IEEE Paper review

Review on IEEE paper

Subject: Comb filter for noise reduction

Topic: Improving Dynamic Response of Active Harmonic Compensator using Digital Comb Filter

Harmonic extraction and control is one of the main elements in active power filter design.To improve the dynamic response of the active harmonic controller, a comb-filter is the solution, which can be implemented on a DSP platform. Performance evaluation results for the load transient show a substantial improvement in the dynamic performance of the harmonic compensator compared with the conventional low pass filter-based approaches. The whole system mainly includes voltage source converter and harmonic and reactive power compensator. Resonant current control has also been used widely for reducing the current harmonic distortion in many different distributed generation applications. A resonant harmonic compensator attenuates the current harmonics and is capable of achieving zero steady-state error. The proposed solution in this paper used for improving the dynamic response of a system by using a limited bandwidth compensator with selective unwanted frequency components.

PAPER LINK : http://ieeexplore.ieee.org/document/6872521/

Patent review

PATENT REVIEW

PATENT: United States Patent, SIGNAL PROCESSING SYSTEM AND METHOD FOR
               ENHANCED CASCADED INTEGRATOR-COMB INTERPOLATION FILTER
               STABILIZATION

PATENT NUMBER: US5880973

In this patent, overflow detection circuit monitors the output of the CIC (Cascaded Integrated Comb)interpolation filter for an overflow condition which is indicative of filter instability. When an Overflow Condition is detected, the reset circuitry automatically resets the filter, thereby eliminating the instability. CIC interpolation filters are constructed as a series of ‘N’comb filter circuits (subtractors), followed by a Zero pad circuit, followed by a series of “N” integrator circuits(adders). The comb section of the filter operates at a low sample rate, Whereas the integrator section of the filter operates at a high sampling rate. Typically, the interpolation factor (R) is set between 2 and 10,000. The number of stages of filter increases accuracy. Implementing a CIC interpolation filter generally involves the three requirements that the filter must be reset before use, addition and subtractions must be performed using modulo arithmetic and the Word size in the subtracters, adders, and storage elements must be large enough to accommodate the arithmetic gain of the filter. This circuit checks whether overflow is there or not,if no then it will proceed further and if yes then it will reset filter. This invention detects instabilities in the CIC interpolation filter by continuously monitoring the output samples of the filter.

PATENT LINK: https://www.google.ch/patents/US5880973

DSP PROCESSOR

DSPs are processors or microcomputer whose hardware, software and instruction set are optimised for high-speed numeric processing application. It is capable of handling large mathematical operations easily while using real-time signals for the manipulations of the data. In this experiment, we performed with basic operations of the DSP processor C2000 such as addition, subtraction, bit manipulation, etc. The coding platform used was Code Composer Studio.

FIR filter design using FSM

In this experiment, we have done FIR filter design using Frequency Sampling Method. In this method, impulse response h(n) is obtained using desired frequency response. The given frequency response is sampled at a set of equally spaced frequencies to obtain N samples. Sampling the continuous frequency response Hd(w) at N points essentially gives the N-point DFT of Hd(wk). A Set of samples determined from Hd(w) are identified as Discrete Fourier Transform Coefficients. It is denoted by H(k). The filter coefficients are obtained by IDFT.

Linear phase FIR filter designing using Windowing method

In this experiment, we have designed linear phase FIR filter using windowing method. Different types of windows are used to design FIR filter, which includes Hamming window, Hanning window, Bartlett window, Blackman window and Kaiser window. Here, we used Hanning window for designing of filter. The parameters used for designing the FIR filters are attenuation in stop band(As),attenuation in pass band (Ap),pass band frequency, stop band frequency and sampling frequency. We have written scilab code with reference of designing method done in class. Advantage of the windowing method is it requires minimum amount of computational effort, so window method is simple to implement.

Butterworth Filter

After completing methods of finding output of filter,we moved towards the design part of filter.In this experiment ,we have studied butterworth filter,low pass and high pass.we observed that butterworth filter have monotonic amplitude response in both pass band and stop band.Quick roll-off around the cut-off frequency,which improves with increasing order.Analog LPF poles lie on LHS of s-plane and digital poles lie inside the circle.These conditions denote stability of filter.

Chebyshev filter

Chebyshev filters are analog or digital filters having a steeper roll-off and more passband ripple than Butterworth filters. Chebyshev filters minimize the error between the idealized and the actual filter characteristic over the range of the filter but with ripples in the passband.The Chebyshev filter has a smaller transition region than the same order Butterworth filter, at the expense of ripples in its pass band. Because of the passband ripple inherent in Chebyshev filters, the ones that have a smoother response in the passband but a more irregular response in the stopband.This filter requires less hardware than butterworth filter.

OVERLAP ADD METHOD AND OVERLAP SAVE METHOD

OSM (Overlap Save Method) and OAM(Overlap Add Method) are linear filtering method. In this practical, we have performed OSM and OAM using 13 point input sequence x(n) and 4 point h(n). The obtained output of OSM and OAM method is same for given input sequence.In OAM, the input signal was decomposed into non overlapping blocks and their corresponding y(n) was calculated using linear convolution and then these signals are combined to obtain the final output.
In OSM, the input is decomposed into overlapping blocks and their corresponding y(n) is calculated through circular convolution and then these signals are concatenated to give the final output.

FAST FOURIER TRANSFORM

Fast fourier transform is fast method of computation than discrete fourier transform(DFT).It requires less number of calculation than Discrete Fourier Transform(DFT), which includes real addition and multiplication. FFT uses parallel processing hence it has high operation efficiency.In this experiment , we have taken cases of 4 point and 8 point sequence. The output of FFT is in bit reversed format. we have perfomed this experiment using C program of DITFFT(Decimation in time FFT).

DISCRETE FOURIER TRANSFORM

In this experiment,first a 4 point sequence was taken and it's DFT was found out using C program.Then this same sequence was appended by 4 zeros to form a 8 point sequence and further 4 point sequence was expanded by adding alternate zeros between two samples.It's DFT output was found out using C code of DFT. Magnitude spectrums were plotted for each case.As value of N increases, frequency spacing and approximation error decreases as well as resolution of the spectra increases.Expansion of signal in time domain giver compressed spectrum in frequency domain.

CONVOLUTION AND CORRELATION

our first practical of semester six was convolution and correlation.The program for this practical was given in C language. I have substituted the values of x(n) and h(n) in program .Final output length of y(n) for linear convolution is equal to L+M-1. For circular convolution it is maximum value from L(length of x(n)) and M(length of h(n)).Circular convolution gives aliased output.
Auto correlation of signal is even signal and cross-correlation of input signal with delayed signal is same as advanced auto correlated input signal.Correlation was used to test the degree of similarity between two signals.