THD Measurement Using Narrow Band Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for estimating total harmonic distortion (THD) in electrical distribution systems are inaccurate at low signal levels relative to broadband background noise, leading to increased errors as current decreases, which is problematic for power distribution systems and other applications where precise harmonic content monitoring is crucial.
Innovation Solution
A multi-pass method using a narrow band filtering algorithm, such as the Goertzel algorithm, is employed to filter out broadband noise and accurately determine THD by analyzing signal measurements at specific harmonic frequencies, reducing computational requirements and improving accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If conventional broadband spectrum analysis is used to estimate THD, then the method is simple to implement, but measurement precision deteriorates at low signal levels due to broadband noise contamination
Solution Approach 1:
The broadband spectrum is segmented into narrow frequency bins centered at harmonic frequencies. Instead of analyzing the entire broadband spectrum at once, the method divides it into discrete frequency components using FFT, then selectively processes only the bins corresponding to harmonic frequencies (H1, H2, H3, etc.). This segmentation isolates the harmonic content from broadband noise, improving measurement precision while maintaining computational feasibility.
Solution Approach 2:
The method applies local quality enhancement by focusing computational resources on specific frequency regions of interest (harmonic frequencies) rather than uniformly processing the entire spectrum. Each harmonic frequency bin is individually identified and processed with higher precision, while broadband noise in non-harmonic regions is effectively ignored. This localized approach improves THD measurement accuracy without proportionally increasing overall computational complexity.
2Reliability
If broadband noise is included in THD calculation, then all frequency components are captured, but measurement precision worsens due to noise contamination in the numerator
Solution Approach 1:
The method extracts only the relevant harmonic frequency components from the broadband spectrum while discarding the broadband noise portions. By identifying and isolating the energy at specific harmonic frequencies (H1, H2, H3, etc.) through FFT bin analysis, the method separates the useful signal information from the contaminating noise. This extraction process removes the harmful broadband noise contribution from the THD calculation numerator, improving accuracy while maintaining reliability of the measurement.
3Use of energy by moving object
If signal level is low relative to broadband noise, then power consumption is reduced, but measurement precision deteriorates as noise dominates the broadband spectrum
Solution Approach 1:
By segmenting the broadband spectrum into discrete frequency bins and selectively analyzing only those bins corresponding to harmonic frequencies, the method maintains measurement precision even when the overall signal level is low. The segmentation allows the system to ignore the dominant broadband noise and focus computational effort on the weaker harmonic signals, enabling accurate THD measurement at low power consumption levels where conventional broadband methods would fail.
Data Source
AI summary
A method of measuring the total harmonic distortion in an electrical distribution system. The distribution system is sampled regularly to generate a set of data. The data is filtered using a narrow band filtering algorithm to measure the energy in fundamental and other harmonic frequencies. Due to the filtering, the energy in harmonic frequencies can be measured without interference from broadband noise, which provides an improvement in the measurement of total harmonic distortion at low current or voltage levels. A method is provided to sum the energy in identified frequencies in a multi-pass configuration, such that only a subset of all monitored frequencies are filtered and summed in each pass, with the balance being filtered and measured in subsequent passes. After all subsets are measured, the total harmonic distortion is calculated.