Building Automation Systems (BAS) offer a powerful means for energy optimization, fault detection, and improved occupant comfort. However, in practice, many operations and maintenance (O&M) teams encounter ongoing challenges, such as manpower shortages or turnover that severely limit BAS effectiveness. Consequently, buildings miss out on energy savings opportunities, and risk unplanned system failures that disrupt operations. Project ARIA (Artificial Responsive Intelligent Assistant) will field-demonstrate an artificial intelligence (AI)-driven “Virtual BAS Operator” that could help overcome some of these challenges by automatically sensing system conditions and automatically implementing optimized responses. The goals of this project are to verify that autonomous optimization can:

• Autonomously detects faults
• Makes timely adjustments
• Continuously improve system performance

Project ARIA is an AI virtual BAS operator employed to directly address Department of War's (DoW) BAS challenges. An AI-driven virtual BAS operator closes a gap in current BAS practice: turning automated detection into immediate corrective action. Instead of waiting for alarms to flag faults, the AI operator optimizes operation, dynamically adjusts settings, and prevents alarms before they occur. The core innovation is the ability to automatically correct, improve and predict rather than simply report. Its focus on continuous optimization, fault response, and maintenance forecasting addresses two fundamental O&M problems: unattended BAS systems and limited technician resources. The technology is layered onto the existing BACnet infrastructure and needs only an outbound encrypted link. The pre-trained model is fine-tuned with one month of site data and is bounded by a safety layer that verifies each action against site-specific limits. The project team will analyze energy, comfort, and work-order impacts; success is defined as automated execution of nearly all routine commands, significant reductions in both energy use and reactive maintenance, and transparent operation that base technicians trust and accept.

Widespread implementation of virtual BAS operator technology could significantly reduce heating, ventilation, and air conditioning energy consumption at DoW installations. While automated fault correction reduces unplanned service workload, facility managers would also benefit from fewer vendor callouts and smoother integration into O&M and performance contracts. Beyond cost, the technology could also restore the performance of under-utilized BAS assets, improve space conditions critical to mission readiness, and deliver auditable fault logs that simplify compliance and reporting, without adding personnel or replacing existing controls. For facility managers who lack access to a full-time BAS operator, this AI-based virtual approach could maintain optimal system functionality and energy efficiency without adding personnel. (Anticipated Project Completion - 2028)