The detection of transformer noise requires a combination of acoustic measurement techniques, signal analysis, and standard compliance to ensure accuracy and reliability. Below are the common detection steps and methods:
1. Pre-Measurement Preparation
Environmental Requirements:
Background noise should be at least 3 dB lower than the transformer noise (ideally 10 dB lower) to avoid interference.
Measurements should be taken in calm weather conditions (no rain, wind speed < 5 m/s) to minimize environmental effects.
Equipment Calibration:
Use a sound level meter compliant with standards (e.g., IEC 61672-1 Class 1) or a noise analyzer, and calibrate it before testing.
Transformer Operating Conditions:
The transformer should operate at rated voltage and load to ensure stable noise generation (e.g., no-load, 50% load, full-load conditions).
2. Measurement Point Arrangement
Standard Distance:
According to GB/T 1094.10 or IEC 60076-10, measurement points are typically placed 0.3 m or 2 m from the transformer's reference sound-emitting surface, at half the transformer's height (but not lower than 0.25 m).
For large transformers, multiple points should be evenly distributed around the perimeter (e.g., 3-5 points per side).
Simplified Method:
If close-range measurement is impractical, points can be placed at 1-3 times the transformer's height, but the positions must be documented.
3. Noise Measurement Methods
A-Weighted Sound Level (dBA):
Use the A-weighting network on the sound level meter (simulating human hearing) to measure equivalent continuous sound pressure level (Leq).
Frequency Domain Analysis:
Perform Fourier Transform (FFT) or 1/3-octave band analysis to identify dominant frequency components (e.g., 100 Hz, 200 Hz, and other power-frequency harmonics).
Focus on low-frequency noise (<500 Hz), as transformer noise is typically dominated by 100 Hz and its harmonics.
Sound Pressure vs. Sound Intensity Methods:
Sound Pressure Method: Simple but susceptible to environmental reflections.
Sound Intensity Method: Effective in eliminating background noise interference (requires a sound intensity probe).
4. Data Processing and Analysis
Background Noise Correction:
If the difference between background noise and transformer noise is ΔL:
If ΔL > 10 dB, background noise can be ignored.
If 3 dB < ΔL ≤ 10 dB, apply correction formulas (e.g., subtract background noise influence).
If ΔL ≤ 3 dB, the measurement is invalid.
Spectrum Comparison:
Compare measured spectra with transformer design parameters to identify abnormal noise sources (e.g., core vibration, loose windings, cooling fan noise).
5. Common Noise Source Identification
Core Magnetostriction: 100 Hz, 200 Hz, and other even harmonics, typically producing a "humming" sound.
Winding Vibration: High-frequency noise (>1 kHz), which may worsen with increased load current.
Cooling System: Fan or oil pump noise (broadband high-frequency components), verifiable by comparing measurements with the cooling system on/off.
6. Standards and Limits
Reference Standards:
International: IEC 60076-10 Determination of Sound Levels.
China: GB/T 1094.10-2022 Power Transformers – Part 10: Determination of Sound Levels.
Industry Requirements: For example, urban substation noise limits (daytime ≤55 dBA, nighttime ≤45 dBA).
Example Limits:
500 kV transformer: No-load noise typically ≤75 dBA (measured 2 m from the enclosure).
7. Advanced Techniques
Acoustic Imaging: Uses a microphone array to locate noise hotspots (suitable for complex structures).
Vibration Correlation Analysis: Simultaneously measures vibration signals (e.g., accelerometers) to analyze noise-vibration relationships.
Online Monitoring: Installs fixed noise sensors for long-term noise trend tracking.
Key Considerations
Avoid reflective surfaces (e.g., walls, equipment); use sound-absorbing materials if necessary.
Record operating parameters (load, voltage, temperature) during measurements.
For abnormal noise (e.g., sudden high-frequency noise), combine with oil chromatography analysis, vibration testing, and other diagnostics.
By following these methods, transformer noise levels can be systematically evaluated, providing a basis for fault diagnosis or noise reduction measures.