本文關(guān)鍵詞：基于拉格朗日松弛法的停電系統(tǒng)機(jī)組恢復(fù)順序優(yōu)化方法研究　出處：《南京理工大學(xué)》2017年碩士論文　論文類型：學(xué)位論文

  更多相關(guān)文章： 停電系統(tǒng) 機(jī)組恢復(fù)順序 拉格朗日松弛法 次梯度 自適應(yīng)次梯度

【摘要】：電網(wǎng)大停電事故后的機(jī)組恢復(fù)問(wèn)題是一個(gè)非線性組合優(yōu)化問(wèn)題。合理的機(jī)組恢復(fù)順序能加快整個(gè)系統(tǒng)的恢復(fù)速度,減少停電損失。隨著電網(wǎng)互聯(lián)規(guī)模的擴(kuò)大,系統(tǒng)安全運(yùn)行面臨考驗(yàn),易發(fā)生大面積停電事故,提高大規(guī)模機(jī)組恢復(fù)控制策略的計(jì)算速度成了研究人員關(guān)注的重要問(wèn)題。目前,機(jī)組恢復(fù)順序的優(yōu)化大多以運(yùn)行經(jīng)驗(yàn)和專家系統(tǒng)為基礎(chǔ)。啟發(fā)式算法,智能算法等方法也相繼應(yīng)用到機(jī)組恢復(fù)順序問(wèn)題之中。專家系統(tǒng)知識(shí)庫(kù)的建立與維護(hù)工作量大,且具有局限性,難以全面獲取系統(tǒng)信息。啟發(fā)式算法和智能算法雖然能在一定程度上加快計(jì)算速度,但是對(duì)于大規(guī)模機(jī)組的計(jì)算時(shí)間仍然較長(zhǎng)。因此,需要進(jìn)一步研究提高機(jī)組恢復(fù)順序問(wèn)題計(jì)算速度的算法。本文提出了基于拉格朗日松弛法的停電系統(tǒng)機(jī)組恢復(fù)順序優(yōu)化的方法,主要完成了以下工作:1.針對(duì)當(dāng)前停電系統(tǒng)機(jī)組恢復(fù)順序數(shù)學(xué)模型復(fù)雜,形式多樣的問(wèn)題,從原理上分析了機(jī)組恢復(fù)的過(guò)程,并在此基礎(chǔ)上簡(jiǎn)化了目標(biāo)函數(shù)和約束條件,從而建立了停電系統(tǒng)機(jī)組恢復(fù)順序的數(shù)學(xué)模型。數(shù)學(xué)模型以機(jī)組的狀態(tài)變量作為目標(biāo)函數(shù)和約束條件的決策變量,不再具體區(qū)分黑啟動(dòng)機(jī)組與非黑啟動(dòng)機(jī)組,形式上得到統(tǒng)一,方便算法的求解。2.針對(duì)傳統(tǒng)算法計(jì)算速度上的不足,提出了將拉格朗日松弛法應(yīng)用于求解大規(guī)模停電系統(tǒng)機(jī)組恢復(fù)順序的優(yōu)化方法,分析了拉格朗日松弛法的數(shù)學(xué)原理,在求解對(duì)偶問(wèn)題的過(guò)程中,結(jié)合可分整數(shù)規(guī)劃問(wèn)題從原理上對(duì)計(jì)算速度的提高給出了合理的解釋。3.針對(duì)拉格朗日松弛法收斂速度慢的問(wèn)題,一方面給出了拉格朗日乘子初值的選取方法,另一方面在優(yōu)化過(guò)程中采用次梯度法和自適應(yīng)次梯度法相結(jié)合的方法迭代拉格朗日乘子,加快了算法的收斂速度,進(jìn)一步減少了計(jì)算時(shí)間。4.通過(guò)對(duì)實(shí)際的新英格蘭系統(tǒng),廣東電網(wǎng)部分分區(qū)系統(tǒng),江蘇電網(wǎng)部分分區(qū)系統(tǒng)進(jìn)行計(jì)算分析,驗(yàn)證了算法的有效性。
[Abstract]:Blackout after the crew recovery problem is a nonlinear optimization problem. The reasonable order of unit restoration can accelerate the recovery speed of the system, reduce the power loss. With the expansion of network interconnection, the safe operation of the system under test, prone to large area power outages, improve the calculation speed of recovery has become an important control strategy concerns of researchers of large scale unit. At present, optimization of plant restoration sequence is mostly based on the operation experience and expert system based heuristic algorithm, intelligent algorithms and other methods have been applied to the unit in recovery order. The expert system knowledge base establishment and maintenance, and has limitations, it is difficult to fully get system information. The heuristic algorithm and intelligent algorithm can accelerate the speed of computation to a certain extent, but the calculation time for large units still Long. Therefore, the need for further research to improve the calculation speed of the order of unit restoration algorithm. This paper presents a method of sequential optimization power system restoration unit based on Lagrange relaxation method, mainly completed the following work: 1. in view of the current power system unit restoration sequence of complex mathematical models, a variety of problems, from the analysis on the principle of process the unit of recovery, and on the basis of simplifying the objective function and constraints, so as to establish the mathematical model of unit power system restoration sequence. The decision variable mathematical model based on the state variable set as the objective function and constraint conditions, no specific distinction between black start units and black start units, forms are unified, for.2. convenient algorithm for the traditional algorithm the calculation speed of the proposed Lagrange relaxation method is applied to solve the large-scale power cut system The order of unit restoration optimization method, analyzes the mathematical principle of Lagrange relaxation method, to solve the dual problem, combined with the separable integer programming problem from the principle of the speed of calculation gives a reasonable explanation for the.3. Lagrange relaxation method for the problem of slow convergence, a method is given to select the Lagrange multiplier value the Lagrange iteration method, on the other hand by the subgradient method and adaptive subgradient method in the optimization process of the combination of the multiplier, to speed up the algorithm convergence and reduce the computational time of.4. on the actual new England system, part division system of Guangdong power grid, analyzed the Jiangsu grid partition system, verification the effectiveness of the algorithm.

【學(xué)位授予單位】：南京理工大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：TM732

【相似文獻(xiàn)】

相關(guān)期刊論文 前4條

1 賴菲;王智微;楊東;黃廷輝;胡洪華;;利用拉格朗日松弛法解決SIS中的負(fù)荷經(jīng)濟(jì)分配問(wèn)題[J];熱力發(fā)電;2007年06期

2 趙慶波,孫嵐,郭燕,高峰;基于拉格朗日松弛法的優(yōu)化調(diào)度系統(tǒng)[J];電力系統(tǒng)自動(dòng)化;2004年18期

3 毛學(xué)志;;水電廠廠內(nèi)經(jīng)濟(jì)運(yùn)行方法研究[J];科技信息;2013年36期

4 ;[J];;年期

相關(guān)碩士學(xué)位論文 前2條

1 張成;基于拉格朗日松弛法的停電系統(tǒng)機(jī)組恢復(fù)順序優(yōu)化方法研究[D];南京理工大學(xué);2017年

2 陶繼平;基于拉格朗日松弛法的調(diào)度算法研究[D];上海交通大學(xué);2014年

，

本文編號(hào)：1385694