光伏热斑图像有效区域分割算法研究

根据光伏热斑图像的特征提出一种基于局部统计特征的有效区域分割算法。以图像局部标准差衡量图像局部灰度一致性,标记图像中局部灰度一致性高的区域,提出基于矩形尺寸及矩形内部有效区域比例的区域判别方法对背景区域进行去除,应用形态学运算处理所得区域,得到最终分割结果。实验结果表明:该算法可准确、有效地对光伏热斑图像有效区域进行分割,且效果优于边缘检测和区域生长算法。     

基于迁移学习的光伏组件鸟粪覆盖检测

该文基于无人机光伏组件可见光图像采集,提出一种基于迁移学习的光伏组件鸟粪检测方法.方法首先基于掩膜区域卷积神经网络(Mask-RCNN)对光伏组件边界进行框选,再利用迁移学习策略,构建光伏组件鸟粪缺陷检测模型,实现鸟粪智能检测.利用上述方法,实现光伏组件鸟粪覆盖检测准确率为96.75％.     

基于卷积神经网络与ECOC-SVM的输电线路异物检测

输电线路悬挂异物会引发输电线路单相接地、相间短路等停电事故,因此本文提出一种基于卷积神经网络与ECOC-SVM的输电线路异物检测方法.首先,本文构建气球、风筝、塑料和鸟巢4种输电线路异物图像数据集;然后采用Otsu自适应阈值分割、形态学处理等方法提取感兴趣区域;再利用DenseNet201提取感兴趣区域的特征;最后对E...     

基于数字图像处理的风电机组叶片裂纹损伤识别方法研究

为实现风电机组叶片损伤检测的高效化、智能化、便捷化,研究一种基于数字图像处理技术的风电机组叶片裂纹损伤识别以及裂纹类型判断和特征参数提取的方法。以无人机采集的风电机组叶片图像为研究对象,通过对比灰度化、滤波、阈值分割等图像处理步骤的多种算法,对形态学处理方法进行改进,首先选用平均值法对叶片图像进行灰度处理,其次使用中值滤波对图像进行去噪处理,再次使用Otsu阈值分割以实现裂纹区域的分割,然后基于改进的形态学方法提取出完善的叶片裂纹损伤区域,最后基于连通域原理完成裂纹区域的框取。基于上述算法设计风电机组叶片裂纹损伤识别系统以实现叶片裂纹图像检测的可视化处理、裂纹类型判断及裂纹特征参数提取等功能。结果表明,该系统对于风电机组叶片裂纹损伤检测具有可靠的识别精度,识别准确率为85%,实现了风电机组叶片裂纹损伤的自动识别与特征参数提取,提高了叶片裂纹损伤的检测效率。     

基于改进Unet的分布式光伏建筑物高精度分割方法

针对屋顶光伏资源评估中难以准确高效地获取建筑物屋顶区域的问题，该文提出一种基于Unet的FPN＿AttentionUnet语义分割网络，用于实现建筑物屋顶的高精度自动提取。该网络融合Soft-Attention注意力机制和双层特征金字塔FPN以提取准确的语义信息，精细化分割结果。Soft-Attention注意力机制用于处理和连接编码部分与解码部分的特征图；双层特征金字塔FPN融合解码部分不同尺度的特征图来获取不同尺度的特征信息。采用无人机获取苏州某区域上空的建筑物数据集和武汉大学WHU公开数据集分别进行训练，训练结果表明：与Unet、AttentionUnet、FPNUnet网络相比，该文提出的FPN＿AttentionUnet在建筑物外轮廓提取中具有更高的精度，有效提高边缘提取效果。在自制数据集中类别像素准确率CPA达95.56%，平均交并比MIoU达91.10%，在WHU公开数据集中分割效果同样优于其他对比网络，所提算法能够有效提升建筑物外轮廓边缘的分割精度。最后以河海大学常州校区为例，利用提出的算法从无人机图像中分割建筑物，评估指定区域的光伏发电量与光伏组件安装潜力。     

