摘要:電力負(fù)荷預(yù)測(cè)是電力生產(chǎn)、調(diào)度部門的重要工作之一，它的實(shí)質(zhì)是根據(jù)預(yù)測(cè)對(duì)象的歷史數(shù)據(jù)建立相應(yīng)的數(shù)學(xué)模型，描述其發(fā)展規(guī)律。因此，能否對(duì)電力負(fù)荷進(jìn)行可靠預(yù)測(cè)直接關(guān)系到電力行業(yè)的經(jīng)濟(jì)效益。

　　　本文采用了基于灰色理論與BP神經(jīng)網(wǎng)絡(luò)相結(jié)合的電力負(fù)荷預(yù)測(cè)方法。文中首先介紹了電力負(fù)荷預(yù)測(cè)的原理、起源以及當(dāng)前國(guó)內(nèi)外電力負(fù)荷預(yù)測(cè)領(lǐng)域的概況，同時(shí)介紹了目前常用的幾種電力負(fù)荷預(yù)測(cè)方法，并對(duì)新引入電力系統(tǒng)的灰色預(yù)測(cè)理論及其相應(yīng)的各種模型的基本原理做了簡(jiǎn)要的探討。然后用三種不同模型（灰色GM(1,1)模型，一元線性回歸模型，等權(quán)平均組合模型）進(jìn)行預(yù)測(cè)，然后把預(yù)測(cè)值作為訓(xùn)練樣本對(duì)神經(jīng)網(wǎng)絡(luò)進(jìn)行訓(xùn)練，并用訓(xùn)練好的網(wǎng)絡(luò)進(jìn)行負(fù)荷預(yù)測(cè)。由于該預(yù)測(cè)模型采用了灰色理論與BP神經(jīng)網(wǎng)絡(luò)相結(jié)合的預(yù)測(cè)算法，吸取了二者的優(yōu)點(diǎn)，避免了單一預(yù)測(cè)模型所存在的預(yù)測(cè)風(fēng)險(xiǎn)。

　　　從對(duì)重慶市某供電局2007各月電量負(fù)荷預(yù)測(cè)實(shí)例中可以看到，較灰色單一預(yù)測(cè)模型而言，該方法的預(yù)測(cè)精確度和穩(wěn)定性都有明顯的改進(jìn)，理論上是可行的，并對(duì)電力行業(yè)進(jìn)行計(jì)劃擬定和生產(chǎn)運(yùn)行均具有一定的指導(dǎo)和應(yīng)用意義。

關(guān)鍵詞：灰色理論；BP神經(jīng)網(wǎng)絡(luò)；負(fù)荷預(yù)測(cè)；精度

Abstract：Power load forecasting power production, scheduling department is one of the most important work, its essence is according to the historical data of the forecasted object establish corresponding mathematical model, described the law of development. So, can make to the electric power load forecasting for reliable directly related to the economic benefits of electric power industry. 

    In this paper based on the Grey theory and the BP neural network of power load forecasting method, this paper firstly introduces the principle of power load forecast, the origin and current domestic and international power load forecasting the general situation of the field, and introduced the common several power load forecasting method, and the introduction of new power system Grey prediction theory and its corresponding all sorts of model of the basic principles do was briefly discussed in this paper. And then in three different model (Grey GM (1, 1) model, the linear regression model, a right to such as average combination model) to carry on the forecast, and then the predicted as the training sample the neural network training, and training good network load forecasting. Because of the prediction model using the Grey system theory and the BP neural network with the combination of prediction algorithm, drawing on the strengths, to avoid the forecast model of the existing prediction risk.  

    In Chongqing from a power supply bureau was produced 2007 power load forecasting example can see, a Grey forecast model in the prediction accuracy and stability of the method are the obvious improvement, this algorithm in theory and practice is feasible, and the electric power industry for planning and production operation has certain directive and all applications.

Keywords: Grey theory; The BP neural network; Load forecasting; Precision