In critical industries such as power generation, petrochemicals, and metallurgy, large fans for ventilation, blowing, and cooling are indispensable. The health of their key components, like impellers and bearings is paramount to operational safety and efficiency.
Case Details
In critical industries such as power generation, petrochemicals, and metallurgy, large fans for ventilation, blowing, and cooling are indispensable. The health of their key components, like impellers and bearings is paramount to operational safety and efficiency. Traditional manual inspections often fail to detect subtle, developing faults, leading to unexpected downtime and costly repairs.
The Challenge: Proactive Fan Health Management
The primary goal is to move from reactive to proactive maintenance. The system must reliably detect and diagnose common fan issues, including:
- Imbalance or Cracked Impeller Blades: Often manifested as elevated vibration at the fan's rotational speed (1x RPM).
- Bearing Wear: Identified by specific high-frequency vibration signatures.
- Structural Loosening: Detected through abnormal harmonic components in the vibration spectrum.
- General Mechanical Faults: Revealed by changes in the overall acoustic noise profile.
The Solution: A Smart Monitoring System Architecture
Our solution is built on a robust hardware and software foundation, designed for harsh industrial environments.
Core Components:
- Sensors: Three IEPE accelerometers are strategically placed: one at the impeller center (measuring both axial and radial vibration), one on the fan bearing housing, and one on the fan casing to monitor overall vibration. An IEPE microphone is also mounted on the casing to capture acoustic data.
- Signal Acquisition: The Y37 IEPE Input Module is the perfect interface for these sensors. Its four independent channels simultaneously acquire high-fidelity signals with 24-bit resolution. It provides the necessary constant current excitation (3.5mA) for the sensors and features a wide bandwidth (0.6 Hz – 20 kHz), essential for capturing both low-frequency imbalance and high-frequency bearing tones.
- Edge Processing & Intelligence: The SmartFan Controller acts as the brain of the system. This powerful industrial computer, equipped with a multi-core ARM processor and a neural processing unit (NPU), hosts the entire application. It receives the digitized data from the Y37 module, runs diagnostic algorithms, and handles communication.
System Workflow:
The IEPE sensors convert physical vibration and sound into electrical signals. These signals are fed into the Y37 module, which conditions and digitizes them. The Y37 module, plugged directly into the SmartFan Controller, streams the high-resolution data. The SmartFan Controller then processes this data in real-time, converting raw vibrations into actionable insights.
Intelligent Diagnostics at the Edge
Running on the SmartFan Controller's Linux-RT operating system, custom diagnostic software performs continuous analysis. The logic includes:
- Imbalance Detection: Monitoring the vibration spectrum for a dominant peak at the fan's fundamental rotating frequency.
- Bearing Fault Detection: Analyzing the high-frequency range for known characteristic defect frequencies associated with the specific bearing.
- Loosening Detection: Identifying the presence of multiple harmonics (2x, 3x RPM, etc.) in the spectrum, which often indicates a loose component.
- Acoustic Anomaly Detection: Triggering an alert if the overall sound pressure level exceeds a predefined threshold, suggesting a potential issue.
This edge-based analysis eliminates the need for constant high-volume data transmission, allowing for near-instantaneous local alerts and control actions.
Operational Benefits and User Interface
The system provides a clear, web-based dashboard built with tools like Node-RED, giving plant personnel an immediate overview of fan health. The interface displays key metrics such as overall vibration velocity, bearing condition index, and noise levels. A traffic-light system (Green/Amber/Red) offers an at-a-glance status.
When a potential fault is detected, the system generates an immediate alert. Maintenance teams can access detailed spectral data to diagnose the root cause, schedule repairs during planned outages, and prevent catastrophic failures. Furthermore, by tracking vibration trends over time, the system enables predictive maintenance, forecasting when a component will need service.
