RMS Current Measurement Using Noise-Decorrelated Sample Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional RMS current measurement systems suffer from high errors at lower current levels due to auto-correlated noise, which is not effectively reduced by averaging, leading to inaccurate results.
Innovation Solution
A noise-decorrelated RMS current algorithm that multiplies each current measurement sample by a delayed version of itself, followed by a frequency-dependent magnitude correction filter to reduce noise and correct for frequency-dependent attenuation, thereby improving the accuracy of RMS current measurements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional auto-correlation algorithm is used to compute RMS current, then measurement coverage is maintained across full current range, but measurement precision deteriorates at lower current levels due to auto-correlated noise
Solution Approach 1:
The patent extracts and removes the auto-correlated noise component from the RMS calculation by using a different computational approach. Instead of auto-correlating each sample with itself (which creates noise correlation), the algorithm uses a cross-correlation approach where current samples are multiplied with delayed versions of themselves, breaking the noise correlation while preserving the signal measurement capability across the full current range.
Solution Approach 2:
The patent changes the computational parameters of the RMS algorithm by introducing a delay factor (one sample period) in the multiplication operation. This parameter change transforms the auto-correlation operation into a cross-correlation operation, which fundamentally alters the noise characteristics from auto-correlated to uncorrelated, thereby improving measurement precision at low current levels while maintaining full-range coverage.
2Reliability
If auto-correlation algorithm is used, then RMS current can be calculated across full current range, but noise reduction through averaging becomes ineffective
Solution Approach 1:
The patent extracts the harmful auto-correlation property from the noise and removes it by using a cross-correlation algorithm. By multiplying current samples with delayed versions rather than with themselves, the algorithm eliminates the noise correlation that prevents effective averaging, thereby enabling noise reduction through averaging while maintaining reliability across the full current range.
Solution Approach 2:
The patent introduces a periodic delay element (one sample period) into the calculation algorithm. This periodic shift breaks the direct auto-correlation relationship and transforms the noise characteristics, allowing averaging operations to effectively reduce noise while maintaining measurement reliability across all current levels.
3Measurement precision
If conventional I2-samples accumulator is used with 240A maximum range, then full-scale measurement capability is achieved, but measurement error increases greatly at lower current levels
Solution Approach 1:
The patent substitutes the conventional mechanical/mathematical auto-correlation operation with a cross-correlation operation that uses delayed sample multiplication. This substitution fundamentally changes how the I2-samples accumulator processes data, replacing the noise-amplifying auto-correlation mechanism with a noise-reducing cross-correlation mechanism that maintains full-scale capability while dramatically improving lower-end measurement accuracy.
Solution Approach 2:
The patent changes the operational parameters of the accumulator by introducing a time delay (one sample period) into the multiplication operation. This parameter modification transforms the accumulator's behavior from amplifying auto-correlated noise to effectively averaging uncorrelated noise, thereby reducing measurement error across the full current range while particularly improving lower-end accuracy.
Data Source
AI summary
Systems and methods are provided for improving the operation of a computer or other electronic device that utilizes root-mean-square (RMS) measurements, e.g., RMS current measurements, by reducing error in the RMS measurement. A series of measurement samples are received at a processor, which executes a noise-decorrelated RMS algorithm including: calculating a current-squared value for each measurement sample by multiplying the measurement sample by a prior measurement sample in the series (rather by simply squaring each measurement sample as in conventional techniques), summing the current-squared values, and calculating an RMS value based on the summed values. The processor may also execute a frequency-dependent magnitude correction filter to correct for frequency-dependent attenuation associated with the noise-decorrelated RMS algorithm. The calculated RMS value has a reduced error, particularly for lower-end current measurements, which may improve the operation of the computer or electronic device that utilizes the RMS value.