Performance Improvement using Simulation Tool in a Tiles Production Facility

In recent times, the globalization of markets due to the improvement of utilization of resources has significant impact on the manufacturing systems. International competition forced companies to establish efficient and effective production facilities that provide best possible outcome in terms of profitability, throughput and lead time. In this background, every industrial facility is in competition to remain competitive in the market and be flexible for future changes. It is not easy to change the environment of any facility in order to check the results beforehand therefore simulation is a viable option. Simulation tools are used to help analyze performance measures for improvement. This paper presents a case by analyzing a facility in order to improve its production using simulation tool. In this study, simulation model for the production facility of Fort Ceramic Company is developed from its current state and then changes are made to its inter-arrival time. Models for both improved and current systems are developed in the simulation tool. Data analyzed in the simulation tool shows better results that represent a significant improvement in productivity, cycle time and throughput time to optimizethe system.It is also observed that by decreasing the inter arrival time there is increase in the throughput which improves the revenue.

To increase the efficiency of a real system and to optimize the processes, it is essential to study the performance measures. Simulation is one option to study the performance of a real system without bringing any change in the real system. With the help of simulation, it is very likely to observe and understand the behavior of the real system.
Main purpose [1] of simulation is that it helps in decision making which means to achieve desired objectives efficiently. Simulation tools are capable of analyzing the shop floor under different scenarios by just changing values and performance measures. Another research [2] used discrete simulation model of transport system to improve variation number and size of vessels in order to get efficient and effective pattern. Simulation is used [3] for shop floor control system that controls current status and information provided by controller. Another approach [4] used simulation to visualize verification of production plans for integration of Enterprise resource planning and scheduling. In the literature; various roles of simulation are emphasized in the effective planning and scheduling of the manufacturing systems. However, it is also essential to conduct a study to investigate the difficulties in efficient flow of material that generates waste in the production process.
After the research gap identification a case study is performed from a local ceramic industry. It has been observed that in the production department, moving the tiles from one machine to another one is not handled properly and big waste is generated during the transportation. It creates difficulties in the efficient flow of material. Furthermore, parts entering the production department are wasted mostly due to cracks and improper check and balance which affects the profitability of the facility. One of the reasons for cracks is that there is much space between each machine which makes the travel time for tiles higher than needed. Thus complete analysis is required in the production facility to measure the production performance. Therefore, in this study simulation model is developed for the tiles production to analyze the production performance and the ways to reduce the waste generated.
In this research simulation is used to measure performance because it is helpful and valuable tool in manufacturing. It allows the system behaviors to be tested and learnt. In addition, simulation modeling provides secure, low cost and fast analysis of the real system. More specifically, in this study, performance measures such as inter-arrival time, cycle time, throughput and WIP (Work in Process) is calculated to identify waste and improve overall production.
The rest of the paper is divided into following sections: next section describes the literature work in this area; then the methodology and data collection of the production system is presented followed by simulation model, results and discussion. Finally, the last section concludes our work and presents recommendations for improvement.

Literature Review
This research study focuses on process improvement using simulation as a tool. In today's world simulation has gained tremendous popularity because of its extensive use in design and information about process parameters of manufacturing systems under different conditions. Simulation allows pre analysis of system before implementing any technology.
A new application introduced by [5] provides simulation algorithm that is capable of producing all the necessary geometric information about the machining process.
Result of simulation helps experts [6] in companies to optimize quality method. A comparison research [7] explores techniques to analyze small samples from large trace files to find their difference using modeling.
Simulation models of spherical and polyhedral grains to analyze railway ballast in design and maintenance used by [8] where the models were compared to give different shear stress. Simulation plays main role in flexible manufacturing for random job arrival where [9] proposed Tardiness estimating method that provided better results.
Ribino et. al. [10] used simulation to avoid unproductive bottlenecks in warehouse optimization at design stage to analyze behavior of automatic logistic warehouse.
However, the above research has not discussed the difficulty in efficient flow of material that generates waste and its impact on revenues.
Performance of manufacturing systems can be efficiently analyzed by using simulation models and tools. Saez et.
al. [11] extracted discrete and continuous variable data from machines in real time and inserted in simulation tool to analyze its performance in virtual environment. A case study conducted by Abdul Malik and Rajgopal [12] represent simulation model to show production lead time and work in process inventory differences. Another research study demonstrated [13] simulation results regarding production performance analysis and conducted exception diagnosis in manufacturing processes. Hwang et. al. [14] proposed a model that Sproedt et. al. [15] presents decision support for two elements which includes discrete event simulation approach considering all relevant eco-efficiency drivers within a production system. Huang [16] concludes that the capacity of the industry is far more than utilization and productivity improvement using performance measures. Modern approaches like rapid prototyping using virtual models in simulation play vital role in manufacturing systems [17].

