Node Document

建筑光伏组件的积灰检测与输出功率预测模型研究

Research on Dust-accumulation Detection and Output Power Prediction Model of Building Photovoltaic Modules

To avoid encountering 504 error when downloading big files,please take CAJ download or PDF download by chapter as your first priority.

【Author in Chinese】 从光杰

【Supervisor】 于军琪

【Author's Information】 西安建筑科技大学, 智能建筑, 2018, 硕士

【Abstract in Chinese】 随着绿色建筑、零能耗建筑的发展及光伏发电技术的成熟,建筑光伏发电系统的应用日益广泛。然而长期的积灰会导致建筑光伏发电系统的输出功率大幅下降,影响其有效使用。由于建筑光伏发电系统的输出功率具有波动性和随机性,并网后将给电网的安全稳定运行带来问题,为降低影响,需对光伏系统的输出功率进行准确的预测。实际中,大量的光伏输出功率预测方法主要考虑气象因素的影响,忽略了积灰的影响,从而导致输出功率预测精度降低。因此,考虑积灰因素的光伏输出功率预测方法研究具有重要价值。论文以建筑光伏系统为研究对象,研究积灰对其输出特性的影响,提出光伏组件表面积灰的检测方法,建立了积灰程度评价模型,并在此基础上构建光伏输出功率衰减模型。综合气象因素和积灰影响因素,建立光伏短期输出功率预测模型,以提高光伏输出功率的预测精度。论文研究内容:首先,对光伏组件表面积灰的物化特性进行实验分析研究,得到:建筑光伏组件表面积灰主要来源于土壤和道路扬尘,部分积灰可能来源于“雾霾”;开发建筑光伏组件输出特性监测系统,采集了积灰组件与清洁组件的发电参数及气象数据,对比分析了积灰对光伏组件输出特性的影响,分析发现:连续积灰12天的光伏组件输出功率衰减了8.87%;同时,在积灰累积的初期,光伏组件输出功率衰减较快,随着时间的推移,输出功率衰减逐渐趋缓。其次,提出了建筑光伏组件表面积灰检测方法和积灰程度评价模型。通过滤波和增强预处理积灰光伏组件的红外图像,使用改进的Otsu分割算法对红外图像进行处理,并良好的识别光伏组件表面积灰区域,取得了满意的识别检测效果。最后,结合积灰因素和气象因素,分别建立了光伏输出功率BP神经网络和SVM回归预测模型。利用两种预测模型对不同天气类型下的光伏输出功率进行预测,预测结果表明:对多云和阴天的光伏输出功率预测,BP神经网络预测模型的预测效果最好;而对于晴天,SVM回归预测模型的预测效果最佳,预测精度最高。总之,研究积灰对建筑光伏组件的影响及其检测技术,可为的后期的清洁维护提供技术支持;同时,准确预测建筑光伏组件的输出功率,对电网安全稳定运行及电网的操作管理与优化调度具有重要的意义。

【Abstract】 With the development of green building and zero-energy building as well as maturity of photovoltaic(PV)power generation technology,and the application of building PV power generation system is increasingly widespread.However,Long-term dust-accumulation can lead to a significant drop in the output power of the building PV power generation system,affecting its effective use.Because of the output power of the building PV power generation system have the volatility and randomness,and after grid connection,it will bring problems to the safe and stable operation of the power grid.Thus in order to decrease the effect,it needs to accurately forecast the output power of the PV system.Actually,a large number of prediction methods of PV output power mainly consider the influence of meteorological factors,neglecting the effect of dust-accumulation,which leads to a decrease in the output power prediction accuracy.Hence,the research of PV output power prediction method considering the effect of dust-accumulation has important value.This paper takes the building PV system as the research object,and studies the effect of dust-accumulation on its output characteristics.To puts forward the detection method of surface area ash of PV modules,establishes the evaluation model of dust-accumulation degree,and builds the PV output power attenuation model on this basis.Integrate the meteorological factors and the factors that affect ash deposition,establish a PV short-term output power prediction model to improve the prediction accuracy of PV output power.The content of the paper is as follows:At first,the experimental analysis and analysis of the physicochemical characteristics of PV modules surface area ash results in: The gray surface area ofbuilding PV modules mainly comes from the soil and road dust,and some of the accumulated dust may originate from haze.Development of the building PV module output characteristics monitoring system,collected the power generation parameters and meteorological data of the dust-accumulation assembly and the cleaning assembly,compared and analyzed the effect of dust-accumulation on the output characteristics of PV modules,and analyzed and found that the output power of PV modules with continuous accumulation of dust for 12 days decrease by 8.87%;At the same time,at the initial stage of accumulation of accumulated dust,the output power of PV modules decays rapidly.As time passes,the output power decays gradually.Secondly,the surface area dust-accumulation detection method and the evaluation model of dust-accumulation degree of building PV modules were proposed.Through the filtering and enhancement of preprocessed infrared images of the dust-stacked PV modules,the improved Otsu segmentation algorithm was used to process the infrared images,and the gray surface area of the PV modules was well identified,and a satisfactory detection effect was achieved.Finally,combined with dust-accumulation factors and meteorological factors,the PV output power BP neural network and SVM regression prediction model were established respectively.Two prediction models are used to predict the PV output power under different weather types.The prediction results show that: for the cloudy overcast PV output prediction,the BP neural network prediction model has the best prediction effect.For the sunny days,the SVM regression prediction model has the best prediction effect and the prediction accuracy is the highest.In short,the study of the effect of dust-accumulation on building photovoltaic components and their detection techniques can provide technical support for the later clean-up maintenance.At the same time,accurately predicting the output power of building PV module is of great significance to the safe and stable operation of the power grid and the operation management and optimal dispatch of the power grid.

Snapshot search of full-text: 

Knowledge network:

Citation network of current document
网页聊天
live chat
在线营销
live chat