舰船科学技术  2017, Vol. 39 Issue (1A): 73-75   PDF    
大型舰船电力系统中潮流算法应用研究
刘丽     
沈阳职业技术学院 电气工程学院, 辽宁 沈阳 110045
摘要: 在传统的大型舰船电力系统中,常常会由于需求功率的快速增大,而出现不稳定的停车现象,对整个航行的安全造成非常大的安全隐患。因此有必要对常见的电力系统的综合性能进行优化,并设计合理的监测系统,对整个系统的运行状态进行监督。本文在分析舰船电力系统的基础上,对一般情况下的潮流抑制算法进行研究,并建立舰船电力系统的潮流优化模型,通过对此非线性模型的求解,从而在误差可控范围内,对电力系统的潮流进行预测,降低了系统的故障发生率。
关键词: 舰船电力     潮流算法     非线性系统    
Research on application of load flow algorithm in large ship power system
LIU Li     
Institute of Electric Engineering Shenyang Polytechnic College, Shenyang 110045, China
Abstract: In the traditional large-scale ship electric power system, it is often due to the rapid increase of demand power, and the phenomenon of unstable parking. Which will cause the entire navigation security is very large security risks. Therefore, it is necessary to optimize the general performance of the common power system and design a reasonable monitoring system to supervise the running status of the whole system. Based on the analysis of the ship's power system, this paper studies the power flow suppression algorithm in general situation and establishes the power flow optimization model of the ship power system. By solving this nonlinear model, And the trend of the power system is forecasted and the failure rate of the system is reduced.
Key words: ship electric power     power flow algorithm     nonlinear system    
0 引言

舰船电力需求越来越高,迫切需求开发出一款性能稳定,故障率低的船舶电力系统,而在电力系统中,潮流分布常常对整个系统的效能有着非常大的影响[1 -3]。因此本文的主要目标是建立电力系统的最优潮流模型,对其网络结构中的潮流分布和电能传输效率进行研究,从而建立其合理的潮流分布模型。本文从多个角度分析电力系统潮流模型中的非线性规划问题,并通过仿真对问题解的误差进行分析,仿真结果表明,本文提出的潮流优化算法已经达到较为优秀的水平。

1 电力系统最优潮流模型

船舶电力系统的最优潮流模型建立过程如下:

1)确定电力系统中所有发电机组的输出功率容量[4]

2)确定系统中用电设备的功率分布;

3)对电力网络分布结构进行建模。

通常所谓的电力系统潮流最优模型,主要是为了实现以下目标:在电力系统的网络结构已经确定的情况下,对每个用电设备的调度策略进行优化;下面建立电力系统的目标函数和约束方程[5 -6]

由矩阵表示的电力系统目标方程为:

$ f{(x)_{\min }} = \sum\limits_{i \in {S_G}} {\left( {{a_{2i}}P_{{G_i}}^2 + {a_{1i}}{P_{{G_i}}} + {a_{0i}}} \right)} = {F_{COST}}\text{。} $ (1)

式中ai 为电力系统发电机组的消耗系数,注意(i=1,2,0)。

设计系统的潮流方程fx)满足:

$ h(x) = {P_G} - {P_D} - e. * (Ge - Bf) - f. * (Gf - Be)\text{,} $ (2)

式(2)中的约束条件为:

$ g(x) = \left[ {\begin{array}{*{20}{c}} {{g_p}}\\ {{g_v}} \end{array}} \right]\text{,} $ (3)

式中:gp =pGI ISG ${g_v} = {e^{.2}} + {f^{.2}}$

用矩阵表示为:

$ J(x) = h{(x)^T} = \left[ {\begin{array}{*{20}{c}} {\displaystyle\frac{{\partial P}}{{\partial {P_G}}}} & {\displaystyle\frac{{\partial P}}{{\partial e}}} & {\displaystyle\frac{{\partial P}}{{\partial f}}} \end{array}} \right]\text{,} $ (4)

上式中满足:

$ \begin{aligned} & \frac{{\partial {P_I}}}{{\partial {P_G}}} = {I_{{n_g}}},I \in {S_G} \text{,}\\ & \frac{{\partial P}}{{\partial e}} = - (EG + FB) - {D_{iag}}\left[ {Ge - Bf} \right] \text{,}\\ & \frac{{\partial P}}{{\partial f}} = - (EB + FG) + {D_{iag}}\left[ {Gf + Be} \right] \text{。} \end{aligned} $ (5)