图像处理在光伏局部阴影中的应用

以局部阴影条件下的光伏组件为研究对象,通过对光伏组件不同位置进行遮挡的实验,得到不同阴影条件下的阴影面积,并对所得数据进行对比分析。利用图像增强、边缘检测等图像处理技术来识别阴影区,其中由于光伏板上表面条纹的影响,采用中值滤波去噪,且通过连通区域法计算出光伏板上阴影面积和百分比。实验结果表明在实验空间d_1=65 cm,d_2=150 cm内,边缘检测法可准确分割阴影,快速准确地计算出阴影面积,不会因膨胀使得阴影面积增大,致使数据不准确。精确的阴影面积更有利于研究光伏的输出特性,降低光伏阵列最大功率点追踪的难度。     

基于灰度聚类的光伏红外热斑图像分割方法

针对光伏红外热斑检测问题，提出一种曲线拟合结合图像聚类的热斑红外图像处理方法。首先，将图像灰度变换处理后，采用高斯最小二乘拟合确定聚类中心点；然后，针对传统模糊C均值噪声鲁棒性较差的特点，加入邻域空间影响以及用核距离代替欧式距离的模糊C均值算法对热斑图像进行聚类；最后，根据拟合图进行灰度多阈值分割。实验结果表明：该方法能将光伏组件损坏程度进行量化，区域进行分层，抑制红外图像噪声，提高检测热斑效率的同时，分割准确率可达86%以上。     

基于红外图像和改进MobileNet-V3的光伏组件故障诊断方法

为了提高光伏系统的可靠性和性能，提出一种基于红外图像和改进MobileNet-V3的光伏组件故障诊断方法。首先，分析开源光伏组件缺陷图像及其存在的问题；然后，基于存在的问题，对光伏组件红外缺陷图像进行图像增强、数据增强处理，使红外图像满足图片可用性及样本数量丰富性的要求；最后，对基本MobileNet-V3网络进行改进，实现光伏组件故障分类。实验结果显示：与传统CNN、基础MobileNet-V3相比，所提故障分类方法不仅准确率高、诊断速度快，且对各种故障类别的识别率高，具有较好的实用性和应用价值。     

融合视觉特征的光伏组件语义分割模型研究

针对光伏组件红外图像的分割问题,使用MobileNetv2作为DeepLabv3+的主干特征提取网络并使用位置通道注意力模块减少背景干扰,引入混合条带池化对ASPP模块进行优化,帮助模型进一步捕获全局和上下文信息。针对检测困难的屋顶光伏组件设计DeepLabv3-T网络,在上述改进的基础上融入纹理信息进行选择性背景抑制,实现光伏组件的精确分割。在PV＿large和PV＿roof数据集上进行实验证明该文方法优于现有技术,DeepLabv3-T相较于DeepLabv3+,mIoU值分别提高了2.74%和10.53%。此外,设计消融实验表明各个改进模块的有效性。     

山地光伏电站的近场阴影模拟研究

结合复杂山地地形条件下光伏组件的布置方案，基于PVsyst软件和Helios 3D软件建立了山地光伏电站三维近场阴影仿真模型，以模拟计算近场阴影遮挡损失修正系数；另外根据项目建设区域内实际的光伏组件倾角和光伏方阵方位角，结合光伏电站朝向替代计算，得到近场阴影遮挡损失修正系数；最后对3种计算方案得到的近场阴影遮挡损失修正系数与该项目实际值进行对比。研究结果表明：由于山地光伏电站项目建设区域地形中的坡度、坡向变化对光伏组件倾角、光伏方阵方位角存在影响，导致山地场景下的近场阴影遮挡损失修正系数比平地场景下的更大。提出的Helios 3D模拟计算方案和公式计算方案同样适用于其他山地光伏电站，可以对特定地形条件下光伏发电系统的阴影情况进行模拟，以评估光伏组件倾角、光伏方阵方位角对近场阴影遮挡损失修正系数的影响，可为精细化评估山地光伏电站中光伏组件布置方案对近场阴影遮挡损失修正系数的影响程度提供合理的分析、研究方案。     

