIPTION This work is evaluating the impact of introducing radio frequency identification technology for tracking and identification of items on the shop floor in a job shop manufacturing environment.implementation of the RFID system without predicting its effect on the operational performance of the system can be very disruptive to its operation.measures like throughput, cycle time, work-in-process, and resources utilization are among the measures selected to evaluate the impact of such implementation. An aluminium tag is produced and attached to each item, after its assembly for identification purposes. The aluminium tag is made by engraving certain numbers on a blank aluminium sheet. This tag is used also in the tracking process in order to meet the delivery due dates. Production department sets the production plan and monitors its progress on daily basis to update the next day plan by tracking items on the shop floor using these aluminium tags. Management is convinced that implementing an RFID system for identification and tracking of about 200 items moving between workstations and departments every day is the best solution. Where, a basic RFID system comprised of tags and readers only is considered for implementation in the near future. However, it is not quite clear how would that affect the overall performance of the system. 3. MODEL DEVELOPMENT A simulation model for the factory has been developed using the ExtendSim OR 7, from Imagine That, Inc.The simulation model has been developed under the following assumptions: The factory produces only 5 products. Travel times between departments and rework of items due to quality checks are ignored as RFID implementation will not affect them. Processing times are in minutes; where, one working week is equivalent to 3,600 minutes. The developed models all have a built-in database that holds all data required for routing the products through the different production stages and the processing times at each stage. This data is read from the database and stored as attributes for each item introduced to the model as shown in Figure

3 Simulation Setup Figure 1: Setting different processing parameters for the five products. The following parameters have been set for the execution of the simulation models: Production run length is set to be ten working years; where ten working years = 10 year 52 weeks/year 6 days/week 10 hours/day 60 min/hr = 1,872,000 minutes. Warm up period is assumed to be one working year which is equivalent to 187,200 minutes. The number of replications has been set as 20 replications for each model or scenario. Scenarios Three models were built and simulated to assess the impact of introducing the RFID technology on a number of performance measures. The first model is the base model that represents the current situation of the manufacturing facility; where aluminium tags are used to track and identify items manually. In the second model, RFID solution was proposed to replace the current identification system by attaching RF tag to each item after its assembly and using portable readers to track and identify items. The final model is basically the same as the proposed model presented earlier; however, a reduced number of resources are used in this model. This would result in a reduction of operating expenses; yet, the performance measures used in this work (introduced in the next section) must be re-evaluated. Performance Measures The performance measures that are evaluated in this study are the output, TH (Throughput), CT (Cycle Time), WIP (Work In Process), resources utilization, and average waiting time in queues. Figure 2 shows the part of the model responsible for calculating these measures. Improvements in these measures should aim at increasing the output, TH, and resources utilization and; on the other hand, it should also aim at decreasing the CT, WIP and average waiting time in queues. 213

4 Figure 2: Calculating the key performance measures. 4. SIMULATION RESULTS AND ANALYSIS Simulation runs of the three models presented earlier are carried out. Table 1 summarizes the performance measures values for the base model and the two proposed scenarios. Table 1: Summary of results of the base model and proposed scenarios. Performance Measure Base Model Proposed Scenario 1 Proposed Scenario 2 Output 15,860 15,840 15,850 Throughput Cycle Time 18,680 15,860 16,650 Work in Process Resources Utilization 28.15% 26.01% 37.75% Average Waiting Time in Queues After examining the three models, it is clear that both of the proposed scenarios are better than the base model except in the output and throughput which are nearly the same in the three models due to the presence of a bottleneck in the shop floor. Figure 3 illustrates the improvement indices calculated for the two proposed scenarios compared to the base model.the first proposed scenario compared to the base model, performed better on cycle time, work in process which have decreased by 15%; however, resources utilization has decreased by 7.5% which could be considered as a good indication because resources should be reduced and hence reduces the cost, while almost the same on output, throughput, and average waiting time in queues. By comparing the second proposed scenario with the base model; cycle time improved by almost 11%, WIP also improved by almost 11%, and resources utilization improved by 34%. However, waiting times in queues increased and resulted in poorer performance by almost 3%. Outputs and throughput remained almost unchanged. To conclude, there is a trade-off between utilization of resources and the remaining measures, which are cycle time, WIP level, and average waiting time in queues. 214

