Cutting HVAC Fault Costs Without Replacing the BMS

Background

Our client is a regional commercial property management firm operating 18 Class A office buildings totaling approximately 2.8 million square feet across Connecticut, New Jersey, and southern New York. Their building stock spans construction vintages from 1987 to 2019, with a heterogeneous mix of BMS vendors — Automated Logic, Siemens Desigo, and Johnson Controls Metasys — across the portfolio.

The firm's maintenance team of 24 technicians had been managing reactive cycles driven by tenant complaints and monthly energy utility reports. HVAC fault detection was manual: technicians reviewed BMS alarm logs during site visits, which occurred on a scheduled weekly or biweekly basis depending on building size.

The Challenge

• $28K average annual energy waste per building from undetected faults — simultaneous heating and cooling, valve leak-by, and fan-curve drift went undetected between scheduled site visits
• Reactive maintenance model — three chiller failures in the prior 18 months required emergency replacement costing $180K, $94K, and $66K respectively; all three had precursor anomalies detectable in the BMS time-series data that no one was monitoring continuously
• No cross-portfolio visibility — each BMS ran its own vendor-proprietary cloud dashboard; comparing energy performance or fault rates across buildings required manual export and spreadsheet assembly
• Capital replacement budget pressure — the board had approved a $2.2M BMS replacement plan for 6 buildings; the client wanted to know if the existing systems could be extended with better analytics before committing

Our Solution

• AiTBMS analytics overlay — deployed a vendor-agnostic cloud historian on top of the existing Automated Logic, Siemens, and Johnson Controls BMS network; no field hardware changes, no controller replacements; integration via BACnet/IP and OPC-UA at the building edge
• Fault detection and diagnostics (FDD) library — 64 fault rules covering AHU, chiller, cooling tower, VAV, and boiler plant; rules adapted to each building's actual sequence of operations rather than generic templates
• Dollar-weighted alert prioritization — all active faults ranked daily by estimated energy waste at the client's actual utility rates; maintenance dispatch queue driven by financial impact, not alarm severity code
• Per-chiller predictive models — trained on 24 months of historical BMS data; flags chiller performance deviation 7–14 days before it becomes a measurable energy event or approaches trip thresholds
• Portfolio-level executive dashboard — single-pane view of all 18 buildings with ranked fault lists, energy benchmarking by building vintage and size, and monthly sustainability metrics for tenant ESG reporting

Results

• $340K in avoided repair costs in year one — predictive model flagged chiller short-cycling at two buildings 11 days before projected failure; proactive service calls cost $18K combined vs. estimated $220K emergency replacement
• 19% reduction in energy spend across the portfolio — annualized in months 7–12 vs. prior-year baseline; simultaneous heating-and-cooling faults alone accounted for 8.4% of total reduction
• Zero field hardware changes — the $2.2M BMS replacement plan was deferred; the board approved a 3-year extension of existing controllers with AiTBMS as the analytics and alerting layer
• Maintenance dispatch efficiency +34% — technician site visits now driven by dollar-weighted fault queue rather than scheduled rounds; same team covering 18 buildings with 3 fewer emergency call-outs per month
• Tenant ESG reporting automated — monthly energy intensity and carbon metrics now generated automatically per building and per floor; three tenants renewed citing improved sustainability transparency