【作者】 孙金凤；

【导师】 王学伟；

【作者基本信息】 北京化工大学 ， 计算机应用技术， 2008， 硕士

【摘要】 随着变电站自动化技术的不断进步,巨量数据的传输与存储已成为电力系统中一个非常重要且亟待解决的问题,而解决这一难题的最为有效的方法就是对监测数据进行压缩。鉴于小波变换优良的时频特性,近年来提出的基于小波变换的数据压缩算法不仅获得了较高压缩倍数,而且还能精确地重构原始信号,但目前来说,该类算法仍存在一些理论和技术上的问题影响了实际应用,如一代小波压缩算法计算量大、不利于硬件和实时实现,二代小波Euclidean分解不唯一、运算量减小有限且提升过程存在破坏小波滤波器正交性的问题等。针对这一现状,本文对基于小波变换的电力数据压缩解压算法及压缩数据的远程传输技术进行了深入地理论研究和实验验证。主要工作如下:1、首先整体介绍了数据压缩技术在电力系统中的应用方法,阐述了数据压缩的原理,引入了压缩算法的评价标准,完成了电力信号数学模型的建立,并采用Matlab编程对信号模型加以实现。2、对现有的阈值压缩算法进行了改进,首次提出了一种纪录非零小波系数的位图压缩新算法,相比传统方法,该法在同等大小的均方误差下,获得了更高的压缩倍数。3、深入研究了一代小波滤波器组的整数提升过程,在此基础上,首次将整数小波变换与改进的阈值压缩算法相结合,提出了整数小波阈值压缩算法,有效解决了一代小波压缩算法实现复杂、运算速度慢、变换结果为浮点数等问题,使算法在重构信号的能量恢复系数和均方误差以及压缩算法的实时性三方面都要优于一代小波压缩算法。4、针对电力数据特点,将Ray-Period无损压缩算法引入到整数小波阈值压缩算法中,提出了一种高效的基于整数小波变换的Ray-Period压缩算法,该算法对小波变换得到的低频系数采用Ray-Period压缩,在不增加均方误差的条件下,使整数小波阈值压缩算法压缩倍数进一步提升。5、对一代小波滤波器结构和系数进行优化,提出了一种高效的整数算法小波变换的构造及实现方案,并将其应用于电力系统数据的压缩,从根本上解决了一代及二代小波应用于电力数据压缩时存在的浮点运算问题,大大提高了压缩算法的实时性,并尝试将该压缩算法扩展应用到图像数据的压缩领域,初步探讨了灰度及真彩色图像的压缩方法。6、分析比较当前远程传输的方式,选用了GPRS无线通信网络来传输电力系统监测数据,大大提高了数据传输的灵活性。设计了基于GPRS技术的远程数据传输方案,并从GPRS通信模块选择、网络接入方式及网络传输协议等方面详细阐述了该方案,在此基础上还完成了数据传输协议上层应用协议的设计,进一步增强了数据传输的可靠性。7、完成了基于GPRS的远程数据传输系统的软硬件开发平台的搭建。其中监测终端采用基于ARM920T内核的S3C2410A型开发板,并经开发板串口与MC55 GPRS通信模块相连,操作系统选用Windows CE。应用EVC4.0开发工具C/C++编程语言设计开发了具有友好图形界面的数据压缩、解压及GPRS数据传输程序,其中GPRS数据传输采用多线程编程技术实现,并进行了大量的网络传输测试,结果证实了本文传输程序的可靠性和稳定性。数据的解压缩主要是对接收到的数据进行小波反变换重构出原始信号,这一部分没有单独讲述,其原理和流程图在讨论数据压缩时一并列出。文中大量的实验结果证实了本文算法的正确性和有效性。更多还原

【Abstract】 With the advancement of transformer substation automatization technology, the amount of data gathered by such monitoring devices has become huge in size. Hence tremendous amount of data transmission and storage has become an issue of utmost importance and need to be solved urgently in power system. The most effective method to resolve this problem is to compress the monitored data. Considering the excellent features of wavelet transform in the time and frequency domain, the data compression algorithm based on wavelet transform was proposed in recent years. This algorithm could not only obtain high compression multiple, but also reconstruct original signal accurately. But according to the present research, there were still some theoretical and technical problems that limited the practical applications of this algorithm. For instance, the compression algrithm based on first generational wavelet transform demanded large amount of computation, which made it not feasible to hardware and real-time implementation; when using second generational wavelet transform, the Euclidean decomposition was not one and only, the reduce of computation was limited and the orthogonality of wavelet filter bank was weakened during the process of lifting.Considering the problems above, the electric power data compression and decompression algrithms based on wavelet transform and the remote transmission technology of compressed data were deeply researched in theory and tested with mass experiments in this paper, the main tasks are below:1、Firstly, the method to apply data compression technology in electric power system was introduced. The principle of compression and the evaluation criteria of compression algorithms were discussed. And the mathematical models of electric power signals were established.2、A novel bitmap compression algorithm to record nonzero wavelet coefficients was proposed for the first time. Compared with traditional threshold value algorithm, the proposed algorithm can obtain higher compression multiple when the reconstructive signal with the same mean square error.3、On the basis of analyzing integer lifting process in depth, an improved integer wavelet threshold compression algorithm was proposed for the first time. It can effectively settle the problems, such as the complex realization, the low speed and the floating-point resulting filtered outputs, which are brought by the compression algrithm based on first generational wavelet transform. Compared with first generational wavelet compression, the performance of new approach is enhanced at three aspects, which are the energy recovery coefficient, the mean square error of reconstructive signal, and real-time property of compression algorithm.4、Considering the features of power system data, a Ray-Period compression algorithm based on integer wavelet transform was proposed, where the low-frequency data was processed by lossless Ray-Period algorithm. The algorithm further increases the compression multiple of integer wavelet threshold compression with the same mean square error.5、An efficient algorithm to build integer arithmetic wavelet transform (IAWT) was proposed by improving the traditional wavelet filter structure and coefficients, which solves the problem of the floating-point computations essentially. By applying IAWT to compress data in power system, the computing speed of algrithm is greatly increased. Then the compression method of gray and true color image was researched preliminaryly.6、Through compared the present ways of remote transmission, GPRS was chosen to transmit the monitoring data in power system, which greatly increased the flexibility of data transmission. Then the remote transmission scheme based on GPRS was proposed and demonstrated in detail from the selection of GPRS module, network access and transmission protocol. Based on the above, the application level protocol was designed, which further enhances the reliability of transmission.7、The software and hardware development platform of remote data transmission scheme based on GPRS was built, among which the monitoring terminal was an embedded development board using S3C2410A processor with ARM920T kernel and Windows CE operation system, the MC55 module was connected to the board via serial communication interface. Using EVC4.0 and C/C++, the programs that consist of data compression, decompression and GPRS data transmission were designed and realized, a friendly graphical interface was given and multithreading technology was adopted in GPRS transmission. Mass of network transmission testing result confirms the reliability and stability of our transmission programs.The compressed data receiving could be reconstructed with the aid of inverse wavelet transform. So the section of data decompression wasn’t discussed solely in this paper, its principles and flow charts had been introduced together with data compression. The experimental results prove that the algorithms proposed in this paper are correct and effective.更多还原