EC.GAIA — AI Energy Optimization Engine | EC.DATA

EC.GAIA — AI-Powered HVAC & Refrigeration Fault Detection

EC.GAIA is the AI/ML intelligence layer of EC.DATA, launching Q2 2025. It monitors pressure, temperature, and amperage across HVAC and refrigeration systems — using AI to detect the 12 most common faults weeks before catastrophic failure — plus predictive demand forecasting, automated setpoint optimization, and natural-language energy insights. Three sensor streams: Pressure (P), Temperature (T), Amperage (A). Key statistics: 12 HVAC faults detected; 3 sensor streams; AI predictive engine with 94.7% accuracy; 47,000+ faults learned; 2.1 million patterns stored.

AI Capabilities

Predictive demand forecasting with 95%+ accuracy at 15-minute intervals
Automated HVAC setpoint optimization based on weather, occupancy, and tariffs
Natural-language energy Q&A — ask questions about your building's performance
Anomaly classification: equipment fault, occupancy change, or weather event
Energy conservation measure (ECM) impact simulation
Automated monthly performance summaries with AI-generated insights

Technology

An AI copilot trained on energy data, not general chat

EC.GAIA is the AI layer that runs continuously against every telemetry stream from EC.EMS, setpoint from EC.BMS, and invoice inside your EC.DATA tenant. Unlike general-purpose chatbots, GAIA is purpose-built on operational energy data: its reasoning knows what a coincident peak is, how to adjust for degree-days, and when a drop in chiller COP is statistically significant.

GAIA surfaces recommendations automatically — "your AHU-3 minimum fresh-air setpoint is 15 percentage points above design, costing about $420/month" — and also answers natural-language questions like "show me every site where last month's demand charge was in the top decile for its building type".

Every GAIA recommendation is traceable to the underlying data and assumptions, so an energy manager can challenge the logic, adjust the parameters, and only then act. It augments rather than replaces the humans who own the savings.

The 12 most common HVAC & refrigeration faults detected by EC.GAIA

GAIA monitors pressure, temperature, and amperage patterns to detect these critical faults before system failure:

Low Refrigerant (CRITICAL) — refrigerant charge below operating range, detected via suction pressure drop and superheat rise; causes compressor overheating and potential burnout.
Condenser Fouling (HIGH) — debris on condenser coil raises discharge pressure and reduces heat rejection; energy consumption rises 15–30% before failure.
Compressor Valve Fault (CRITICAL) — worn or broken compressor valves reduce pumping efficiency; detected via abnormal pressure ratios and amperage patterns.
Evaporator Icing (HIGH) — ice buildup on evaporator coil blocks airflow; causes suction pressure drop and eventual system lockout.
TXV Hunting (MEDIUM) — thermostatic expansion valve oscillation causes unstable superheat and efficiency loss; detected via suction pressure oscillation pattern.
Liquid Line Restriction (HIGH) — blockage in the liquid line causes flash gas before the expansion device; reduces capacity and efficiency significantly.
Non-Condensables (MEDIUM) — air or moisture in the refrigerant circuit raises condensing pressure and temperature; identified via abnormal pressure-temperature relationship.
Refrigerant Overcharge (MEDIUM) — excess refrigerant causes high discharge pressure and subcooling; can flood the compressor with liquid refrigerant.
Electrical Fault (CRITICAL) — abnormal amperage draw indicates motor, contactor, or capacitor failures; risk of complete system shutdown or fire.
Airflow Issues (MEDIUM) — restricted or insufficient airflow across coils reduces heat transfer; detected via abnormal supply/return temperature differentials.
Sensor Drift (LOW) — calibration error in pressure or temperature sensors causes incorrect readings; can mask real faults or trigger false alarms.
Oil Return Failure (HIGH) — insufficient oil circulation back to the compressor causes lubrication failure and premature compressor wear.

Six specialized AI agents working in concert

GAIA deploys six specialized AI agents that continuously feed intelligence into the platform to make it progressively smarter:

Fault Detector — continuously scans all sensor streams for anomaly signatures across pressure, temperature, and amperage data.
Anomaly Classifier — classifies detected anomalies into specific fault types with confidence scoring and severity assessment.
Maintenance Scheduler — prioritizes repairs based on fault severity, progression rate, and business impact, optimizing technician dispatch.
Energy Optimizer — identifies energy waste from developing faults and calculates cost impact of delayed vs. immediate repair.
Alert Dispatcher — routes enriched alerts to the right people through the right channels with the right urgency level.
Pattern Learner — every fault detected and pattern observed continuously trains the model, making predictions more accurate over time.

Learning statistics: 47,000+ faults learned; 2.1 million+ patterns stored; 94.7% accuracy rate.

Daikin alliance — enriched fault alerts with part numbers

EC.DATA's partnership with Daikin enriches every alert with specific part numbers, pinpoints where the problem is in the system, and suggests potential fixes ranked by likelihood.

Generic alert example: "High Discharge Pressure Detected — Unit: RTU-03, Discharge Pressure 385 PSI (normal 250–300). Action: check condenser and refrigerant levels." A technician must manually diagnose the root cause and find parts.

GAIA + Daikin enriched alert example: "Condenser Fouling — High Discharge Pressure. Unit: Daikin VRV IV Model RXYQ10T. Fault Location: outdoor condenser coil. Root Cause: debris accumulation on condenser fins (92% confidence). Suggested fixes ranked: 1. Clean condenser coils with approved coil cleaner. 2. Inspect condenser fan motor — Part #KFD-420A. 3. Check/replace air filter — Part #BAE101A442."

Three Daikin alliance features: Specific Part Numbers — every alert includes exact Daikin replacement part numbers so technicians arrive prepared. Pinpoint Fault Location — identifies exactly where in the system the problem is occurring (compressor, condenser, evaporator, or controls). Ranked Fix Suggestions — potential fixes ordered by likelihood, reducing diagnostic time from hours to minutes.

"GAIA powers every module, turning our platform into a smarter, more valuable decision-making engine."

— Roberto Flores, CEO, EC.DATA