RESEARCH METHODOLOGY
Referring to Fig. 1

Description of Each Section in Production Process
The press molding machines presses the desired powder with high pressure of 200 bars. After pressing the powder, the tile called green biscuit is produced.

MODEL IN SIMIO
The data collected from industry is analyzed using As the production department is automated with some manual works, that is why the exact data was provided by the company.

TIMES, TRANSPORTATION TIME
The following data was collected from the tiles production. Table 1 shows the data of inter arrival time, processing times and Transportation times.

Tiles Pressed by Press Machine (Per Shift)
The amount of tiles which the press machine pressed per shift is taken.The average data is taken from industry for use in simulation model for press machine. Table 2 shows the data of tiles pressed by press machine per shift.

Throughput
Throughput is defined as the total good parts produced per unit time. So the throughput provided by the forte industry is shown in Table 3.

No. Description
Time (sec) Line

Wastes Produced at Production Lines
The big issue of Forte Ceramics industry is that the wasted parts are very high. The data collected for this purpose is average data and was collected at specific points. The data also shows the waste produced before entering machines. Table 4 shows the wastes produced before different points and the overall waste parts per shift.
All these objects were used to develop the simulation model. The machines were arranged in the SIMIO model as in the real system. After feeding all the collected data in the SIMIO software the final model was developed.

Modeling of the Production Process in SIMIO Software
SIMIO software was used as a tool to study the current production process and analyzed the cycle time, throughput, work in process, inventory and production rate of tiles. Wastes were identified at different workstations and the reasons behind them. Simulation model of the process is shown in following Fig. 3(a-c).
The snapshots were taken after verification and validation.

Verification
The verification phase involves that whether the collected data was right and the model represents the real system.
Since the data was provided by the industry that is why the data is actual and correct.

Validation
In the validation phase the results from simulation model and the real system are checked for any differences. So the simulation results were compared with the real system results.

. STATUS PIE CHART OF COMPLETE SIMULATION MODEL
The error in the results in Table 5 is due to the wear and tear in the machines in the real system, personal errors and environmental errors. The differences of the results are small in percentage so the model is valid. Table 6 shows that with inter-arrival time of 7.15sec in real system the wasted ones are 6289 parts which is 26.6% of total parts (17343) entering the system. Wastage of more than 26% is not bearable for any industry. This study focuses on reducing the waste with least changes in production process in order to cope up with the waste that significantly impacts the profit.

Amount of Production of Different Categories of Tiles
There are four types of tiles such as classics, standard, commercial and sub-standard tiles produced in Fort Ceramics industry. The four types depend on the quality and demand of tiles. Classic being the top quality is 66% of overall production, standard tile is 17.60%, and commercial is 7.4% and sub-standard is 4% of production. The following data about the four types is shown in Table 7.

RESULTS
After the simulation model, the results are analyzed to find the respective processes. These results include the performance measures, comparison of simulation results with real ones, throughput, and work in process and production quantity.

Performance Measures
Performance measures are quantifiable indicators which show the performance of organization that how the organization is achieving its desired objectives. Different managers in organization routinely review various performance measures and assess such things as production, demand and operating efficiency in order to get a more objective sense of how their business is operating and whether improvement is required. Table 8 shows the performance measures from simulation model results. Table 8 shows that the work in process of the current system is 2268.25parts/shift and the cycle time is 0.7813sec for each part. The throughput is basically good parts produced per hour that is 2006.4 and the waste produced per shift is 5737parts. The waste produced is too high and it is a big issue in frontier forte ceramics.

Comparison of Simulation Result with Real System
After completing the simulation model it gives the result so the comparison between real system results and simulation model results are given in Table 9 and also shown in Fig. 6. Where it can clearly be depicted that SIMIO and real system results are almost same with little difference. The graph in Fig. 6 shows two lines red and blue, the blue line represents real system data that is currently being produced at FORT Ceramics industry while the red line represent SIMIO results that has been obtained after putting data in software.
The result in Fig. 7 shows that there are small differences in the real system data and simulation results that might be due to system errors. As the difference is less than 10% that is within the acceptable range therefore, the data is valid.

