發(fā)布時(shí)間：2021-02-24 10:58

【摘要】：近年來,隨著我國經(jīng)濟(jì)的快速發(fā)展,人們對(duì)電網(wǎng)運(yùn)行水平提出了更高的要求,大規(guī)模的降損措施被應(yīng)用到電網(wǎng),對(duì)于改善電源浪費(fèi)起到了積極作用。再者,隨著全球能源危機(jī)的加劇,十二五計(jì)劃提出堅(jiān)持把建設(shè)資源節(jié)約型、環(huán)境友好型社會(huì)作為加快轉(zhuǎn)變經(jīng)濟(jì)發(fā)展方式的重要著力點(diǎn)。為了積極響應(yīng)黨中央的政策,建立節(jié)約型社會(huì),最大限度地降低電能在傳輸過程中產(chǎn)生的損耗,開展線損預(yù)測研究變得尤為重要。配電網(wǎng)連接著發(fā)電系統(tǒng)、輸電系統(tǒng)和用戶,隨著電力系統(tǒng)的快速發(fā)展,配電網(wǎng)線損的不可避免性、復(fù)雜性和不確定性給電力系統(tǒng)的安全運(yùn)行及電能質(zhì)量帶來了嚴(yán)峻的挑戰(zhàn)。因此,準(zhǔn)確地對(duì)線損率進(jìn)行預(yù)測,幫助相關(guān)工作人員制定符合現(xiàn)實(shí)情況的考核指標(biāo)和規(guī)劃,有效發(fā)揮電能價(jià)值,已經(jīng)成為了當(dāng)前亟需分析和解決的實(shí)際課題。深入研究線損率預(yù)測,對(duì)提高電力系統(tǒng)協(xié)調(diào)運(yùn)行能力,促進(jìn)電能持續(xù)健康發(fā)展具有十分重要的意義�；�于歷史理論線損率數(shù)據(jù),本文對(duì)理論線損率展開了以下研究：1.神經(jīng)網(wǎng)絡(luò)預(yù)測模型能較好的應(yīng)對(duì)序列的波動(dòng)性且具有良好的精準(zhǔn)度,因此該非線性模型被廣泛的應(yīng)用。利用神經(jīng)網(wǎng)絡(luò)預(yù)測模型對(duì)理論線損率進(jìn)行預(yù)測,利用馬爾可夫?qū)�理論線損率預(yù)測誤差進(jìn)行修正處理,建立RBF-馬爾可夫模型,預(yù)測理論線損率。2.SVM(支持向量機(jī))收斂速度快,學(xué)習(xí)能力強(qiáng),泛化能力好,能有效預(yù)測序列的變化趨勢。以支持向量機(jī)原理為基礎(chǔ),建立支持向量機(jī)回歸模型,實(shí)現(xiàn)理論線損率直接預(yù)測。3.對(duì)支持向量機(jī)的模型參數(shù)優(yōu)化問題進(jìn)行研究。分析懲罰因子C和核參數(shù)σ的作用及對(duì)支持向量機(jī)性能的影響。運(yùn)用遺傳優(yōu)化算法,提出基于遺傳算法優(yōu)化支持向量機(jī)預(yù)測模型(GA-SVM),解決SVM建模時(shí)存在的弊端,并利用GA-SVM預(yù)測模型實(shí)現(xiàn)理論線損率預(yù)測。4.結(jié)合RBF-Markov模型和遺GA-SVM模型,研究理論線損率的概率預(yù)測模型。針對(duì)概率預(yù)測模型中求解概率密度難這一關(guān)鍵問題,采用非參數(shù)核密度估計(jì)方法估計(jì)理論線損率的概率密度函數(shù),最終建立概率預(yù)測模型求得置信區(qū)間。
[Abstract]:In recent years, with the rapid development of China's economy, people put forward higher requirements for the level of power grid operation. Large-scale loss reduction measures have been applied to the power grid, which has played a positive role in improving power waste. Furthermore, with the aggravation of the global energy crisis, the 12th Five-Year Plan puts forward that the construction of resource-saving and environment-friendly society should be regarded as an important point to accelerate the transformation of economic development mode. In order to respond positively to the policies of the CPC Central Committee, establish a conservation-oriented society and minimize the loss of electric energy in the transmission process, it is particularly important to carry out the research on line loss prediction. Distribution network is connected with generation system, transmission system and users. With the rapid development of power system, the inevitable, complexity and uncertainty of distribution network line loss bring severe challenges to the safe operation and power quality of power system. Therefore, the accurate prediction of line loss rate, the help of relevant staff to formulate assessment indicators and plans in line with the actual situation, and the effective use of electric energy value, has become a practical issue that needs to be analyzed and solved. It is very important to study the prediction of line loss rate for improving the coordinated operation ability of power system and promoting the sustainable and healthy development of electric energy. Based on the historical theory line loss rate data, this paper carries out the following research on the theoretical line loss rate: 1. Neural network prediction model can deal with the volatility of the sequence and has a good accuracy, so the nonlinear model is widely used. Neural network prediction model is used to predict the theoretical line loss rate and Markov model is used to correct the theoretical line loss rate prediction error. The RBF- Markov model is established. 2.SVM (support Vector Machine) converges fast, has strong learning ability and good generalization ability, and can effectively predict the change trend of the sequence. Based on the principle of support vector machine, the regression model of support vector machine is established, and the direct prediction of theoretical line loss rate is realized. The optimization of support vector machine (SVM) model parameters is studied. The effects of penalty factor C and kernel parameter 蟽 on the performance of support vector machines are analyzed. By using genetic optimization algorithm, the support vector machine prediction model (GA-SVM) based on genetic algorithm is proposed to solve the disadvantages of SVM modeling, and the theoretical line loss rate prediction is realized by using GA-SVM prediction model. 4. Combined with RBF-Markov model and posthumous GA-SVM model, the probabilistic prediction model of theoretical line loss rate is studied. In order to solve the problem that probability density is difficult to solve in probabilistic prediction model, the nonparametric kernel density estimation method is used to estimate the probability density function of theoretical linear loss rate, and the confidence interval is obtained by establishing the probabilistic prediction model.
【學(xué)位授予單位】：安徽工程大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2015
【分類號(hào)】：TM732


本文編號(hào)：2298688