【作者】 滕小波；

【导师】 耿相铭；

【作者基本信息】 上海交通大学 ， 信号与信息处理， 2009， 硕士

【摘要】 由于电网中非线性器件和设备的广泛使用,电力系统中产生了大量的高次谐波,本文主要研究并设计实现电力系统中电力谐波参数的实时检测。确切掌握电网中谐波的运行状况,有利于防止谐波的危害、维护电网的安全运行。本文首先介绍了当前电力谐波参数检测的快速傅立叶变换方法,FFT算法技术比较成熟、易于实现,但算法本身很难做到同步采样和整周期截断,由此会造成频谱泄漏和栅栏效应,影响谐波检测结果。论文引入加窗插值FFT算法进行修正,对加窗插值FFT算法进行了较为深入的分析,并给出较详细的数学推导计算过程。其次,较详细介绍了电力谐波检测系统的DSP+CPLD硬件平台,通过采用锁相环实现硬件同步采样,保证严格同步和减小频谱泄漏;采用TMS320C6713B浮点DSP实现加窗插值FFT算法的运算,实时检测电力谐波的参数;采用CoolRunner-II系列XC2C512 CPLD控制系统的外围器件。再次,简要介绍系统前端信号调理、数据采集和外围模块硬件设计部分,以及CPLD逻辑控制部分,包括PLL锁相跟踪、ADC数据采集、本地温度监控、LCD液晶显示,以及CPLD与DSP的外部存储器接口(EMIF)数据通信等模块。重点介绍了DSP系统软件设计部分,包括DSP芯片配置、DSP外设初始化、增强型直接存储器访问(EDMA)、乒乓缓冲传输等底层软件设计,以及算法的设计实现。DSP软件开发采用芯片支持库CSL配置外设、方便代码移植,采用DSP/BIOS实时多任务操作系统管理DSP的线程优先级调度,包括EDMA硬件中断HWI、信号处理算法任务TSK等,它们之间通过信号量保持线程间的同步和通信。最后使用Matlab仿真分析了矩形窗(信号截断)和非整周期采样对FFT算法的影响,并分析比较窗函数和加窗正弦信号的频谱,表明Hanning窗或Blackman窗、整周期采样可以有效地减小频谱泄漏。随后在硬件平台上验证了加窗插值FFT算法的正确性和有效性,表明加Blackman窗的FFT插值算法在非整周期采样时仍然具有较高的精度,幅值误差为e-4数量级,相位误差为e-3数量级。仿真及硬件调试结果表明,算法的精度和实时性都能很好地满足电力系统谐波分析和谐波检测的实际要求。更多还原

【Abstract】 Because of large-scale employment of non-linear devices and equipments, power system generates a lot of harmonics. The main purpose of the paper is to research, design and realize electric harmonic parameter real-time measurement of power system. Mastering harmonic state accurately is in favor of preventing harmonics harming power grid, maintaining power grid safe running state.At first, Fast Fourier Transform method is introduced for current electric harmonic parameter measurement because it is mature and realizable. However it is difficult to perform synchronized sampling and integral period truncation, and the results will be disturbed by spectrum leakage and fence effect caused consequently. So it is necessary to correct FFT algorithm by adding window and introducing interpolation algorithm. The paper analyzes the interpolating windowed FFT algorithm in depth, and gives accurate mathematic deduced process and calculated formula.Next, DSP & CPLD hardware platform is presented for the electric harmonic measurement system, which employs PLL to realize hardware synchronized sampling to ensure accurate synchronized sampling and diminish spectrum leakage, employs TMS320C6713B floating-point DSP to realize interpolating windowed FFT algorithm and measure harmonic parameters, and employs CoolRunner-II XC2C512 CPLD to control peripheral devices logically.Then, the schematic diagram about front-end signal conditioning, data acquisition and peripheral module is introduced. Logic control of CPLD includes PLL frequency trace module, ADC data acquisition module, thermometer I2C bus module, LCD display module, and CPLD communicating with DSP external memory interface (EMIF) module. DSP system software design is emphasized, which includes DSP chip configuration, peripheral initialization, enhanced direct memory access (EDMA) ping-pong buffering transmission. DSP software development employs Chip Support Library (CSL) to configure on-chip peripherals, and employs real-time multi-task operating system DSP/BIOS to manage and schedule DSP threads, including EDMA hardware interrupt (HWI) and signal processing algorithm task (TSK) maintaining synchronization and communication between threads by semaphore. The hardware platform ensures data acquired by ADC flow into DSP internal memory through CPLD.At last, the paper introduces, simulates and analyzes rectangle window’s (signal is truncated) and non-integer-period sampling’s disturbance to FFT algorithm, compares and analyzes window function’s and windowed sine signal’s spectrum using Matlab. The results show adding Hanning window or Blackman window, integer-period sampling can diminish spectrum leakage effectively. Finally, interpolating windowed FFT algorithm is realized on DSP & CPLD hardware platform. And it shows adding Blackman window and interpolating FFT algorithm has high precision even non-integer-period sampling, and amplitude error is e-4 magnitude and phase error is e-3 magnitude. The simulation and hardware debug show that harmonic measurement theory in this paper is correct; hardware platform, system bottom software and algorithm design is also correct. And it meets the requirement of power system harmonic analysis for high-precision and real-time situation. The whole system has a valuable application for power industry.更多还原