Validation of DfMA and Data Platform in Data Strategy Beyond BIM and Its Future

MICHIAKI INUZUKA: Good afternoon. Thank you very much for attending our session. We are going to present validation of DfMA and data platform in the data strategy beyond BIM and its future. Let us introduce ourselves.

My name is Michiaki Inuzuka. I joined Daiwa House in 2008, and have been working on architectural design of various kinds of buildings, such as commercial facilities, hotels, nursing care homes, logistic warehouses, factories, and so on. If you have ever stayed at the hotel in Nagoya during your business trip to Japan, it might be a hotel I designed.

As I went through a transition of design phase, from 2D to 3D to BIM, I am very excited to think about how BIM will develop further in the future. Hope I can share my excitement with you today.

KAITO SAWA: My name is Kaito Sawa. I am honored to be here and to have so many professionals in the audience I studied digital design, digital fabrication, and BIM at University and joined Daiwa House in 2020. Although it's only been three years, I have been working towards the transformation of the construction industry and the theme of DfMA and IC, industrialized construction.

Today, I'd like to share some of our efforts and our vision for the future. With your feedback, we'd like to further evolve our activities.

MICHIAKI INUZUKA: This is our company profile. Daiwa House Industry, Company Limited. Head office is located in Osaka, Japan. The company was established in 1947, and started its operation in 1955, about 70 years of history. Number of employees are around 50,000. The net sales in the last fiscal year resulted in 4.9 trillion yen, which is around $33 billion. If converted as a fixed rate, 145 yen to $1.

In the 1950s, when we began operations, the post-war reconstruction in Japan was in full swing. With the rapid economic growth, many buildings, schools, residences, offices, et cetera, were needed. In 1955, our founder developed a prefabricated building, utilizing the standardized components, being shown on your right in the slide.

It enabled to deliver large quantity of buildings in short construction periods. Based on the philosophy of proposing and implementing solutions to the needs of the times, Daiwa House has grown significantly. Here is a short video to show our broader business areas currently working on.

Throughout this session, we would like to share the content, as listed here. Our English is not fluent, at all, but we hope we can convey some of the elements that are resonate with you and meaningful for our industry to move forward. Now, I would like to hand over the next section to Sawa-san.

KAITO SAWA: Thank you, Inuzaka-san. Now, among four running objectives, I'm going to cover these three points, exploring some of the important elements to apply DfMA in their project.

First of all, what is DfMA? As many of you, as many of you attending this class may already know, DfMA is an approach in design with manufacturing and assembly in factories and construction sites in mind. Taking a deeper consideration on how to build in the early stages of design, with [INAUDIBLE] method, more optimized, industrialized construction can be achieved.

As Inuzaka-san introduced our company earlier, we have developed prefabrication technology, since the beginning, and have promoted industrialized construction in housing business. On the other hand, in building a business, we still rely on a conventional approach in design, , manufacturing and construction process, except for a few products.

The shortage of labor in the construction industry has become a social issue. We also feel the current process is limited in its ability to meet the rapidly increasing needs for housing and buildings globally. [INAUDIBLE] construction is a solution for such needs and challenges, and we believe DfML will provide significant growth for innovation.

To proceed with our case study on DfMA, we put our focus on these two points. The first is concurrent decision making and the second is productization.

The first point is about transforming processes to achieve concurrent decision making. The left is a traditional process. Data is passed from design to manufacturing to construction.

In contrast, the process on the right the future process. Each stakeholder shares data in the entire project with data at the center of the process. [INAUDIBLE] process from [INAUDIBLE] for decision making to concurrent decision making, is a key.

The second point is to establish a workflow to achieve the building configurations, based on the productized data. In a typical construction process, designed to accommodate manufacturing data on a case by case basis. Our future process, on the right, uses predefined manufacturing data for design. Productized the components may be considered to enable the building design, with standardized parts as much as possible.

The use of productized data supports the concurrent decision making described on the previous slide. Having the support from [INAUDIBLE] this slide summarizes the capabilities required to apply the FMA even more effectively. In the case study, we focus on these two themes as a foundation for applying DfMA.

We will now follow the flow of the hypothesis testing conducted as a case study. It consists of three steps. Let me explain step one.

We have defined the future process by focusing on the early stages of design, product development, and project development. said those processes are very important to apply productization concept and to study the building configuration based on productized data.

The first process is to create the product, or kit of parts, to be used in the design of each project. We designed prefabricated panels and units to achieve the industrialized construction.

In this process, we started to define the product concept. The product must be defined, not only in terms of appearance and function, but also based on the FMA principles.