对上述的目标函数求一阶偏导:

$ \nabla f(x) = \left[ {\begin{array}{*{20}{c}} {\displaystyle\frac{{\partial f(x)}}{{\partial {P_G}}}}\\ \\ {\displaystyle\frac{{\partial f(x)}}{{\partial e}}}\\ \\ {\displaystyle\frac{{\partial f(x)}}{{\partial f}}} \end{array}} \right] \text{,} $ (6)

其中:

$ \begin{array}{l} \displaystyle\frac{{\partial f(x)}}{{\partial {P_G}}} = 2{a_2}. * {P_G} + {a_1} \text{,}\\ \\ \displaystyle\frac{{\partial f(x)}}{{\partial e}} = 2\sum\limits_{} {G. * (\overline E } - \overline E ') \text{,}\\ \displaystyle\frac{{\partial f(x)}}{{\partial f}} = 2\sum\limits_{} {G. * (\overline F } - \overline F ') \text{。} \end{array} $ (7)

约束条件为:

$ \nabla g(x) = \left[ {\begin{array}{*{20}{c}} {{I_{{N_g}}}} & 0 & 0\\ 0 & 0 & {\displaystyle\frac{{\partial {g_v}}}{{\partial e}}}\\ 0 & 0 & {\displaystyle\frac{{\partial {g_v}}}{{\partial f}}} \end{array}} \right] \text{,} $ (8)

其中:

$ \frac{{\partial {g_v}}}{{\partial e}} = 2Diag[e]\;\;\frac{{\partial {g_v}}}{{\partial f}} = 2Diag[f] \text{,} $ (9)

对潮流方程进行二阶偏导求解:

$ \begin{aligned} {\nabla ^2}h(x)y = \left[ {\begin{array}{*{20}{c}} 0 & 0 & 0\\[10pt] 0 & {\displaystyle\frac{{{\partial ^2}h(x)}}{{\partial {e^2}}}y} & {\displaystyle\frac{{{\partial ^2}h(x)}}{{\partial e\partial f}}y}\\[10pt] 0 & {\displaystyle\frac{{{\partial ^2}h(x)}}{{\partial f\partial e}}y} & {\displaystyle\frac{{{\partial ^2}h(x)}}{{{\partial ^{}}{f^2}}}y} \end{array}} \right] \text{,} \end{aligned} $ (10)

其中满足:

$ \begin{aligned} \frac{{{\partial ^2}h(x)}}{{\partial {e^2}}}y = & \frac{{{\partial ^2}h(x)}}{{\partial {f^2}}}y =\\ & - \overline G .*\overline Y + \overline B .*\overline {{Y_N}} + ( - \overline G .*\overline Y + \overline B .*\overline {{Y_N}} )' \text{,}\\ \frac{{{\partial ^2}h(x)}}{{\partial e\partial f}}y = & \left( {\frac{{{\partial ^2}h(x)}}{{\partial f\partial e}}y} \right)' =\\ & (\overline B .*\overline Y + \overline G .*{\overline Y _N}) - (\overline B .*\overline Y + \overline G .*{\overline Y _N})' \text{,}\\ & \overline Y = \left[ {\begin{array}{*{20}{c}} {y1} & \ldots & {y1}\\ \vdots & \ddots & \vdots \\ {{y_N}} & \cdots & {{y_N}} \end{array}} \right] = y(1toN){I_N} \text{。} \end{aligned} $ (11)
2 潮流算法中非线性规划的求解

对电力系统的潮流方程进行非线性求解,从而可以得到一个具有约束条件的解,因此设计如下的数学模型:

$ \begin{aligned} & \min f(x)\;h(x) = 0\;\;\underline g \leqslant g(x) \leqslant \overline g \text{,}\\ & x \in {R^n},h(x) = \left[ {{h_1}(x), \cdots {h_m}(x)} \right]' \text{,}\\ & g(x) = \left[ {{g_1}(x) \cdots {g_r}(x)} \right]' \text{,}\\ & x \in {R^n},f:{R^n} \to R,h:{R^n} \to {R^m};y \in {R^m} \text{,} \end{aligned} $ (12)

通过拉格朗日进行变换:

