Edge AI Safety Intelligence · BESS
Your cooling systemis degrading yourbatteries right now.Your BMS can't see it.
Edge AI predictive maintenance for the BESS systems that existing monitoring misses — HVAC health, environmental conditions, off-gas detection, thermal anomalies. On-site processing with sub-second response.
✓ No cloud dependency for safety-critical decisions. Pilot-ready in 90 days.
65%
of BESS failures originate from operational systems — not cell defects
5–20 min
off-gas warning window before thermal runaway — most sites have no analytics
72%
of BESS failures occur within the first two years of operation
The challenge
Major BESS fires start outside cells.
Your monitoring is looking in the wrong place.
Your BMS watches cells. Cloud analytics model degradation. But the cooling system, the container environment, the gas detection — the systems responsible for real-world fires — have no AI monitoring. EPRI's 2024 analysis is clear: only 11% of failures trace to cell defects. The rest involve cooling, environmental factors, and integration. And 72% of those failures occur within the first two years.
73% of O&M staff report monthly technical issues
01 — MONITORING GAP
65%
of BESS failures · from operational systems, not cells
Your BMS Is Watching the Wrong Layer
Desire → Problem
You want to know the health of every system that keeps your batteries safe — cooling performance, environmental conditions, gas detection, thermal management across the entire container.
Instead, your BMS monitors cell voltage, current, and temperature. Cloud analytics model degradation curves. But the cooling system, the container environment, the gas detection — the systems behind the majority of real-world failures — have no AI monitoring. EPRI found that only 11% of BESS failures trace to cell defects. The remaining 65% involve cooling, environmental factors, control integration, and assembly. You're monitoring the 11%. Nobody is monitoring the 65%.
Transformation
BMS covers cells only. Cloud analytics model degradation. Cooling and environment assumed to work.
↓
Edge AI monitors the battery-adjacent systems — HVAC, environment, gas detection — that existing platforms ignore.
02 — THERMAL DEGRADATION
2×
degradation rate · for every 10°C above optimal
Your Cooling Is Failing Slowly — And Destroying Battery Life
Desire → Problem
You want your thermal management to maintain optimal conditions throughout the asset's life — not just at commissioning, but through years of compressor wear, refrigerant loss, and efficiency decline.
COP degradation is slow and invisible. A system delivering COP 3.5 at commissioning may operate at 2.0 three years later — 75% more electricity for the same cooling. Batteries experience higher temperatures, faster degradation, and the asset owner absorbs higher auxiliary costs — all without a single alarm. On a €50M battery asset, 40% lifespan reduction from thermal mismanagement means €20M+ in accelerated replacement. Your BMS sees the temperature rising. It cannot see that the cause is a slowly failing HVAC system.
Transformation
No COP tracking. No HVAC efficiency trending. Cooling degradation invisible until failure.
↓
Continuous COP monitoring and HVAC health scoring. Detect cooling decline months before it affects battery life.
03 — ENVIRONMENTAL HAZARD
75%+
humidity levels · inside containers during HVAC cycling
Condensation, Humidity, Water Ingress — the Documented Fire Causes
Desire → Problem
You want environmental conditions inside your BESS containers monitored continuously — humidity, dew point, condensation risk on critical surfaces — not assumed to be safe between inspections.
DNV confirmed condensation on electrical components from faulty humidity control as a direct fire cause. Korean ESS investigations found condensation plus dust caused insulation breakdown. Moss Landing burned from water ingress. Research shows humidity inside BESS containers routinely exceeds 75% during HVAC cycling — and nobody measures it. Standard BMS has no humidity sensors. Every HVAC cycle creates humidity spikes that operators never see. This is a documented fire cause hiding in plain sight.
Transformation
No humidity sensors. No condensation monitoring. Environmental conditions unknown between inspections.
↓
Real-time humidity, dew point, and condensation probability on critical surfaces. Detect the mechanism behind documented BESS fires.
04 — DETECTION WINDOW
5–20 min
warning before thermal runaway · most sites have no analytics
The Only Early Warning — Buried in Alarm Noise
Desire → Problem
You want gas detection that provides high-confidence warnings of thermal runaway — ranked by severity, correlated across sensors, with recommended actions and compliance documentation.
Off-gas detection provides 5–20 minutes before thermal runaway — the only proven pre-runaway detection method. But raw sensor data generates false alarms from environmental interference, HVAC cycling, and sensor drift. Without analytics, operators face alarm fatigue or ignore genuine warnings. A 10-minute warning is worthless if buried in daily noise. Most BESS sites have gas sensors installed but no integrated analytics — just binary threshold alarms that cry wolf.
Transformation
Binary gas alarms. High false-positive rate. No pattern analysis. Alarm fatigue.
↓
AI-driven gas pattern recognition with multi-gas correlation and confidence scoring. Fewer alarms, higher trust, faster response.
Running BESS assets without monitoring the cooling system, the container environment, or gas detection analytics? Let's see what edge AI can reveal about your operational blind spots.
Start the conversation →Our solutions
Six applications. One edge platform.
No cloud dependency for safety.
An edge-resident monitoring platform that detects HVAC degradation, thermal anomalies, environmental hazards, and off-gas events — on site, in real time. Custom hardware, embedded AI, and six detection engines purpose-built for the BESS systems your BMS and cloud analytics assume are working.
