Consider Discrete Event Simulation

The technique can provide supply chain and process improvement insights.

By Dayana Cope, Eastman Chemical Company

Share Print Related RSS

Contrary to what the name might suggest, discrete event simulation (DES) isn't just for "widget" manufacturing. Indeed, it has particular value in the chemical industry. Yet, chemical makers have been slow to adopt this methodology. So, in this article, we'll look at what DES can do.

The chemical industry widely uses process simulation and inventory optimization tools, and clearly understands that modeling has a very important role in decision-making. However, there's a missing link between the focused analysis performed with these tools and the need for a clear comprehensive view of the entire supply chain and how factors like continuous operations, discrete processes, inventory planning and logistics affect specific key performance metrics. DES supplies this missing link by providing a method to quantify these interactions and their effect on the bottom line. Furthermore, it enables decision-makers to assess "what if" scenarios to explore the different optimal configurations that can be achieved or the possibilities of yet-to-be built plants and processes.

The Value of Simulation
Simulation consistently ranks as the most useful and powerful of mathematical-modeling approaches. This stems in part from its ability to handle high fidelity models of very complicated systems. Recent advances in computing power and the ease of use of "commercial-off-the-shelf" simulation software bolster its appeal. In addition, the flexibility provided by hybrid continuous/discrete simulation software has paved the way to an increase in viable applications, including for the chemical industry.

Most chemical processes are continuous, but discrete operations also are common at plants. Enabling processes such as procurement and logistics are discrete, as are packaging, raw material arrival and asset maintenance. So when modeling the chemical industry from end to end, it's important to include elements of both continuous and discrete change.

A chemical maker's supply chain must be responsive, reliable and flexible. Production processes often are campaigned to ensure nonstop operation. Premature halts — due to lack of raw materials or other supply chain failures — often can incur high costs. A missed shipment of energy or feedstock materials could cause a total plant shutdown, resulting in a loss of millions of dollars. To mitigate this risk, chemical companies often carry large inventories of raw and intermediate materials and finished goods. Compounding this, customers are implementing lean strategies that, in effect, push inventory demands back onto chemical makers.

Supply chains are complex systems to model due to their uncertain and highly variable nature. The integration and interrelation of suppliers, suppliers' suppliers, customers, etc., mean that something taking place at one company could greatly affect supply chain activities. Another complication is that supply chains are dynamic. Changes such as an enterprise leaving or another joining the chain are common. Therefore, decision-makers need a methodology that allows for timely and efficient updating to reflect such changes. In addition, supply chains span numerous physical locations — and necessary information must come from all these sites.

With decision support tools based on mathematical models, spreadsheets or process map methodologies, decision-makers must contend with lots of assumptions that hardly ever hold true. Using a number such as an average for systems with variation isn't the solution. That approach fosters significant inefficiencies like production backlogs and unbalanced capacities — and also reduces revenues due to lost sales (where production couldn't meet demand) and higher inventory and other costs. The bottom line is that not taking variability into account costs money.

The power of supply chain simulation lies in its ability to provide the essential level of realism and utility required for accurate modeling. That's why companies in many industries have adopted DES as the methodology of choice when tackling supply chain decision-making.

A Powerful Tool
Consider, for example, the supply chain of energy and feedstock raw materials as well as the material handling and storage equipment required to transport and inventory materials before use in chemical production. A DES model provides invaluable insights into the reliability of the plant as a whole when it's used to stochastically analyze the impact of probabilistic failures associated with suppliers, procurement processes, logistics (rail, trucks, etc.) and equipment; uncontrollable events, such as natural disasters, also can be incorporated to observe their long-term effects. Furthermore, by modeling contingency plans for responding to failures, we can accurately identify the vulnerabilities of the supply and recommend areas for improvement. Simulations also can determine the distribution of the probability of plant failure over x number of years, and provide quantifiable justification for projects to improve overall plant reliability. We used a DES model to assess reliability gains from process changes versus capital improvement projects. The model demonstrated the proposed process changes could significantly boost overall plant reliability, making the major capital improvement projects unnecessary.