Throughput versus Work in Process
After completing the simulation model the actual inter arrival time was 7.15 sec. The performance of the system was then checked by changing the inter arrival time.  Similarly, Fig.8 shows the graph between throughput and WIP where critical WIP can be easily estimated by the peak point. The figure shows the effect of change in interval arrival time on WIP and throughput. Similarly, it shows that when decreasing the inter arrival time the throughput increase till inter arrival time of 6.25 sec and after this the throughput is constant. On the other hand, when decreasing the inter arrival time the WIP increasing.
The yellow dot shows current and real time data as the inter-arrival time is 7.15 seconds.The grey dot shows 2123 parts as throughput with critical work in process of 3304.6 parts. It is called critical because by decreasing the inter arrival time there is no specific change and throughput remains constant as shown in Fig. 8.

Cycle Time versus Work in Process
After completing the simulation model, the actual inter arrival time was 7.15 sec. The performance of the system was then checked by changing the inter arrival time. Table 11 and Fig. 9 show the difference in cycle time and WIP level by changing the inter arrival time.
Similarly, Fig. 9 shows the graph between cycle time and WIP. We have to select the point in cycle time where more parts are produced in least time. The Fig. 8

FIG. 7. RESULTS OF REAL SYSTEM AND SIMULATIONS IN SIMIO
inter arrival time of 6.25 sec and after this the cycle time increasing. On the other hand, by decreasing the inter arrival time the WIP increases. In Fig. 9 the grey dot shows inter-arrival time of 6.25 sec where the work in process is critical having least cycle time. The yellow dot shows the inter-arrival time of 7.15 sec which gives 2268.2 tiles per shift having cycle time of 0.78 seconds.
Therefore, from

Improvements in Production Quantity
After completing the model and finding the critical work in process, the model is checked for the inter-arrival time of 6.25 sec and then the results were compared. Table 12 and Fig. 10 show the improvements between the real time results and the simulated results by reducing the inter arrival times.
Result in Fig. 10    production while the WIP is not increasing from critical WIP level.
The following significant results are achieved that impact the overall production using simulation tool shown in

DISCUSSION
In this study the tiles production of the forte ceramics is conducted. From the Inter Arrival Time, Processing Times and Transportation Time it is found that the big issue of Forte Ceramics industry is that waste is produced in tiles production. That impact their overall production and loss of revenue. SIMIO software was used as a tool to study the current production process and analyzed the cycle time, throughput; work in process, inventory and production rate of tiles. Model was created and simulated in SIMIO of the current production department at different cycle times. The results were analyzed to find the respective processes which improve the production rate to a greater extent as discussed in the following section.
The results in Table 8 show that the work in process of   shift and WIP could also be achieved with reduced inter arrival time. The impact of the overall production using simulation tool is shown in Table 13. It can be seen in the results that by decreasing the inter arrival time, the throughput increases which results an increase of revenue. So when applied with 25% of sale price gives a substantial profit that is a big improvement in the process.
One of the limitations of this study is that the design changes and especially the layout changes are not analyzed in the simulation model. The design changes will have an impact on the overall production of the tiles.

CONCLUSION
In this study simulation model for the production facility of Ceramic Company is developed. The results represent a significant improvement of the production process in productivity, cycle time and throughput time to optimize the system. There is a substantial increase in production of each type of tile by changing the inter-arrival time from 7.15-6.25 seconds. In addition, significant results obtained are summarized here: (a) The critical work in process is achieved at throughput 2123 parts with WIP 3304 parts which almost remains constant after increasing the throughput.
(b) Cycle time of 1.00 sec is achieved at WIP of 3304 parts.
(c) Simulation results show an increase of 646 classic tiles as each piece price is PAK RS. 40.00 which could gives total of PAK RS. 25840.00 increase in sales.
(d) Standard tiles are increased by 208 and each tiles price is 35 Rs that could give total of PAK RS. 7280.00 increase in sales.
(e) Similarly, the increase in sales of commercial and Sub-standard tiles are PAKRS. 1050.00 and PAK RS. 1232.00 respectively.
It is concluded that by decreasing the inter arrival time, throughput can be increased, therefore the real time of 7.15sec is reduced to 6.25sec in SIMIO software. By decreasing the inter arrival time there is an increase in the throughput which improves the revenue. Therefore, data analyzed in the simulation tool shows better results that represent a substantial improvement in the production process.