Semantic Interoperability for Building Data

Summary

A major obstacle to deploying building automation and leveraging artificial intelligence for system-scale optimization is the labor-intensive process of mapping building automation system data points to the relevant software tools. The necessary information about the mechanical systems design, point naming conventions used when installing the sensors and control systems, and other details, are contained in drawings, contract documents, and maintenance staff knowledge. Sometimes it is incomplete or partially wrong.

Achieving national goals for AI investment and emerging technologies like digital twins and grid-integrated buildings will require standardized metadata to enhance AI-driven Fault Detection and Diagnostics (FDD), enabling more accurate, efficient, and intelligent supervisory control processes. The focus of this project is to develop, test standards and other supporting infrastructure for representing the semantic information about the equipment, sensors, and actuators in a machine-readable way. It will include developing software tools to create building-specific models that conform to the standard using a combination of information available in BACnet systems and input from building operations staff.

Description

Objective 
To pursue an effort that results in national and international standards that define concepts and a methodology to create interoperable, machine-readable semantic models for representing building system information for analytics, automation, and control.

Technical Idea
The new technical idea is to adapt Semantic Web standards to the creation of formal models that represent building system components, their relationships in various contexts, and the associated data and control points. The Semantic Web is an extension of the World Wide Web that enables linkage of information, through knowledge graphs, located in different places and the encoding of metadata with that information that describes its semantic meaning. It is based on technologies such as the Resource Description Framework (RDF) and the Shapes Constraint Language (SHACL). These technologies can be applied in intelligent buildings domain to:

1. Represent knowledge from building system data and other sources of information related to building operations;
2. Integrate the multiform of data from various systems and applications; and
3. Enable description logic-based reasoning to support decision making based on integrated data from heterogenous sources.

Building specific models created with this technology will enable building analytics and enterprise knowledge tools to automatically find necessary information to implement applications including:

• (a) automated fault detection and diagnostics,
• (b) control system configuration,
• (c) building system commissioning,
• (d) digital twins,
• (e) optimization of energy use,
• (f) energy audits, and
• (g) smart grid interactions.

Research Plan
ASHRAE has sponsored the development of proposed Standard 223, Semantic Data Model for Analytics and Automation Applications in Buildings. This standard is crucial in the future of smart buildings. The development of the standard will leverage knowledge gained from earlier efforts to partially address this problem in the U.S. and in Europe including: Project Haystack, Brick, Building Topology Ontology (BOT), RealEstateCore (REC), Semantic Sensor Network Ontology (SSN) and Smart Appliances Reference Ontology (SAREF). It will also build on the QUDT ontology for measurements and units.

In addition to standard development, NIST will partner with other industry and national lab partners to develop and test software tools that apply the draft standard to various specific applications envisioned for the standard, aligning with our expertise. Our research team plans to assess the implications of the ASHRAE 223 standard for future smart building innovations and AI-driven applications for buildings. NIST will also participate in efforts to make this standard an ISO standard. This project will enhance major industries in building automation and analytics, which currently recognize the need for these semantic models but have no alternative other than using nonstandard solutions. Furthermore, this project will enable technical advancements for manufacturers, facilitating the development of standard-compliant versions of their equipment models for their clients.