A time-frequency signal analysis system for power quality assessment

The purpose of this work is to develop a time-frequency signal analysis system to detect and recognize different kinds of power quality events or disturbances. To achieve our goal, design of computational signal processing methods is being addressed developing time-frequency tools based on signal algebra operators. The use of signal algebra operators permits the formulation of time-frequency algorithms in a computational framework setting, allowing the search for efficient hardware implementations. Time-frequency formulations are being implemented for embedded system applications using digital signal processing (DSP) and field programmable gate array (FPGA) units. Voltage disturbances are responsible for many disruptions in industrial, commercial and residential power supply systems, causing time and monetary losses. It is necessary to identify fast methods of determining when these disturbances are occurring in order to make correct power quality decisions. Automated information processing systems are needed to assess existing problems. Although a lot of work has been done on power line disturbance assessment, new techniques are desired to address new challenging power quality issues, specially in the area of physical security. Time-frequency signal analysis is a well known tool used in fields such as speech, sonar, and radar processing and is now finding its being in power quality assessment. A time-frequency representation is a two-dimensional representation that shows how the spectral content of a given signal changes with time. There is a genuine interest in this kind of representations for the field of power quality signal analysis systems since they provide more information than the typical one dimensional analysis. To achieve the goals of this work an environment is being created using MATLAB to simulate power line signal disturbances such as swells, sags, harmonic distortions, and outages. This environment computes desired time-frequency representations of selected signals to produce two-dimensional time-frequency characterizations of the disturbances. Time-frequency tools such as the discrete short time fourier transform (DSFT), the discrete ambiguity function (DAF), and the discrete Wigner distribution (DWD) are being used in the MATLAB environment- . These tools we are characterizing the voltage disturbances, since the time-frequency representations reveal patterns or properties that are not readily perceptible in one dimension.