← 返回
光伏发电技术 多物理场耦合 ★ 5.0

多串光伏升压变换器中受电缆不确定性影响的直流母线电容在线退化故障预测

Online Degradation Fault Prognosis for DC-Link Capacitors in Multistring-Connected Photovoltaic Boost Converters Subject to Cable Uncertainties

作者 Geye Lu · Qiuqiong Lin · Dayong Zheng · Pinjia Zhang
期刊 IEEE Journal of Emerging and Selected Topics in Power Electronics
出版日期 2024年11月
技术分类 光伏发电技术
技术标签 多物理场耦合
相关度评分 ★★★★★ 5.0 / 5.0
关键词 多串光伏发电 直流母线电容 故障预测 退化指标 电缆不确定性
语言:

中文摘要

针对多串光伏系统中各直流支路升压变换器的直流母线电容退化故障问题,提出一种基于母线电压和单串电缆电流的综合故障预测方法。通过构建三维平面y=f,结合去耦电缆参数,分别估计电容的等效串联电阻Rx和电容值Cx,并提出融合Rx与Cx⁻¹变化趋势的新型退化指标D,提升评估灵敏度。仿真与实验验证表明,该方法对多串电缆不确定性及工况变化具有强鲁棒性,参数估计最大误差为1.6%,D指标分辨率达1.18 dis/mΩ,且无需直接测量电容电压或电流,易于实现。

English Abstract

In the application of multistring photovoltaic (PV) power generation, each dc string has a boost converter and the degradation fault of its dc-link capacitor can cause power outage. A comprehensive fault prognosis technology is proposed using the measurable busbar voltage and per-string cable current. The challenge is to avoid the effect of multistring cable uncertainties. First, a 3-D plane y = f is optimized with the information of the series resistance ( R_x ) and capacitance ( C_x ) of dc-link capacitor. The mathematical proof is presented that based on y = f , R_x and C_x are estimated separately by decoupling the cable parameter. Second, a novel degradation indicator D is proposed that combines increasing trends of both R_x and C_x^-1 . The D-based evaluation sensitivity is improved compared with the single estimated parameter. Simulation verifies that the D-based performance is robust to multistring cable uncertainties and varying operating conditions of PV module and load. Online experiments upon the grid-connected platform clarify that the maximum error of parameter estimation is 1.6% and the D-based evaluation resolution relative to R_x is 1.18 dis/m . The breakthrough lies in the mitigation of uncertain cable effect using neither voltage nor current across dc-link capacitors. Convenient implementation and robust performance are emphasized.
S

SunView 深度解读

该直流母线电容在线退化预测技术对阳光电源SG系列多串光伏逆变器和ST储能变流器具有重要应用价值。通过仅测量母线电压和单串电流即可实现ESR和电容值的解耦估计(误差≤1.6%),无需额外传感器,可直接集成到现有MPPT控制架构中。融合Rx与Cx⁻¹的退化指标D(分辨率1.18 dis/mΩ)能显著提升iSolarCloud平台的预测性维护能力,在多串并联拓扑中对电缆参数不确定性的强鲁棒性尤其适用于大型地面电站和工商业分布式场景。该方法可延伸至PowerTitan储能系统的DC/DC变换器健康管理,降低因电容失效导致的系统停机风险,提升全生命周期可靠性。