A Study on Integrated Performance Verification Methodology for Energy/Carbon Reduction and Quantitative Calculation of ALFUS-based Autonomy Assessment Model in AI-based Buildings

The adoption of AI-based autonomous control technologies in Building Energy Management Systems (BEMS) is accelerating to achieve carbon neutrality in the building sector. However, existing international Measurement and Verification(M&V) protocols, which rely on static baselines, have limitations in accurately evaluating the dynamic control characteristics of AI and the intermittent contributions of renewable energy. Furthermore, the lack of standards for quantitatively judging the intelligence level of systems makes objective evaluation of the technology difficult. Therefore, this study proposes an integrated calculation model that links energy savings and carbon emission reductions into a single formula and introduces a Renewable Energy Alternative Supply Reflection Coefficient(fr) to prevent double counting. In addition, to evaluate the technical sophistication of AI systems, this study redefines the concept of NIST’s ALFUS(Autonomy Levels for Unmanned Systems) optimized for the building control environment. By presenting detailed score distribution criteria(0-10 points) for three core axes(Human Independence, Mission Complexity, and Environmental Difficulty), this study establishes a framework to quantitatively verify the performance of AI BEMS.