Manufacturing Simulation Software: How Northrop Grumman Expanded Modeling Capabilities with Simio

Introduction

Northrop Grumman, a global aerospace and defense technology company, faced significant challenges in production planning due to the complexity and scale of their manufacturing operations. The company’s Space Park site, responsible for creating intricate and innovative projects like the James Webb Space Telescope, required advanced simulation capabilities to effectively plan and manage their production processes. This case study explores how Northrop Grumman leveraged Simio’s manufacturing simulation software to develop a revolutionary long-term planning model that dramatically expanded their modeling capabilities and improved strategic decision-making.

As Marie Scholl, an industrial engineer at Northrop Grumman’s Space Park site, explains, “We needed flexible capabilities to help us with bringing these projects to life.” The company’s vertically integrated manufacturing processes, spanning from sea to space, demanded simulation tools that could handle both immediate production needs and long-term strategic planning. This case study details how Northrop Grumman’s implementation of Simio’s simulation technology transformed their approach to production planning and resource allocation.

Company Background

Northrop Grumman operates at the cutting edge of aerospace and defense technology, creating complex systems that, in their words, “make the impossible possible.” Their Space Park site is responsible for developing sophisticated technologies ranging from space systems to advanced computing solutions. The company’s manufacturing processes are highly specialized, involving intricate components and innovative production techniques.

“The things that we build are intricate, complex and innovative,” notes Scholl in describing the company’s operations. “We are very vertically integrated, allowing us to make a wide array of nontraditional manufacturing projects, from sea to space to everything in between.”

This complexity creates unique challenges for production planning. With thousands of orders flowing through their manufacturing facilities each month, each with multiple operations and specific resource requirements, Northrop Grumman needed sophisticated simulation tools to effectively manage their production environment. The company’s stakeholders—including manufacturing managers, manufacturing teams, program managers, and executive leadership—all required different levels of insight into production planning and resource allocation.

The Challenge: Limitations of Short-Term Modeling

Northrop Grumman initially developed a short-term production planning model using Simio’s manufacturing simulation software. This model took detailed information from various data sources, built a simulation model, created a feasible production plan, and provided results for day-to-day production planning. The model was effective for planning up to 12 months into the future, offering valuable insights for daily tasks, worker assignments, and priority adjustments.

However, as Scholl explains, “No model is perfect. There are known limitations within this short-term model.” These limitations included:

• Extreme Detail Requirements: The model’s high level of detail made running simulations time-consuming and required extensive data cleanup before implementation.
• Limited Time Horizon: Since data details became less reliable beyond 12 months, the model could only run for a year or less, making it difficult to plan for long-term changes.
• Reactive Rather Than Proactive: The 12-month limitation made the company more reactionary than their approval processes sometimes allowed, particularly for equipment layouts and staffing changes.
• Training Constraints: With predictions only going 12 months out, it was challenging to train people for needed skills from the ground up in time for future production needs.

These limitations created a significant gap in Northrop Grumman’s planning capabilities. As Scholl notes, “We need to know if we should hire more team members or buy more equipment now, so we’ll have the skills and resources available when the time comes. We also want to test out potential layout and facility improvements, and even try bigger plans, like opening an entirely new building.”

The company needed a solution that could provide insights beyond the 12-month horizon while requiring less detailed data input. This led to the development of a completely new approach to simulation modeling.

Solution: A Token-Based Long-Term Model

Developing a New Modeling Approach

To address the limitations of their short-term model, Northrop Grumman’s industrial engineers developed an innovative long-term model using Simio’s token-based simulation capabilities. This approach represented a significant departure from traditional modeling techniques.

“We took a whole new approach to this model starting from scratch,” explains Scholl. “We grappled with manual data, rethought how part families and tasks were created and run, made this model to run for five years at least.”

The token-based simulation approach eliminated the need for detailed visualization of parts moving through the production process. Instead, the model used tokens—virtual entities that carry information and execute process steps—to represent the flow of work through the system. This approach dramatically reduced the computational requirements while maintaining the accuracy needed for strategic planning.

Key Innovations in the Long-Term Model

The long-term model incorporated several innovative approaches to simplify data requirements while maintaining planning accuracy:

• Summarized Worker Availability: Instead of detailed worker schedules, the model used a “working days per month” approach that incorporated lost time and efficiency coefficients.
• Part Family Simplification: Rather than modeling each specific part number, the model grouped similar parts into “part families” with generalized sets of tasks.
• Simplified Bill of Materials: The model used a more generalized approach to materials, focusing on the types of components needed rather than specific part numbers.
• Monthly Bucket Planning: Tasks were assigned to monthly “buckets” rather than specific days or times, allowing for a more strategic view of resource utilization.

As Scholl describes it, “A token is like the invisible person behind the scenes, executing on everything. The process steps tells it to do. Sometimes it splits itself so that multiple tokens can be worked on in parallel.”

This innovative approach allowed Northrop Grumman to create a model that could run quickly while providing valuable insights for long-term planning. “With these unique new ways of modeling, our long-term model was born,” says Scholl. “It successfully gives us a five-year look ahead, requires a lower level of effort to get useful inputs since we need less detailed outputs, and all while having the model run in mere minutes.”

