本文關(guān)鍵詞： 變壓器 在線監(jiān)測(cè) 神經(jīng)網(wǎng)絡(luò)算法 故障診斷　出處：《西南交通大學(xué)》2017年碩士論文　論文類(lèi)型：學(xué)位論文

【摘要】：智能變電站的發(fā)展對(duì)變壓器智能化水平提出了更高的要求,變電站變壓器在線監(jiān)測(cè)與故障診斷技術(shù)也在逐步改善和提高,變壓器智能化水平的提高不僅能提高電網(wǎng)工作人員的效率,更對(duì)電力系統(tǒng)的穩(wěn)定運(yùn)行起著重要作用。對(duì)智能變電站中變壓器的運(yùn)行情況進(jìn)行實(shí)時(shí)在線監(jiān)測(cè),能夠及時(shí)發(fā)現(xiàn)變壓器運(yùn)行過(guò)程中的問(wèn)題,在線故障診斷系統(tǒng)能夠預(yù)測(cè)變壓器潛在故障,在變壓器發(fā)生故障前就及時(shí)找出問(wèn)題,避免由于變壓器故障導(dǎo)致的電力系統(tǒng)運(yùn)行不穩(wěn)定甚至是停電等造成的危害�；贐P神經(jīng)網(wǎng)絡(luò)算法,采用遺傳算法優(yōu)化后作為變壓器故障診斷的方法,通過(guò)相應(yīng)的硬件和軟件實(shí)現(xiàn),完成了變壓器在線監(jiān)測(cè)與故障診斷系統(tǒng)的設(shè)計(jì)。通過(guò)對(duì)人工智能診斷方法的研究,提出采用BP神經(jīng)網(wǎng)絡(luò)算法作為變壓器故障診斷的手段,并基于遺傳算法對(duì)BP神經(jīng)網(wǎng)絡(luò)算法進(jìn)行優(yōu)化設(shè)計(jì),在Matlab軟件平臺(tái)上對(duì)算法做了仿真驗(yàn)證,結(jié)果證明,優(yōu)化后的神經(jīng)網(wǎng)絡(luò)算法準(zhǔn)確性得到了提高。以LPC2214CPU的ARM芯片為核心,設(shè)計(jì)了變壓器油中氣體濃度的數(shù)據(jù)采集和傳輸?shù)挠布到y(tǒng),并將uC/OS-Ⅱ?qū)崟r(shí)操作系統(tǒng)進(jìn)行移植,結(jié)合嵌入式軟件系統(tǒng)實(shí)現(xiàn)整體功能,基于LabVIEW虛擬儀器軟件設(shè)計(jì)了上位機(jī)在線監(jiān)測(cè)系統(tǒng),通過(guò)圖形化編程,設(shè)計(jì)了數(shù)據(jù)采集與DGA分析模塊,基于神經(jīng)網(wǎng)絡(luò)的故障診斷模塊和數(shù)據(jù)庫(kù)模塊,完成了變壓器在線監(jiān)測(cè)與故障診斷系統(tǒng)的人機(jī)交互操作界面設(shè)計(jì)。本文所設(shè)計(jì)的變壓器在線監(jiān)測(cè)與故障診斷系統(tǒng),不僅完成了基于遺傳算法改進(jìn)后的BP神經(jīng)網(wǎng)絡(luò)算法的理論驗(yàn)證,并且在分析實(shí)際需求的基礎(chǔ)上,通過(guò)嵌入式系統(tǒng)對(duì)硬件和軟件部分設(shè)計(jì)實(shí)現(xiàn)了具體的數(shù)據(jù)采集、信息傳輸、在線監(jiān)測(cè)、故障診斷、數(shù)據(jù)庫(kù)存儲(chǔ)等功能,完成了變壓器在線監(jiān)測(cè)與故障診斷系統(tǒng)的設(shè)計(jì),為變壓器在線監(jiān)測(cè)與人工智能診斷的實(shí)際應(yīng)用提供了具體的解決方案。
[Abstract]:The development of intelligent substation puts forward higher requirements for the level of transformer intelligence, and the on-line monitoring and fault diagnosis technology of transformer in substation is also gradually improved and improved. The improvement of transformer intelligence level can not only improve the efficiency of power grid workers, but also play an important role in the stable operation of power system. The on-line fault diagnosis system can predict the potential fault of the transformer and find out the problem before the fault of the transformer occurs. Based on BP neural network algorithm, genetic algorithm is adopted as the method of transformer fault diagnosis. Through the realization of corresponding hardware and software, the design of transformer on-line monitoring and fault diagnosis system is completed. Through the research of artificial intelligence diagnosis method, BP neural network algorithm is used as the means of transformer fault diagnosis. The BP neural network algorithm is optimized based on genetic algorithm, and the algorithm is simulated on the Matlab software platform. The results show that the accuracy of the optimized neural network algorithm has been improved. The core of this algorithm is the ARM chip of LPC2214CPU. The hardware system of data acquisition and transmission of gas concentration in transformer oil is designed, and the UC / OS- 鈪，

本文編號(hào)：1535180