基于S-V特性分析的晶硅光伏组件阴影遮障诊断

针对广泛使用的晶硅光伏组件,通过Simulink建立太阳电池双二极管精确仿真模型,对实际应用中最常见的光伏组件阴影遮挡故障进行多种工况的仿真验证。根据I-V曲线拐点、台阶、曲线下积分面积（S）下降的特征,提出一种基于S-V曲线特性的光伏组件阴影遮挡故障的在线诊断方法。该方法建立S-V曲线,根据S-V曲线分叉点位置可判断光伏组件遮挡情况,通过整体积分面积进而判断遮挡比例。对温度、辐照度进行折算,使该方法在全工况下适用。结合光伏组件功率优化器验证该诊断方法有较高的准确率,并且可准确地判断阴影遮挡面积,具有很高的实际应用价值。     

Performance enhancement of partially shaded solar PV array using novel shade dispersion technique

Solar photo voltaic array (SPVA) generates a smaller amount of power than the standard rating of the panel due to the partial shading effect. Since the modules of the arrays receive different solar irradiations, the P-V characteristics of photovoltaic (PV) arrays contain multiple peaks or local peaks. This paper presents an innovative method (magic square) in order to increase the generated power by configuring the modules of a shaded photovoltaic array. In this approach, the physical location of the modules in the total cross tied (TCT) connected in the solar PV array is rearranged based on the magic square arrangement pattern. This connection is done without altering any electrical configurations of the modules in the PV array. This method can distribute the shading effect over the entire PVarray, without concentrating on any row of modules and can achieve global peaks. For different types of shading patterns, the output power of the solar PV array with the proposed magic square configuration is compared with the traditional configurations and the performance is calculated. This paper presents a new reconfiguration technique for solar PV arrays, which increases the PV power under different shading conditions. The proposed technique facilitates the distribution of the effect of shading over the entire array, thereby, reducing the mismatch losses caused by partial shading. The theoretical calculations are tested through simulations in Matlab/ Simulink to validate the results. A comparison of power loss for different types of topologies under different types of shading patterns for a 4×4 array is also explained.     

A Novel Auto-Scaling MPPT Algorithm based on Perturb and Observe Method for Photovoltaic Modules under Partial Shading Conditions

Maximum Power Point Tracking (MPPT) controller is required in a solar photovoltaic (PV) system to deliver the maximum power to load from PV module. This paper proposes a novel stepped MPPT method to realize a simple MPPT controller, which can track the real maximum power point (RMPP) even under partial shading conditions. The proposed algorithm is started by scanning the characteristic curve of the PV modules to detect the global maximum power point and then the algorithm will be switched to the conventional P&O algorithm to track the true maximum power point. The obtained simulation results, using Power electronic simulation software (PSIM), are compared with those found using the P&O method to confirm the performance of our proposed MPPT method even under non-uniform solar irradiation.     

Study of bypass diodes configuration on PV modules

A procedure of simulation and modelling solar cells and PV modules, working partially shadowed in Pspice environment, is presented. Simulation results have been contrasted with real measured data from a commercial PV module of 209 Wp from Siliken. Some cases of study are presented as application examples of this simulation methodology, showing its potential on the design of bypass diodes configuration to include in a PV module and also on the study of PV generators working in partial shading conditions.     

Forecasting photovoltaic array power production subject to mismatch losses

The development of photovoltaic (PV) energy throughout the world this last decade has brought to light the presence of module mismatch losses in most PV applications. Such power losses, mainly occasioned by partial shading of arrays and differences in PV modules, can be reduced by changing module interconnections of a solar array. This paper presents a novel method to forecast existing PV array production in diverse environmental conditions. In this approach, field measurement data is used to identify module parameters once and for all. The proposed method simulates PV arrays with adaptable module interconnection schemes in order to reduce mismatch losses. The model has been validated by experimental results taken on a 2.2 kW_p plant, with three different interconnection schemes, which show reliable power production forecast precision in both partially shaded and normal operating conditions. Field measurements show interest in using alternative plant configurations in PV systems for decreasing module mismatch losses.     