Implementation: Integrating the Long-Term Model

Implementing the long-term model required careful consideration of how it would complement the existing short-term model. Northrop Grumman developed a clear understanding of the differences between the two models and how they would work together to provide a comprehensive planning solution.

Comparing Short-Term and Long-Term Models

The implementation team created a detailed comparison of the two models to ensure stakeholders understood the appropriate use cases for each:

This comparison helped stakeholders understand when to use each model and how to interpret the results. As Scholl explains, “Several aspects of these models are beneficial to different stakeholders.”

Use Cases and Stakeholder Alignment

The implementation team identified three primary use cases for the long-term model:

1. Current State Analysis: Running the model with current labor and resources to output expected schedule performance and resource utilization.
2. Ideal Headcount Determination: Running the model with infinite workers and current resources to identify how many people are needed to meet the schedule.
3. Resource Optimization: Testing the impact of adding, updating, or removing resources to alleviate bottlenecks.

These use cases were aligned with specific stakeholder needs. Manufacturing managers and teams focused on daily details used the short-term model results. Manufacturing managers and program managers looking at a 1-12 month horizon used the short-term model for annual planning. Program managers, executive leaders, and manufacturing managers interested in beyond-year planning used the long-term model for capacity and headcount requirements.

As Scholl notes, “Industrial engineers will be there for these teams every step of the way to help them use, test and interpret results of both models.”

Results and Benefits

The implementation of the token-based long-term model delivered significant benefits to Northrop Grumman’s production planning capabilities. The model successfully addressed the limitations of the short-term model while providing valuable new insights for strategic decision-making.

Enhanced Planning Horizon

The most significant benefit was the ability to plan five or more years into the future, compared to the 12-month limitation of the short-term model. This extended planning horizon enabled more proactive decision-making for staffing, equipment, and facility planning.

“If the model shows us that there is a huge demand in two years that can’t be completed in time with our current workforce, we have that time to train up needed skills or even have time to hire more people as needed,” explains Scholl. “This is a much more proactive method than the short-term model allows for.”

Improved Model Performance

The token-based simulation approach dramatically improved model performance, reducing run times from hours to minutes. This efficiency gain allowed for more frequent model runs and scenario testing, enhancing the team’s ability to evaluate different planning options.

“Since you just uses tokens, processes, and the data tables, this method allows the model to run rapidly in a matter of minutes,” notes Scholl. This performance improvement was critical for a model intended to simulate five years of production activity.

Reduced Data Requirements

The simplified data approach of the long-term model significantly reduced the effort required to prepare and maintain the model. By using part families, generalized tasks, and monthly buckets, the model required less detailed data while still providing valuable planning insights.

“It’s quicker to make data for this model, but does inherently allow for a slight increase in human error comparison time,” acknowledges Scholl. However, this trade-off was acceptable given the model’s strategic planning focus.

Enhanced Decision-Making Capabilities

The long-term model provided Northrop Grumman with new capabilities for testing and evaluating strategic decisions. The ability to run what-if scenarios for staffing, equipment, and facility changes enabled more informed decision-making for long-term investments.

“These options allow us the flexibility to test out theories digitally and make plans for added resources well in advance,” explains Scholl. This capability was particularly valuable for planning major changes like facility expansions or new program implementations.

Future Directions

Northrop Grumman continues to enhance and expand their simulation modeling capabilities. The company has identified several key areas for future development:

Enhanced Usability

The team is working to improve the usability of the model outputs through customized dashboards and visualization tools. “We utilize Tableau as our visualization tool for end users to view and interact with similar outputs,” explains Scholl. “We customize and cater dashboards to fit their daily needs to quickly guide productivity.”

Process Documentation and Training

To ensure effective use of the models, the team is developing comprehensive process documentation and training materials. “We need to make a user process flow for people to follow,” notes Scholl. “Something outlining the general steps of how the models work, getting and inputting data, how, when or why to run the model, how to read the results, and then ways to improve plans and production based on the results.”

Model Validation and Integration

The team is working to validate the long-term model by comparing its results with the short-term model. “Cross validating our results with the short-term model will prove out its validity,” explains Scholl. “If the numbers generally match between the long-term models first year and the short-term model, then we will know our long-term model is valid.”

Expanded Implementation

Northrop Grumman plans to expand the implementation of the long-term model to additional manufacturing areas. “We need to find a pilot team to adopt the use of our long-term model,” says Scholl. “We want to pick a team that has the best data available and that is willing and excited to use this model.”

Conclusion

Northrop Grumman’s implementation of Simio’s token-based simulation capabilities represents a significant advancement in manufacturing simulation and production planning. By developing an innovative long-term model that complements their existing short-term planning capabilities, the company has created a comprehensive solution for production planning across multiple time horizons.

The token-based approach to simulation modeling has proven to be an effective solution for long-term strategic planning, providing valuable insights with reduced data requirements and improved performance. As Scholl concludes, “From the minute to minute details to broad five year forecasts, industrial engineers will be there for these teams every step of the way to help them use, test and interpret results of both models.”

This case study demonstrates the power of innovative simulation approaches to address complex manufacturing planning challenges. By thinking beyond traditional modeling techniques and leveraging Simio’s flexible simulation capabilities, Northrop Grumman has enhanced their ability to plan for the future and make informed strategic decisions.