$ \begin{aligned} L(x, & l,u,y,z,w,\overline z ,\overline z ) \equiv f(x) - {y^T}h(x) -\\ & {z^T}\left[ {g(x) - l - \underline g } \right] - {w^T}\left[ {g(x) + u - \overline g } \right] - \overline z 'l - \overline w 'u \text{。} \end{aligned} $ (13)

式中 $y \in {R^m},(z,w,\overline z ,\overline w ) \in {R^r}$ 为拉格朗日乘子。

通过最优方程的求导,得到:

$ \left\{ {\begin{array}{*{20}{c}} {\displaystyle\frac{{\partial L(x)}}{{\partial l}} = z - \overline z = 0\text{,}}\\[10pt] {\displaystyle\frac{{\partial L(x)}}{{\partial u}} = - w - \overline w = 0\text{,}} \end{array}} \right. \Rightarrow \left\{ {\begin{array}{*{20}{c}} {\overline z = z} \text{,}\\ {\overline w = - w} \text{。} \end{array}} \right. $ (14)

从而可以得到最优方程解为:

$ \begin{array}{l} \left\{ {\begin{aligned} & {{L_x} = \nabla f(x) - \nabla h(x)y - \nabla g(x)(z + w) = 0} \text{,}\\ & {{L_y} = h(x) = 0} \text{,}\\ & {{L_z} = g(x) - l - \underline g = 0} \text{,}\\ & {{L_w} = g(x) + u - \overline g = 0} \text{,} \end{aligned}} \right.\\ \\ \left\{ {\begin{array}{*{20}{c}} \!\!\!\!\!{{L_l} = LZe = 0} \text{,}\\ \\ \!\!\!{{L_u} = UWe = 0} \text{。} \end{array}} \right. \end{array} $ (15)

式中: $\left( {l,u,z} \right) \! \geqslant 0, w \leqslant 0, y \ne 0, \left( {L,U,Z,W} \right) \in {R^{r \times r}}$ 为对角阵;Lx $\frac{{\partial L}}{{\partial x}}$ ,同时满足 $e \equiv {\left[ {1,1, \cdots 1} \right]^{\text{T}}} \in {R^r}$

3 算法仿真

建立船舶电力系统仿真环境,并就上文提出的算法性能进行仿真验证,由于潮流算法中的多数计算步骤都已经转化为线性方程的求导,因此节约了大量的仿真时间,同时该算法的综合性能得到了提高,稳定性也得到了改善。

图 1所示为原始算法和潮流算法的误差仿真结果。当系统的不确定量缓慢增加后,系统的电压输出也会随之降低,当系统的不确定数超过20后,2种算法之间的仿真误差会随之增加;特别是当不确定数达到35以后,两者的误差值已经达到了0.03。从曲线的平滑程度看,潮流算法的整体误差范围更小,系统的稳定性也更好。

图 1 误差分析 Fig. 1 Error Analysis
4 结语

本文重点研究了船舶电力系统中的潮流分布问题,建立了合理的状态模型,并通过对最优非线性方程组的求解,得到了误差范围可控的分布模型,具有一定的实用价值。

参考文献
[1] 冮明颖, 鲁宝春, 姜丕杰, 等. 基于广义Tellegen定理的小扰动定理在含光伏源的电力系统潮流算法中的应用研究[J]. 电力系统保护与控制, 2015 (18): 31–36.
[2] 刘光梅, 王锡淮, 肖建梅. 基于混合蛙跳算法的舰船电力系统网络重构[J]. 科技风, 2014 (8): 75–76.
[3] 王丛佼, 王锡淮, 肖建梅. 改进差分进化算法在舰船电力系统网络重构中的应用[J]. 船舶工程, 2013 (6): 55–59+67.
[4] SUEYOSHI T, GOTO M. Undesirable congestion under natural disposability and desirable congestion under managerial disposability in U. S. electric power industry measured by DEA environmental assessment[J]. Energy Economics, 2016, 234 : 432–440.
[5] JIANG Ping, MA Xue-jiao. A hybrid forecasting approach applied in the electrical power system based on data preprocessing, optimization and artificial intelligence algorithms[J]. Applied Mathematical Modelling, 2016 (12): 345–350.
[6] Mário César Giacco Ramos, Carlos Márcio Vieira Tahan. An Assessment of the electric power quality and electrical installation impacts on medical electrical equipment operations at health care facilities[J]. American Journal of Applied Sciences, 2009 (64): 65–70.