高级检索

面向盐田高湿腐蚀环境的光伏电站直流侧绝缘故障在线定位

Online Location of DC-Side Insulation Faults in Photovoltaic Power Plants under High-Humidity and Corrosive Salt-Field Environment

  • 摘要: 针对盐田高湿环境光伏直流绝缘监测受杂散电容、行波混叠、泄漏电流掩盖微弱故障信号的痛点,本文提出在线故障定位方案。借助Tikhonov正则化反演介电弛豫参数,剥离湿度扰动并解耦绝缘电阻;依托行波折射系数奇异谱熵与随机森林辨识故障组串,消除行波混叠干扰;采用自适应卡尔曼滤波结合窗斜率补偿,把背景泄漏抑制至5μA以内。现场实测可甄别电化学迁移、缝隙腐蚀等隐性故障,定位误差中位数2.3~3.1m,误报9.7%、漏报4.4%,达成米级故障在线定位预警。

     

    Abstract: Aiming at interference from stray capacitance, overlapping traveling waves and leakage current masking faint faults in DC insulation monitoring of salt-field PV stations, an online localization method is proposed. Tikhonov regularization retrieves dielectric relaxation parameters to decouple insulation resistance from humidity; singular spectral entropy of traveling-wave refraction coefficient plus random forest identifies faulty strings; adaptive Kalman filtering with sliding-window slope compensation restrains background leakage below 5 μA. Field tests detect hidden corrosion faults with median localization error 2.3~3.1 m, false alarm rate 9.7% and missing alarm rate 4.4%, realizing meter-level online early warning.

     

/

返回文章
返回