In the landscape of Industry 4.0, predictive maintenance has evolved beyond vibration analysis and thermal imaging. A new technology is emerging as an essential tool for asset monitoring: the online acoustic camera. But what exactly is it, and how does it work under the hood?
This article provides a technical exploration of online acoustic cameras—their core components, working principles, and how they integrate with modern industrial systems to deliver continuous, intelligent fault detection.
An online acoustic camera is a permanently installed, intelligent monitoring device that continuously visualizes sound sources in real time, enabling automated fault detection and diagnostics in industrial environments.
Unlike handheld acoustic cameras used for periodic spot checks or manual troubleshooting, online versions are engineered for 24/7 fixed installation—mounting on walls, poles, or integrated with robotic systems—to provide uninterrupted surveillance of critical assets such as transformers, pipelines, switchgear, and rotating machinery.
At the heart of every online acoustic camera lies a sophisticated fusion of hardware and software:
Modern online acoustic cameras utilize MEMS (Micro-Electromechanical Systems) microphone arrays, typically consisting of 64 to 128 channels. Compared to traditional electret microphones, MEMS offer:
Superior consistency: Matched phase and sensitivity across all channels
Compact form factor: Enabling smaller, lighter industrial enclosures
Wide frequency response: Typically covering 2 kHz to 65 kHz, spanning both audible and ultrasonic ranges
Environmental durability: Better tolerance to temperature variations and vibration
A high-resolution optical camera (typically 1080p or higher) captures the visual scene. The acoustic data is overlaid onto this video feed, creating an intuitive visual representation of sound sources.
The magic of acoustic imaging lies in beamforming—a signal processing technique that combines signals from multiple microphones to determine the direction of sound sources.
Key technical aspects:
Delay-and-sum beamforming: The most common algorithm, applying time delays to align signals from each microphone before summation
Frequency-dependent resolution: Lower frequencies require larger arrays for precise localization; higher frequencies enable sharper spatial resolution
Real-time processing: Modern DSPs and FPGAs process 130k samples per second across all channels to generate sound maps at 25-30 FPS
The output is a sound map—a color overlay (typically red for high intensity, blue for low) superimposed on the video image, showing exactly where the sound originates.
Online acoustic cameras are optimized for three primary industrial applications:
In high-voltage equipment, insulation degradation causes partial discharges that emit ultrasonic energy. Acoustic cameras can:
Classify PD types: Corona discharge (surface), floating discharge (gap), and surface discharge (along insulation)
Locate the exact source: Pinpoint discharge points within switchgear, transformers, or cable terminations
Monitor trends: Track PD activity over time to predict failure
Pressurized gas leaks generate broadband ultrasonic noise. Advanced systems can:
Detect leaks down to 0.035 l/min at 120KPa from 0.3m
Estimate leak rate and cost: Using proprietary algorithms based on pressure, distance, and acoustic intensity
Localize in noisy environments: Filter out background industrial noise
Rotating machinery emits characteristic acoustic signatures. Online monitoring can detect:
Bearing wear and lubrication failure
Cavitation in pumps
Valve leakage and steam trap failure
What transforms an acoustic camera from a simple sensor into a diagnostic tool is onboard intelligence:
All acoustic data is processed locally on the device:
No reliance on central servers: Minimizes latency and network bandwidth
Instant alerts: Fault detection in milliseconds
Privacy and security: Sensitive acoustic data never leaves the facility
Hertzinno's systems are trained on over 5 million industrial acoustic signatures, enabling:
Pattern recognition: Distinguish between partial discharge, mechanical noise, and environmental sounds
Type prediction: Identify specific fault categories (e.g., corona vs. floating discharge)
Anomaly detection: Flag deviations from normal operating baselines
Advanced AI models create acoustic "fingerprints" for each asset. When a fault occurs, the system matches the signature against known patterns, providing instant diagnosis.
Online acoustic cameras are designed as IIoT devices, supporting standard industrial protocols:
Ethernet/HTTP: Primary interface for configuration and data retrieval
RTSP streaming: Real-time video and sound map delivery to any standard viewer
WebSocket: Real-time API for integration with SCADA, DCS, and custom dashboards
Compact, lightweight designs (approx. 900g to 1.6kg) allow mounting on:
Autonomous Mobile Robots (AMRs) for flexible inspection routes
Unmanned Ground Vehicles (UGVs) for hazardous area inspection
Fixed gantries for production line monitoring
Models like the HZ-FA-371T add a 640×512 resolution thermal imaging module, enabling:
Correlation of acoustic events with temperature anomalies
Dual-validation of partial discharges (acoustic signature + heat generation)
Comprehensive asset health monitoring from a single device
Industrial deployment demands ruggedness:
Protection ratings: IP65/IP66 for dust and water ingress protection
Temperature range: Operation from -20°C to 70°C, storage from -30°C to 70°C
Hazardous area certification: Ex ib IIC T4 Gb (gas) and Ex ib IIIC T80°C Db (dust), suitable for Zone 1/2 and Zone 21/22 environments
| Aspect | Traditional Methods | Online Acoustic Camera |
|---|---|---|
| Coverage | Periodic spot checks | Continuous 24/7 |
| Fault detection | Visual/thermal only | Acoustic + optional thermal |
| Intermittent faults | Often missed | Captured automatically |
| Integration | Manual reporting | SCADA/DCS/API-ready |
| Localization | Approximate | Visual sound map overlay |
In the era of Industry 4.0, online acoustic cameras serve as an essential predictive maintenance tool, transforming sound into actionable insights. By continuously monitoring the acoustic spectrum—from audible mechanical noise to ultrasonic partial discharges—these intelligent devices enable:
Prevention of unplanned downtime through early fault detection
Enhanced safety by identifying hazards before they escalate
Optimized asset performance through data-driven maintenance planning
As sensor fusion advances and AI models grow more sophisticated, the online acoustic camera will become as ubiquitous in industrial monitoring as thermal imaging and vibration analysis are today.
Contact Hertzinno for a technical consultation or to request a demo of the HZ-FA-371 and HZ-FA-371T series.
