【摘要】：針對(duì)傳統(tǒng)聚類算法不易選取初始聚類中心、對(duì)噪聲值較敏感、收斂速度慢及易陷入局部最優(yōu)等問(wèn)題,提出一種基于密度峰值的層次聚類算法對(duì)天氣類型進(jìn)行聚類。首先確定氣象數(shù)據(jù)的密度峰值參數(shù),采用分層聚類算法將氣象數(shù)據(jù)劃分為不同類別,然后利用支持向量機(jī)(SVM)對(duì)未知天氣類型進(jìn)行識(shí)別,最終采用徑向基(RBF)神經(jīng)網(wǎng)絡(luò)建立光伏發(fā)電短期功率預(yù)測(cè)模型。仿真結(jié)果表明,該方法能有效提高氣象類型的分類精度、加快尋優(yōu)速度,提高離群樣本點(diǎn)分離的魯棒性,證明了其在小樣本的情況下具有較高的精度,且在天氣波動(dòng)較大時(shí)仍能較好地實(shí)現(xiàn)功率值的預(yù)測(cè)。
[Abstract]:Aiming at the problem that traditional clustering algorithm is difficult to select initial clustering center, sensitive to noise value, slow convergence speed and easy to fall into local optimum, a hierarchical clustering algorithm based on peak density is proposed to cluster weather types. Firstly, the peak density parameters of meteorological data are determined, then the meteorological data are divided into different categories by hierarchical clustering algorithm, and then unknown weather types are identified by support vector machine (SVM). Finally, the short-term power prediction model of photovoltaic generation is established by radial basis function (RBF) neural network. The simulation results show that this method can effectively improve the classification accuracy of meteorological types, accelerate the speed of optimization, and improve the robustness of the separation of outlier samples. It is proved that this method has a high accuracy in the case of small samples. And the prediction of power value can be realized well when the weather fluctuates greatly.
【作者單位】： 上海電力學(xué)院自動(dòng)化工程學(xué)院;同濟(jì)大學(xué)電子與信息工程學(xué)院;
【基金】：國(guó)家自然科學(xué)基金(61573239) 上海市重點(diǎn)科技攻關(guān)計(jì)劃(14110500700) 上海市電站自動(dòng)化技術(shù)重點(diǎn)實(shí)驗(yàn)室項(xiàng)目(13DZ2273800) 上海市自然科學(xué)基金(15ZR1418600)~~
【分類號(hào)】：TM615
，

本文編號(hào)：2316878