Implementation Of Real-time Data Compression For Power Quality Events In Embedded System

In power systems, power quality disturbances cover a wide spectrum. A higher sampling frequency makes rapid volume increase of monitoring data concerned with them in a short time, whose huge capacity causes much trouble in saving and transmitting. Thus, how to effectively achieve real-time data compression is important and of practical significance. Various power quality events and data compression techniques were firstly discussed in the thesis. A Dual-CPU embedded system was designed for real-time data compression of power quality. It turned out that the system realized the targets satisfactorily.Based on the theory of wavelet transform and features of power quality detection data, a hard threshold method was selected as the main compression algorithm in DSP. The parameter settings are flexible, including threshold value, sampling rate, and compression grades. They can be either the system default settings or inputs by users through the keyboard and display platform provided by ARM. The work is divided by as follows: DSP was responsible for signal collection, wavelet transform and hard threshold compression. Meanwhile, reading keys, displaying LCD, and communications with the Ethernet were implemented by ARM. This scheme can make full use of DSP's fast data-processing capacity and ARM's powerful peripheral management capabilities. The scheme of Dual-CPU simplified the design and debug work, and can be easily upgraded. DSP and ARM were communicated by Dual-port RAM which can provide an effective arbitration mechanism.The system was tested on powerful RTDS. The results were satisfactory, and the indexes, including operating speed, compression ratio, and reconstruction signal distortion and so on, met the requirements of real-time data compression. This system had some advantages, including small size, low cost, friendly interface and effectiveness of real-time data compression. If implemented in power system, it will bring not only a great convenience for data measurement and detection, but also social and economic benefits.The thesis also gives some suggestions on further improvement in hardware and software. They could serve as references for next stage of the project.