5 Improvement Index 40% 35% 30% 25% 20% 15% 10% 5% 0% -5% -10% -0.13% +ve percentage --> Proposed is better -ve percentage --> Base is better -0.06% -0.16% -0.13% 15.10% 10.87% 15.51% 11.28% -7.59% 34.13% Output Throughput Cycle Time WIP Resources Utilization 0.55% -2.98% Waiting Time in Queues Figure 3: Calculated improvement indices for the performance measures of the proposed scenarios. 5. CONCLUSIONS This work has investigated the effectiveness of introducing the radio frequency identification technology for products tracking and identification in a furniture manufacturing shop floor that experiences problems in tracking and identifying its items inside the facility. Three models have been developed, which are: base, first proposed, and second proposed. Runs and experiments with these models have been conducted to compare and analyze their results for the selected performance measures. Scenario 1 model reflects the implementation of a basic RFID system for tracking and identification of products. Compared to the base model, proposed scenario 1 performed better on cycle time and work in process; however, resources utilization has decreased which could be considered as a good indication because resources should be reduced and hence reduces the cost, while almost the same on output, throughput, and average waiting time in queues. Implementation of the RFID system cannot improve the manufacturing processes (throughput cannot be improved); however, it can help in providing a better flow of materials through the shop floor as indicated by the improvement of cycle time and WIP levels. Scenario 2 model is like scenario 1; however, this model is run with fewer numbers of resources. Compared to the base model; CT, WIP, and resources utilization are improved. However, waiting times in queues increased and resulted in poorer performance. Outputs and throughput remained almost unchanged. Compared to scenario 1 model, resource utilization is much better; however, reducing the resources resulted in higher cycle times, higher WIP levels, and higher waiting time in queues. Outputs and throughput remained almost the same. Although this model has been run with fewer resources; yet, CT and WIP are still performing better than the base model. In addition, the model resulted in better performance relative to resource utilization. Further research recommended to be undertaken in other areas such as assessing the applicability of this model on other furniture manufacturing systems or larger furniture firms and comparing different RFID solutions in terms of their impact on the system performance and on the overall cost. ACKNOWLEDGEMENTS This work has been awarded the ExtendSim Research Grant to use ExtendSim TM OR v7 simulation environment for developing the simulation models presented in this paper. 215

6 REFERENCES [1] E. Sahin and Y. Dallery, "Assessing the Impact of Inventory Inaccuracies within a Newsvendor Framework," European Journal of Operational Research, vol. 197, [2] I. Lee and Byoung-ChanLee, "An Investment Evaluation of Supply Chain RFID Technologies: A Normative Modeling Approach," International Journal of Production Economics, vol. 125, [3] J. Worapot, Y. Li, and A.-I. Somjit, "Business Process Analysis and Simulation for the RFID and EPCglobal Network Enabled Supply Chain: A Proof-of-Concept Approach," Journal of Network and Computer Applications, [4] L. C. Lin, "An Integrated Framework for the Development of Radio Frequency Identification Technology in the Logistics and Supply Chain Management," Computers & Industrial Engineering, vol. 57, [5] E. W. T. Ngaia, K. K. L. Moonb, F. J. Rigginsc, and C. Y. Yi, "RFID Research: An Academic Literature Review ( ) and Future Research Directions," International Journal of Production Economics, vol. 112, [6] K. Sari, "Exploring the Impacts of Radio Frequency Identification (RFID) Technology on Supply Chain Performance," European Journal of Operational Research, [7] Y. Z. Mehrjerdi, "RFID-enabled Supply Chain Systems with Computer Simulation," Assembly Automation, vol. 29, [8] Y. M. Lee, F. Cheng, and Y. T. Leung, "Exploring the Impact of RFID on Supply Chain Dynamics," in Winter Simulation Conference, [9] Y. T. Leung, F. Cheng, Y. M. Lee, and J. J. Hennessy, "A Tool Set for Exploring the Value of RFID in a Supply Chain," in Trends in Supply Chain Design and Management, [10] Y. M. Lee, F. Cheng, and Y. T. Leung, "A Quantitative View on How RFID Can Improve Inventory Management in a Supply Chain," International Journal of Logistics, vol. 12, [11] A. Sarac, N. Absi, and S. Dauzere-Peres, "A Simulation Approach to Evaluate the Impact of Introducing RFID Technologies in a Three-Level Supply Chain," in Winter Simulation Conference, [12] C. K. M. Lee and T. M. Chan, "Development of RFID-based Reverse Logistics System," Expert Systems with Applications, vol. 36, [13] C. Saygin, "Adaptive Inventory Management Using RFID Data," International Journal of Advanced Manufacturing Technology, vol. 32, [14] M. Amini, R. F. Otondo, B. D. Janz, and M. G. Pitts, "Simulation Modeling and Analysis: A Collateral Application and Exposition of RFID Technology," Production and Operations Management, vol. 16, [15] K. Efe, V. Raghavan, and S. Choubey, "Simulation Modeling Movable Hospital Assets Managed with RFID Sensors," in Winter Simulation Conference,