MDC-Opt: Modular Data Center Optimizer

Tool Overview

MDC-Opt, short for Modular Data Center Optimizer, is a Dynamo and Generative Design tool for evaluating modularized data center configurations during early design.

The tool helps users estimate and compare key performance outcomes by adjusting module count, number of floors, and rack power density. Based on these inputs, MDC-Opt calculates IT power capacity, cooling load, PUE proxy, and cost per MW.

The purpose of the tool is to support early decision-making for modular data center planning. It allows users to quickly test many design alternatives and understand how modular layout decisions affect power capacity, cooling demand, energy efficiency, and cost performance.

This tool is not intended to produce final engineering values. Instead, it provides a conceptual performance framework that helps users compare tradeoffs between different modular data center options.

What the Tool Does

MDC-Opt generates a simplified modular data center massing model and evaluates each design option using four performance metrics.

The tool helps answer the question:

How can a modular data center be configured to maximize IT power capacity while minimizing cooling demand, facility energy demand, and cost per MW?

Instead of manually testing one design at a time, users can use Generative Design to quickly compare many alternatives and understand the tradeoffs between capacity, cooling, energy efficiency, and cost.

Design Inputs

The user controls four main variables:

These inputs allow the user to test different building sizes, layouts, and power-density assumptions.

Fixed Assumptions

The tool uses simplified metric assumptions to keep the workflow clear and easy to use.

The model uses metric units. Each modular unit is assumed to be 12 meters long by 6 meters wide, with a 4.5-meter floor height. This is approximately 39 feet by 20 feet, with a 15-foot floor height.

For example, if Module Count X = 4 and Module Count Y = 5, the building footprint becomes:

Length = 4 × 12 m = 48 m
Width = 5 × 6 m = 30 m
Footprint = 48 m × 30 m = 1,440 m²

If the Number of Floors = 4, the total floor area becomes:

Total Floor Area = 1,440 m² × 4 = 5,760 m²

Performance Outputs

The tool evaluates four outputs:

How the Tool Works

The Dynamo graph first creates a modular building mass based on the number of modules in the X and Y directions and the number of floors.

Then it calculates:

1. Total building area
2. Data hall area
3. Estimated rack count
4. IT power capacity
5. Cooling load
6. PUE proxy
7. Construction cost
8. Cost per MW

The Generative Design interface then tests multiple combinations of the inputs and ranks the results based on the selected goals.

Main Formulas

For the cost output, the model uses a simplified data-center-specific cost assumption rather than only a building shell cost. This makes the cost metric more appropriate for comparing data center alternatives during conceptual design.

Typical Result

One high-capacity result from the Generative Design study was:

This result represents a high-density modular data center configuration. It provides strong IT power capacity, but it also requires a much larger cooling system.

Design Tradeoffs

The main tradeoff is between IT power capacity and cooling load.

A larger data center with higher rack power density can support more computing equipment, but it also creates more heat. This means the cooling system must become larger and more energy intensive.

Another important tradeoff is between total cost and cost efficiency. A larger building costs more overall, but it may have a better cost per MW if the rack density is high.

The best option is not always the smallest or largest design. The most useful option is the one that balances:

• High IT power capacity
• Reasonable cooling load
• Low PUE proxy
• Low cost per MW

How to Use the Tool

1. Download the project package.
2. Open the Revit file.
3. Open the Dynamo graph.
4. Confirm that the four inputs are visible:
• Module Count X
• Module Count Y
• Number of Floors
• Rack Power Density
5. Open Generative Design in Revit.
6. Select the study type: Modular Data Center Performance Study.
7. Set the goals:
• IT Power Capacity MW = Maximize
• Cooling Load Tons = Minimize
• PUE Proxy = Minimize
• Cost per MW = Minimize
8. Click Generate.
9. Review the generated alternatives and select the preferred design option.

Original Idea and Evolution

The original idea was to create a performance-based generative design tool for a modular data center. At first, I considered using more general building metrics, such as floor area, daylight, and construction cost. As the project developed, I decided to focus on metrics that are more important for data center construction projects.

The final metrics became:

• IT power capacity
• Cooling load
• PUE proxy
• Cost per MW

This shifted the project from a general building massing study to a more data-center-specific performance analysis.

The final tool focuses less on architectural appearance and more on early-stage decision making. The geometry is simplified so that the project can clearly show how different modular configurations affect capacity, cooling demand, energy efficiency, and cost performance.

One important change during development was the cost logic. The first version calculated cost using only floor area and a general construction cost per square meter. This produced a cost per MW that was too low for a real data center. The revised version treats cost per MW as a data-center-specific performance metric, making the output more useful for comparing modular data center alternatives.

Limitations

This tool is intended for conceptual design and early-stage performance comparison.

It does not include detailed engineering for:

• Electrical redundancy
• Backup generators
• Substations
• Utility service capacity
• Detailed HVAC routing
• Structural design
• Fire protection
• Site constraints
• Data hall equipment layout
• Real construction scheduling

The massing model is simplified, and the performance calculations use assumptions. Because of this, the outputs should not be treated as final engineering values. They are best used for comparing design alternatives and understanding tradeoffs.

Why This Tool Is Useful

MDC-Opt is useful because it allows a user to quickly compare many data center design alternatives without manually rebuilding the model each time.

The tool helps users understand how early design decisions, such as module count, floor count, and rack power density, can affect major project outcomes.

For data center projects, these decisions are important because power capacity, cooling demand, and cost efficiency are closely connected. This tool makes those relationships easier to see through Dynamo and Generative Design.

Video Demo

Demo Video - https://drive.google.com/file/d/1GKgBRkbOMVFL9tSqbaupo_pznhZoaZmL/view?usp=drive_link