Prevent fires: Thermal Anomaly + HVAC Power Monitoring
ThermalShield
Thermal & HVAC Anomaly Detection
Continuous monitoring of temperature gradients, hotspot formation, and HVAC power consumption patterns across racks and enclosures. AI distinguishes normal thermal cycling from emerging imbalances — cooling loss, blocked airflow, compressor strain, or localized cell stress. HVAC power trending detects when your cooling draws more energy and delivers less capacity.
ThermalShield catches spatial thermal anomalies that threshold-based alarms miss — hotspots developing 30 minutes before they cascade. HVAC power analysis reveals the efficiency degradation that silently accelerates battery aging without triggering a single BMS alarm. Prevents the compressor failure → cooling loss → thermal stress → safety event chain that caused the Victorian Big Battery fire in 2021.
Protect assets: Gas Detection + Environmental Monitoring
EnviroGuard
Gas Analytics & Environmental Intelligence
Integrated off-gas analytics provide pattern-based thermal runaway early warning — not binary threshold alarms. Environmental monitoring tracks humidity, dew point, and condensation probability on critical surfaces in real time, detecting the exact mechanism behind documented BESS fires cited by DNV and Korean ESS investigations.
Higher-confidence gas warnings with fewer false positives. Elimination of condensation-driven insulation failures that standard BMS makes invisible. Environmental compliance documentation for insurance and NFPA 855. Detects humidity spikes during HVAC cycling, water ingress, and enclosure seal degradation — before they cause short circuits or worse.
Optimize operations: Compressor Health + COP Tracking
CoolingSense
Compressor Health & Cooling Efficiency
Current signature analysis and vibration trending for HVAC compressors detect bearing wear, refrigerant loss, and mechanical imbalance weeks to months before breakdown. COP tracking reveals the slow, invisible efficiency decline — a system at COP 3.5 at commissioning operating at 2.0 three years later, consuming 75% more electricity for the same cooling output.
No unplanned cooling outages. A compressor replacement scheduled for Tuesday instead of an emergency failure at 2 AM on a 38°C Saturday. COP trending makes the invisible decline visible — research shows optimizing cooling reduces maximum cell temperature differences from 31°C to 3.5°C and increases COP fourfold. Extended battery life, reduced auxiliary energy costs, and documented HVAC performance for warranty and insurance.
How it works
Four layers. Edge-resident.
No cloud dependency for safety.
No BMS integration. No SCADA modification. No IT approvals. The platform deploys alongside existing equipment with dedicated sensors, processes everything on site, and delivers actionable alerts — without touching your control systems or OT network.
Layer 1
Sense
Dedicated sensors inside and around BESS containers — temperature arrays for spatial gradients, gas detectors (VOC, CO, CO₂, H₂), humidity and dew point sensors, HVAC power meters, compressor vibration and current sensors. Non-intrusive. OEM-agnostic.
Layer 2
Edge
Promwad-designed edge computing unit at the BESS site. Six detection engines run in parallel with sub-second latency. No internet required for safety-critical decisions. Ruggedized for industrial environments.
Layer 3
Detect
Six anomaly detection engines: thermal gradient, HVAC power, gas pattern, environmental risk, compressor health, COP trending. Each learns baselines specific to the installation. Detects rate-of-change patterns and spatial gradients — not just threshold violations.
Layer 4
Act
Structured alerts with context — what was detected, where, severity, likely cause, recommended action. Integration with SCADA, CMMS, email, SMS via standard APIs. Fewer notifications, higher diagnostic value.
No BMS integration needed
The platform operates independently alongside your existing systems. No control changes, no IT approvals, no OEM coordination.
Protocols: Modbus TCP, MQTT, OPC-UA
Standard industrial connectivity. Integrates with existing SCADA, CMMS (SAP PM, Maximo), email, and SMS workflows.
NFPA 855 / insurance-ready
Edge processing, encrypted transport, compliance-ready monitoring documentation. Designed to support insurance requirements for enhanced BESS monitoring.
Get started
Prove the value on
one asset in 90 days.
We don't ask for a fleet-wide commitment. Pick your most critical BESS asset — and let edge AI prove what your existing monitoring is missing.
01
Technical consultation — 30 min
We learn about your BESS assets — fleet size, OEM configuration, HVAC setup, current monitoring stack. You get a preliminary assessment of monitoring gaps. No commitment.
02
On-site sensor deployment
Our engineers install dedicated sensors and the edge computing unit on one BESS asset — temperature arrays, gas detectors, environmental sensors, HVAC monitors. Days, not months. No BMS integration required.
03
90-day monitoring period
Continuous data collection across all six detection engines. Weekly status reports. At the end: baseline health assessment, detected anomalies, environmental risk profile, HVAC efficiency analysis, and fleet-wide deployment recommendation with ROI model.
Let's discuss your BESS monitoring gaps
Tell us about your storage assets — fleet size, HVAC concerns, current monitoring. We'll prepare a tailored assessment.
Built by
Promwad engineering team
3M+
Devices shipped
20+
Years engineering
500+
Products developed
100+
Engineers across EU
© 2026 Promwad Electronics. All rights reserved.