Introduction In the previous article, “Beyond Level of Detail into AM/FM,” we discussed the current development of the Digital Twin concept in the architecture, engineering, construction, and owner (AECO) industry, focusing on end users’ demands. In addition, a “3x3” rule was intro - duced to highlight the flexibility and adaptivity a user-oriented Digital Twin should include. This article takes a deep dive and breaks down the general Digital Twin framework into five application modules: Asset Operation, Asset Maintenance, Safety, Real Estate, and Sustain - ability, with the capability of data analytics and reporting, respectively. The collaboration of these five modules is an essential component to support the project’s operational phase from both operation and main - tenance (O&M) and data perspectives. Modularized Digital Twin Framework The new digital era welcomes technologies like data science, state-of- the-art gaming engines, the Internet of Things (IoT), Artificial Intel - ligence (AI), and so on. Realistically, an ideal Digital Twin platform would better connect with, and take advantage, of multiple systems/ services. The list includes but is not limited to: Enterprise Resource Planning (ERP), Customer Relationship Management (CRM), Inte - grated Building Management System (IBMS), Fire Protection and Life Safety (FP&LS), space management, Integrated Work Management Systems (IWMS), and Computerized Maintenance Management Sys - tem (CMMS). As shown in Figure 1, the modularized framework consists of three layers: Data Collection & Systems, Data Processing, and Business Intelligence, which allow data consumption to support each system while enabling business intelligence through a well-connected data processing gateway. Data and standardization should be implemented across all connected channels via data gateways and a data mart. Data mining, analyzing, reporting, or machine learning integration would be utilized to fulfill the demands of all stakeholders during the operational stage. Finally, AI could be used as the ideal human-computer inter - action (HCI) process to understand users’ natural language, collect, and analyze corresponding data from related connected channels, and provide answers to the user. We will discuss AI applications in-depth in the upcoming articles. The five identified modules that can encompass a holistic Digital Twin are Asset Operation, Asset Maintenance, Safety, Real Estate, and Sus - MODULAR DATA ANALYTICS AND REPORTING IN FACILITY MANAGEMENT - FIVE MODULES By George Broadbent, Dr. Jeff Chen, Dr. Eve Lin with Kai Yin, and Jiayi Yan
The Modularized Framework for a Digital Twin. Credit: George Broadbent, Dr. Jeff Chen, and Dr. Eve Lin
tainability. Each of the modules has been defined to work separately and collaboratively to support the operational stage as illustrated in Figure 2. We will begin to see how the virtual geometry of the building/ facility blends with the operational data through these five modules. It is through this merging that the long-term benefits of BIM are realized. The goal of BIM implementation in the operational stage is to portray a digital representation of the real world, the Digital Twin. While traditional BIM for design and construction will have a useful life of zero to five years (i.e., design and construction), the Digital Twin (BIM extended into operations) can expand it from 20 to 30 years, sometimes even 50. Asset Operation Module The Asset Operation Module performs general management practices on equipment running across the facility, enhances overall produc - tivity, and reduces asset downtime. The module uses the integrated data model, analytics, and management workflow. Compared with a traditional O&M workflow, this data-driven workflow can detect and predict potential risks and enhance the facility’s performance. The operation monitoring tools are introduced to help the user record, analyze, and identify risks, failure probability, and related conse - quences with data aggregated from the entire facility, a single room, an asset system or a single piece of equipment. This can overcome the drawbacks of a traditional fixed O&M workflow. A user predefined problem, cause, remedy (PCR) data library can be linked with specific asset classes or locations to provide more suitable decision-making suggestions. Meanwhile, real-time sensor readings, if available, need to be integrated and standardized via the data gateway into the global data mart. Machine learning-based algorithms will examine these data
48
csengineermag.com
september 2021
Made with FlippingBook Annual report