- Semiconductor
It should come as no surprise that our data-driven business culture is significantly influencing the design and construction of semiconductor manufacturing facilities, or “fabs” as they are known in the industry.
As the coin of the realm in most aspects of business, data has ushered in real progress in the semiconductor industry. Together with 3D modeling, it has enabled the development of accurate and data-rich engineering and construction digital twins that have, in turn, enabled workflows such as offsite manufacturing (OSM) and modularization to speed up fabrication, assembly, and installation, to improve on-site safety, and to reduce the overall expense and duration of construction for the fabs.
Digital twins: Streamlining semiconductor fab design & construction
Over the past five years or so, digital twins (the crossroads of 3D mapping, data, analytics and systems) have paved the way for offsite manufacturing and prefabrication, leading to better, faster and more efficient design of semiconductor facilities. This technological advancement spans the entire facility lifecycle, from design and engineering, to construction, equipment installation and operations – all the way to redesign or renovation, re-tooling, demolition or recycling, to end of life.
Data-rich digital twins are still a relatively new concept in the context of semiconductor fab design and construction. As a result, the data component is sometimes given short shrift in the project management process. Organizing a wide-reaching project like this requires sufficient upfront planning to fully understand the role of data from the very beginning.
Digital twins are the culmination of numerous data-driven technologies. Laser scanning and 3D reality capture, for example, have greatly improved existing 2D and 3D documentation, enabling virtual design and construction workflows. The illustration below shows how data drives the digital twin concept from beginning to end, to yield a “smarter” overall manufacturing process.
-
Read more
Artificial Intelligence and Machine Learning have also made inroads in semiconductor construction. These technologies can perform simulations and create business intelligence reports on large datasets in areas such as building components, materials and labor costs, procurement and logistics schedules, and suppliers or supply chain information.
All of this has resulted in better information to the right project stakeholders at the right time, to track project health and performance, statuses, issues, and trends along the way. Coupled with offsite manufacturing and modular construction, often the only action left in the construction process may be field module assembly for much of the overall facility and systems.
Unfortunately, it is currently common to see inaccurate and incomplete information and data from fab facility owners, as well as from their equipment suppliers – especially at the start of the project planning and design phases. This often results from a desire to move as quickly as possible in the process, foregoing the time required to develop and use data appropriately.
The right time for data
From my experience with numerous fab projects over the years, I can say that too often, there is a rush to get started on a project. This rush can come at the expense of gathering required data and information upfront, and effectively deploying it into the applicable data-related technologies.
So, at the risk of belaboring the subject, here are some pieces of advice that will help make your next project a success.
Data has upfront value
Data and information is currency. It has value. What some may not fully grasp, however, is that the value of data actually can be realized at the front end of the project.
Designers and other project stakeholders need to take all the time required to gather information from the client, and to check, validate, and refine the data at the outset of the project. The more high-quality, accurate data and information you can get at the start of the project, the greater the payback will be throughout the entire project life cycle.
-
Read more
Don’t rush
As a corollary to the above point, it’s important in project planning to build in a scheduled phase of the project for data and information gathering. That means checking and specifying with the fab owners, ensuring that the data is high quality, accurate, and ready to go before the first production person touches it.
That can take some time – ideally, during the first few weeks or so at the start of a project. Don't feel obliged to start design production as soon as the contracts are signed. That's where many projects fall down.
The team typically wants to move forward aggressively, mobilizing engineers and contractors as quickly as possible. Then two months or more into the project it’s often discovered that there wasn’t adequate data available from the start. As a result, you could find yourself having wasted hundreds of hours with a team of designers and engineers waiting around for information – or worse, you may already have performed significant work on incorrect or inaccurate information.
By not falling victim to an unnecessary need to rush into the project production, you can perform quality and completeness checks on the data and information. It is important to collaborate with the owner to be sure you have the highest quality data set possible. During this phase, it may become apparent that specific smaller aspects of the project should be undertaken first and evaluated before committing to full-scale design production.
Integrate standards
Design and construction projects are driven by industry codes and standards for physical facilities. But there are also standards related to data and information management for a project. For fab and facilities projects, the relevant standards are ISO 19650 and SEMI F122 standards. Let’s take a brief look at each.
The International Organization for Standards’ ISO 19650 provides a framework for managing information throughout a project’s lifecycle, using Building Information Modeling (BIM). The standard is comprised of five parts, covering different aspects of BIM, including planning, design, construction, operation, and maintenance. The overall goal is to ensure that the right information is available to the right people at the right time, enhancing collaboration and reducing risks.
Managed by SEMI, the SEMI F122 standard is otherwise known as the "Guide for Facilities Data Package (FDP) for Manufacturing Equipment Installation and Building Information Modeling (BIM).” It provides guidelines for facility and manufacturing equipment information, data, 2D drawings and 3D models of manufacturing equipment needed for fab planning and facility optimization for the installation of the manufacturing equipment. This is critical source data and information that feeds entire fab planning, design and construction projects, as well as the installation of manufacturing equipment.
These two standards come together first to design and construct the facility, then to install the manufacturing equipment, and then further feed into the fab operations and manufacturing. Therefore, integrating these standards before the project begins is fundamental to its success, and it is made easier because of the high quality data you have gathered along the way.
Data and information, and the technologies utilizing it, are dramatically evolving semiconductor fab design and construction landscape. By taking the steps outlined here, we can make the most of all available technologies, and enable better project outcomes for fab owners and suppliers.