Next, we use the wireframe approach for modeling. Wireframe is to represent the defined concept and pass a bill, based on the frame, as a master assembly or template assembly. The parts can be designed as a parametric product. Dimensions can be changed following the defined rules.

We also added attribute information to those products, for instance, weight. It will be updated referring to the master.

At this point, products cannot be used for building configuration, as they are. The model for production is converted into a model for building configuration and managed as a library to be used by designers.

The second process is a product evaluation process based on data. Designing products from DfMA principles requires skill to determine and build production ready componentization, considering manufacturing, as well as on site assembly. To support decision making by various stakeholders, dashboards will be deployed, showing data efficiently and effectively.

The dashboard will be maintained in the environment, accessible to all stakeholders. The sale process is a building configuration process by products. Create a concept design based on customer requirements and selects the product that satisfies specifications of the design.

The families associated with the product data assessed and referenced. Product selection is performed, checking the product performance, shown on the dashboard. With the approach, that designer can build the design, using producer products, without having to enter each component configuration from scratch.

In case or components of the building cannot be configured by already defined products, we need a process to request a new product. The process is the evaluation process for the configured building. We talked about product evaluation in process too.

Similarly, we can prepare a dashboard for the building to support decision making. The dashboard calculates KPIs for the building, by reporting it to the KPI for the product. By referring it to KPIs, we can make changes to more advantageous design. Also, the verification history can be saved, shared, and compared by all stakeholders. That's all for future processes we defined in the case study.

Next, I'd like to explain how we are going to achieve the process we shown in the video. We call it DDP DfMA data platform. At the bottom, the data layer, project data, and product data shown in the red box across three link. The linkage allows us to configure buildings with product data and to evaluate performance with dashboards.

Data is managed in the cloud and can be accessed using protocols defined in the middle layer, the service layer, through the presentation layer at the top.

The data is available to those who need it, when they need it. So far, we have discussed the process definitions and to proceed with our validation steps. As a next step, we have examined the feasibility of the process and the barriers for implementation. Then analyzed and evaluated the results.

The project we've selected for this verification is a timber apartment about 300 square meters. Since [INAUDIBLE] has not been matched upright to timber buildings and processes are fragmented in Japan, without the benefits of applying DfMA could be significant. In our [INAUDIBLE] we focus on three points.

Try to quantify the effect as much as possible. First, how much time can be saved in creating building manufacturing data in design phase with the productization concept? Second, how much of duplicated works can be eliminated by using data? Third, how much risk of inconsistency in design phase can be reduced by applying our future process?

The scope of the validation is illustrated in this slide. The process rolls from left to right. First, project related to data and factory data were collected, shown in green. Then future processes one to four were executed in parallel with the actual project execution, shown in orange. And finally, a prototype of the part of the products was made, shown in blue.

From this slide, I will explain three verifications conducted one by one. The first case is with the future process we define how much we can reduce the time to build manufacturing data in each project design phase. In this verification, we executed the future process, in parallel with the traditional process, and we compare the time to build a manufacturing data for the entire building.

In our variation, executing the defined feature process has resulted in about 80% reduction in the time required to build the manufacturing data. Most of them can be built simply by selecting and aligning the productized data from a library, according to the required performance. We could eliminate the time to create fabrication drawings from scratch. [AUDIO OUT]

The manufacturing data for products must be defined and ready to be manufactured. When prototyping some [INAUDIBLE] we created design data and product data. However, the data could not be linked and we had to rebuild the manufacturing data. This result reminded us the importance of data integration across departments in productization.

Next, let's look into the second verification. Here, we investigated on how much duplicate tasks we can reduce by implementing the future process we have defined. We identified the tasks done in the previous phase but the similar tasks repeated in later process. We examined whether such duplicated work could be eliminated by applying the future process.

Through this verification, the total time saved in the design phase was about 20 hours. The cost estimation only needs to address project specific elements and quantities are already known from product data. The time spent discussing on how to bear can all be deduced by reusing the product data, eliminating the same studies for those productized area.

Although the saving time was about 5% of the total design time for the entire building, if the scope of reusable data is expanded and applied to larger buildings, the impact could be larger. In order to ensure this time saving effect, it is important to accumulate reliable and high quality, reusable data.

Finally, the third validation was to verify how much the risk of inconsistency could be reduced in each project design phase. By constructing the building with productized components, with it, it is possible to resolve inconsistencies in the design stage, considering each component level, not just between architecture design and structure design.

If the crash is found during the panel production stage, it causes on site corrections and panel remaking, resulted in schedule delays and additional costs. In this validation, we could automatically detect a crash at the inside corner of the external wall panel. We could also identify an error in design, where it couldn't be configured by available panels.