基于并联光伏组件外特性的最大功率点跟踪方法

当阴影条件变化时,并联光伏组件的全局最大功率点(MPP)会随之改变.为了实现太阳能发电最大化,要求最大功率点跟踪(MPPT)方法始终能实时而准确地锁定住并联光伏组件的全局MPP.不同阴影条件下并联光伏组件会呈现不同的外特性特征,如多阶梯的电流电压特性以及多峰值的功率电压特性.基于此现象,该文提出一种基于并联光伏组件外特...     

A comprehensive MATLAB Simulink PV system simulator with partial shading capability based on two-diode model

This paper proposes a comprehensive MATLAB Simulink simulator for photovoltaic (PV) system. The simulator utilizes a new two-diode model to represent the PV cell. This model is known to have better accuracy at low irradiance level that allows for a more accurate prediction of PV system performance during partial shading condition. To reduce computational time, only four parameters are extracted for the model. The values of R_p and R_s are computed by an efficient iteration method. Furthermore, all the inputs to the simulators are information available on standard PV module datasheet. The simulator supports a large array combination that can be interfaced to MPPT algorithms and power electronic converters. The accurateness of the simulator is verified by applying the model to five PV modules of different types (multi-crystalline, mono-crystalline, and thin-film) from various manufacturers. It is envisaged that the proposed work can be very useful for PV professionals who require simple, fast, and accurate PV simulator to design their systems. The developed simulator is freely available for download.     

A high efficiency photovoltaic module integrated converter with the asymmetrical half-bridge flyback converter

A module integrated converter (MIC) for a photovoltaic (PV) cell is important part of power conditioning system (PCS). It performs maximum power point tracking of a PV cell to generate the power as much as possible from solar energy. There are several methods for connection between the PV modules and the MICs. In order to avoid partial shading effects, converter-per-module approach was proposed. The MIC that performs maximum power point tracking (MPPT), if it is low efficiency, is no use. The MIC whose output is connected to the output of PV module was proposed for high efficiency. However, there are some problems. In this study, an asymmetrical half-bridge flyback converter is proposed instead of the original flyback converter with same method to solve the problems. The proposed MIC was built to verify the performance. The new topology using soft switching technique showed good performance for the efficiency. At the higher power, the efficiency of the proposed converter is higher than existing converter.     

基于图像识别的高温熔渣流量测量方法研究

熔渣流量是影响高温熔渣离心粒化效果的关键因素。为了测量熔渣流量,提出了一种流量测量方法,利用高速摄影机拍摄液柱流动图像,采用边缘检测、阈值分割等方法,获得液柱直径及流速,计算流量,采用常温甘油溶液作为测试对象进行实验,研究该方法测量精度的影响因素。结果显示,长曝光时间、合理的拍摄帧率、短物距有利于提高测量精度;同时,感兴趣区域（ROI）的位置为距管口2~5倍管径处、ROI长度为1/3~2/3倍管径时,测量精度较高。通过对实际高温熔渣流动过程的检测,其测量误差约9.0%,证明了该方法的实用性。     

Calculating the shading reduction coefficient of photovoltaic system efficiency using the anisotropic sky scattering model

The front-row shading reduction coefficient is a key parameter used to calculate the system efficiency of a photovoltaic (PV) power station. Based on the Hay anisotropic sky scattering model, the variation rule of solar radiation intensity on the surface of the PV array during the shaded period is simulated, combined with the voltage–current characteristics of the PV modules, and the shadow occlusion operating mode of the PV array is modeled. A method for calculating the loss coefficient of front shadow occlusion based on the division of the PV cell string unit and Hay anisotropic sky scattering model is proposed. This algorithm can accurately evaluate the degree of influence of the PV array layout, wiring mode, array spacing, PV module specifications, and solar radiation on PV power station system efficiency. It provides a basis for optimizing the PV array layout, reducing system loss, and improving PV system efficiency.