DES models can also provide insights on the operational requirements of chemical systems and the impact of logistics on these requirements. DES models often are used to answer questions such as:

• What are the effects on chemical production of adding or removing multiple suppliers, transportation, storage and treatment?
• What are the minimum requirements needed to allow for a continuous production outflow?
In addition, DES can enhance Six Sigma process-improvement efforts. Consider, for example, a debottlenecking project. The simulation could address questions like:
• How is the upstream or downstream process affected by this improvement effort? Will the bottleneck move upstream or downstream?
• Will the improved process require additional resources or inventory when seasonality or variability is taken into account?

Simulation also reduces Six Sigma practitioners' reliance on static models or prototypes and post-design rework. This often leads to quantifiable benefits — reduced time and cost — within the product development lifecycle as well as improved Six Sigma project cycle times.

Six Sigma and DES complement each other. Statistical tools support the Six Sigma methodology while simulation provides the next level of statistical rigor, accuracy and robustness required for many complex Six Sigma efforts where error in analysis can't be tolerated [1,2].

Furthermore, DES sometimes is the only viable solution for complex problems such as the optimal scheduling for campaigned products. A simulation-based optimization model enables observing the tradeoffs associated with shorter campaign runs, paired with efficiently set up processes and decreased inventory costs. Optimally scheduled campaign runs bring us one step closer to leaner, more responsive chemical processes.

DES isn't limited to existing operations. Modeling the discrete and continuous elements of a new process as a system can provide important insights for the design of future plants or processes.

A new chemical plant or process design often requires large investments in inflexible and expensive capital assets. Simulating the plant or process is a great way to validate design assumptions, analyze dependencies and analytically quantify initial capital expense as well as the rollout timeframe. Simulation models help in "derisking" capital decisions by providing an early proving ground to test design specifications such as warehousing space, equipment and labor requirements. Moreover, animated simulation models offer a powerful way to visualize a future plant or process and communicate details to stakeholders and others. We developed such a model to test the design specifications of a proposed capital project for a new plant. We used the model to analyze the effect on yearly throughput of varying process times in key equipment, and to determine the impact of increased processing times on warehousing space requirements and the amount of equipment needed. Having this information early in the design phase of a capital project reduces the risk of over- or under-estimating investment.

Simulation has its downsides, though. DES modeling may be hard for nonspecialists to grasp. This is an important drawback because understanding the methodology behind results often is a prerequisite to stakeholder "buy in." The effort needed for data collection and analysis also is a significant hurdle. Simulation frequently demands collection of large amounts of data to obtain valid models of the system. Furthermore, the data collected must be analyzed and transformed into statistical distributions. The techniques involved often require the knowledge of a statistics or simulation expert. The substantial effort associated with data collection and analysis frequently translates into huge investments in time, resources and money for initial implementation of simulation modeling methodologies. Then, once developed, the model must undergo validation and verification to ensure its usefulness and accuracy — this also necessitates a sizable time investment.

Add DES To Your Toolset
Chemical makers definitely should consider DES. It's a powerful tool in areas such as supply chain analysis and process improvement. DES enables decision-makers to obtain accurate results that take into account the uncertainty, dynamic behavior and distributed nature of supply chain environments. Simulation also complements Six Sigma by providing the next level of statistical rigor, accuracy and robustness required for many complex Six Sigma efforts, and can result in improved Six Sigma project cycle times.


DAYANA COPE, Ph.D., is an operations research analyst for Eastman Chemical Company, Kingsport, Tenn. E-mail her at dcope@eastman.com.

References
1 Ferrin, D.M., Miller, M.J., and Muthler, D., "Lean Sigma and Simulation, So What's the Correlation," Proceedings, 2005 Winter Simulation Conference, Orlando (2005).
2. Luce, K., Trepanier, L., Ciochetto, F., and Goldman, L., "Simulation and Optimization as Effective DFSS Tools," Proceedings, 2005 Winter Simulation Conference, Orlando (2005).

Share Print Reprints Permissions

What are your comments?

Join the discussion today. Login Here.

Comments

No one has commented on this page yet.

RSS feed for comments on this page | RSS feed for all comments