Digital Design of DWPT Technique on FPGA for Power System Harmonics Estimation

Real-time detection and quick amplitude estimation of power system harmonics are significant issues among various power quality indices in the electrical power distribution system. In this paper digital design of the discrete wavelet packet transform (DWPT) technique is presented for accurate and fast amplitude estimation of power system harmonics in real-time applications. The DWPT technique provides a uniform frequency bandwidth of the power supply waveform at the output end. Consequently, the input signal's frequency spectrum is evenly divided into eight equal frequency bands that give accurate amplitude estimation of the harmonics by the proposed technique. Generally, the DWPT technique gives the output in the form of coefficients, so the proposed digital architecture incorporates arrangements for giving root mean square values directly. The DWPT based wavelet decomposition tree has been developed with the help of MAC-based wavelet filters that gives an advantage in saving the number of multipliers compared with the polyphase structure on the FPGA board. Also, the downsampling operation performed at each decomposition level reduces the wavelet coefficients by half which minimizes the computations involved in it. The digital design of the proposed DWPT technique has been implemented on the Xilinx Artix-7 FPGA AC-701 board with Xilinx System Generator (XSG)/Vivado design suite 18.3. Thereafter, standard experimental test signals have been generated on the MATLAB/Simulink software platform to validate the robustness and accuracy of the proposed digital hardware under various experimental test conditions, proving their applicability in real-time estimation of power system harmonics.