If we apply further products based configuration at the design stage, we could reduce subsequent process delays and additional costs, effectively. Furthermore, we tried to verify the process through the actual panel prototyping based on design data. Unfortunately, the attempt was failed, due to the high dependency of the manufacturing equipment.

Here is a summary of our validation. First, we saw the possibility of up to 80% reduction in building manufacturing data in each project design phase. Second, we could save 20 hours, eliminating duplicated tasks. Third, with a future process, building can be configured by producable products during the design phase, and we can eliminate the risk of inconsistency.

Through this case study, we run some key elements to apply DfMA. First, the importance of data integration across organizations. It is essential to enable productization and product based design configuration. Second, the importance of reliable, quality data management. To promote reusing data, it is important to accumulate reliable, high quality data, and maintain property

Finally, consistency between design and production. Productization can not be effective if design data does not flow directly to production. In summary, how we create, accumulate, assemble, and maintain data, we will enable the DfMA for effectively in our company.

So far, I have covered three learning objectives. Next, I'm going to hand over to Inuzuka-san to talk about the last topic, our data strategy.

MICHIAKI INUZUKA: Thank you, Sawa-san. As we mentioned at the beginning of this session, we have been developing and constructing products in housing business for almost 70 years, focusing on prefabricated housing. About four years ago, we launched a project to revisit and reinforce our DNA, industrialized construction, having support from Autodesk consulting.

Through the project, we realized once again, the importance of the way we create and accumulate data. As we dig deeper into the methods and concepts of DfMA, in this chapter, we share our data strategy, including new restarted project with Autodesk Consulting this year.

As Sawa-san's team worked on the productization process for building components in its basis of DfMA. By implementing this process, the design and manufacturing of assemblies for factory production can be more effectively implemented. And then once [INAUDIBLE] buildings is established, the conceptualization, design, and simulator of complete built environment asset can meet downstream project requirements.

Furthermore, connected construction will simplify field operations and supply chain logistics through a digital site and assembly process. Common data environment is a cloud based platform to support the design and delivery of products and projects. It's centralized data storage and access, according to defined protocols.

As we discussed in the previous slide, there is a close relationship between BIM and [INAUDIBLE] Similarly, between DfMA approach and data strategy. We have described our model as [INAUDIBLE] of BIM and construction data, where data has grown in a circular manner throughout the entire construction life cycle.

Data storage and DfMA are similarly related to each other, we believe. In order to achieve true data driven transformation, we are working on a project to achieve data centric world. This slide shows how improved the use of data and information can significantly increase the company's capabilities and ability to execute its strategy.

In our project, we have defined three stages, naming the significant improvements in digital readiness at stages of transformation. As the stages progress, more advanced technologies and capabilities are acquired and utilized. Each stage has solutions that provide more advanced capabilities, as the stage progresses.

Stage one, focusing, focuses on structuring data in functional areas and centrally managing the data on the platform. Analytics in functional areas are possible at this stage.

Stage two will focus on the integration of design and production assembly through the platform, leading to enterprise diagnosis and analysis. Stage three will involve further integration of design, manufacturing, construction, and operations. The introduction of simple AI or machine learning enables predictive analysis across the business.

Beyond that point, we envision more advanced generative AIs, relating independently and collaboratively, for even greater generative design and make capabilities. Based on such data strategy, we have started to work on establishing the foundation of data among business units, as shown in the preview. Our goal is to build an environment where necessary data can be accessed by necessary people, in necessary timing, with security.

I saw some mentioned earlier, there are issues scattered in the current workflow in linking design data and production data. It is essential to transform through conventional workflows and processes, by building a mechanism centered on application independent data. [INAUDIBLE]

Now, we'd like to wrap up this session. Today, we have talked about the items in this slide. Based on DfMA concept, we define four future processes and tested their effectiveness, learning the process, along with the conventional processes. With the result, we have learned some important elements in applying DfMA. We also touched upon the data platform required to achieve greater value and effectiveness, as well as our data strategy.

At the beginning, I mentioned, we have about 50,000 employees in Daiwa House. Many people from many countries work together. With various business operations, the data accumulate. We may differ, depending on the region and the nature of the business. Data required for each, may also differ.

At the same time, changes in the needs of society and changes in technology are accelerating with each passing year. We believe to respond quickly and flexibly to such changes, we need to accelerate our transformation with data at the core, further developing the data center, the data centric project with Autodesk. We will accelerate our growth.

Last year, in AU 2022, we set forth the concept of digital Daiwa House, by promoting data centric initiatives. We try to create new value by confronting and solving various issues, not only internally, but also on a global scale.

Thank you very much for your kind attention.