Real-time power quality waveform recognition with a programmable digital signal processor

Power quality (PQ) monitoring is an important issue to electric utilities and many industrial power customers. This paper presents a DSP-based hardware monitoring system based on a recently proposed PQ classification algorithm. The algorithm is implemented with a Texas Instruments (TI) TMS320VC5416 digital signal processor (DSP) with the TI THS1206 12-bit 6 MSPS analog to digital converter. A TI TMS320VC5416 DSP starter kit (DSK) is used as the host board with the THS1206 mounted on a daughter card. The implemented PQ classification algorithm is composed of two processes: feature extraction and classification. The feature extraction projects a PQ signal onto a time-frequency representation (TFR), which is designed for maximizing the separability between classes. The classifiers include a Heaviside-function linear classifier and neural networks with feedforward structures. The algorithm is optimized according to the architecture of the DSP to meet the hard realtime constraints of classifying a 5-cycle segment of the 60 Hz sinusoidal voltage/current signals in power systems. The classification output can be transmitted serially to an operator interface or control mechanism for logging and issue resolution.