FPGA-based intelligient recognition system that identifies real time power quality problems

In this thesis, intelligient, embedded system-based power quality recognition systems have been developed to identify power quality problems. The intelligient recognition systems that are developed by using FPGA have an online structure and aim at identifying and classifying real-time power quality problems. In the designs that are developed, the recognition systems, which simultaneously determine power quality events and power quality disturbances in power systems, have been created by using FPGA. Three different intelligient recognition approaches exist in the developed online recognition systems. These approaches are signal processing techniques that include Fourier transform, wavelet transform and Hilbert Huang transform. Decision tree algorithm has been used for classification in these three different that are suitable for online recognition. A graphical interface software is created in computer to monitor the results obtained from the recognition systems visually and save them. The other approach realized in this thesis is the offline recognition applications that have been created for power quality events and power quality disturbances. In these approaches, intelligient recognition applications have been realized by using feature vectors obtained from online recognition systems. The least squares support vector machines, extreme learning machines and artificial neural networks classifiers have been used in offline applications. A prototypical power system has been created in laboratory for the purpose of testing FPGA-based online intelligient recognition systems, determining accuracy rates and assessing their performances. Different power quality event types have been created within broad parameter intervals on this model. When the results obtained from the online and offline recognition systems developed within the framework of this thesis are examined and compared with the literatüre, it can be seen that hardware and software designs have a very efficient, fast and high performance. Key Words: Power Quality Problems, FPGA, Signal Processing, Smart Recognition Systems