风机油液磨粒在线监测系统的研究与设计

发布时间：2018-06-19 16:50

本文选题：数字信号处理 + 磨粒监测　；参考：《北京交通大学》2017年硕士论文

【摘要】：摘要:在风力发电机运行过程中,其内部的部件会相互接触摩擦产生磨损颗粒,这些磨粒随着润滑油在油路中循环。通过监测风力发电机油液中的磨粒信息可以确定合适的换油时机以及判断风力发电机的磨损程度,从而增加风力发电机的安全性及延长其使用寿命。本文针对磨粒信号检测存在的难点,提出了磨粒信号检测算法,研究并设计了一套风机油液磨粒在线监测系统,实现对风力发电机油液中的磨粒在线监测。本文研究的对象为风机油液磨粒在线监测系统,分别从系统的设计需求、磨粒信号检测算法、系统的总体设计三个方面展开研究。(1)提出了风机油液磨粒在线监测系统的总体设计框架,由磨粒传感器、磨粒信号数据处理、上位机三部分组成。磨粒信号数据处理部分需要具备对磨粒数据的采集、数据处理、数据存储、数据通信等功能,根据各项功能,提出相应的方法策略,并根据系统设计的需求进行选择。(2)分析了磨粒信号检测存在的难点,主要包括高频信号干扰、基线漂移、尖峰干扰、惯性假峰、波形截断等问题,这些问题的存在会造成磨粒多检和漏检。利用多项式拟合移动平滑滤波算法对信号进行滤波,消除高频信号干扰;利用寻峰算法解决基线漂移、尖峰干扰、惯性假峰的问题,从采样信号中寻找到磨粒波形的波峰和波谷;提出两类分类算法解决波形截断的问题,根据波峰和波谷的信息对磨粒信号分类,提取出真实磨粒数据信息。(3)从软件和硬件上设计并实现了风机油液磨粒在线监测系统。在软件上,实现了通过DMA的方式搬移采样数据、利用I2C总线将数据存储进铁电存储器中、通过Modbus通信协议实现与上位机的通信等功能;硬件方面,设计了以DSP为核心的系统硬件电路。本文通过实验对磨粒信号检测算法中的参数进行设置、通过数据连续性测验验证了 DMA搬移数据的可行性、通过滤波测试验证了滤波算法的有效性、通过磨粒检测实验验证了磨粒检测算法对于直径大于100μm的铁磁磨粒和直径大于500μm的非铁磁磨粒具有较好的检测效果、通过上位机监测的结果验证了整个系统的有效性。
[Abstract]:Absrtact: during the operation of wind turbine, the internal components of wind turbine will contact each other to produce wear particles, which circulate in the oil circuit with lubricating oil. By monitoring the information of abrasive particles in the oil of wind turbine, we can determine the right time of oil exchange and judge the degree of wear of wind turbine, so as to increase the safety of wind turbine and prolong its service life. Aiming at the difficulties in the signal detection of abrasive particles, this paper puts forward an algorithm for detecting the signal of abrasive particles, studies and designs a set of on-line monitoring system of wear particles in fan oil, which can realize the on-line monitoring of abrasive particles in the oil of wind turbine. The object of this paper is the on-line monitoring system of oil abrasive particles in fan, which is based on the design requirements of the system, the signal detection algorithm of abrasive particles, The overall design framework of the on-line monitoring system for oil abrasive particles in fan is proposed, which consists of three parts: abrasive particle sensor, abrasive particle signal data processing and upper computer. The data processing part of abrasive particle signal needs to have the functions of collecting, processing, storing, communicating and so on. According to each function, the corresponding method and strategy are put forward. According to the requirements of the system design, the paper analyzes the difficulties of the wear particle signal detection, including high frequency signal interference, baseline drift, spike interference, inertia false peak, waveform truncation and so on. The existence of these problems will result in multiple detection and missed detection of abrasive particles. The polynomial fitting moving smoothing filter algorithm is used to filter the signal to eliminate the high frequency signal interference, and the peak seeking algorithm is used to solve the problems of baseline drift, peak interference and inertial false peak. Two kinds of classification algorithms are proposed to solve the problem of waveform truncation. According to the information of wave peak and trough, the wear particle signal is classified. The real abrasive particle data information is extracted. The on-line monitoring system of fan oil abrasive particles is designed and implemented from the software and hardware. In software, the data is moved by DMA, the data is stored in ferroelectric memory by I2C bus, and the communication with host computer is realized by Modbus protocol. The system hardware circuit based on DSP is designed. In this paper, the parameters of abrasive signal detection algorithm are set through experiments, the feasibility of DMA moving data is verified by data continuity test, and the validity of filtering algorithm is verified by filter test. The testing results show that the algorithm is effective for ferromagnetic abrasive particles with diameter greater than 100 渭 m and non-ferromagnetic abrasive particles with diameter greater than 500 渭 m, and the effectiveness of the whole system is verified by the monitoring results of upper computer.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM315

【参考文献】