【作者】 马士聪；

【导师】 高厚磊； 徐丙垠；

【作者基本信息】 山东大学 ， 电力系统及其自动化， 2008， 博士

【摘要】 随着社会经济的不断发展和人民生活水平的日益提高,用户对供电可靠性的要求越来越高。在对电力体制进行市场化改革的国家,供电可靠性是电力监管委员会监管和考核供电企业的关键指标。因此,各供电企业十分重视电力线路故障定位技术的开发与应用,以缩短故障查找与修复时间,提高供电可靠性和用户满意度。我国6～66kV中压电网普遍采用中性点非有效接地方式。采用这种运行方式,在发生单相接地故障(小电流接地故障)后,允许带故障运行一段时间,不影响负荷供电,提高了供电可靠性。但对于中性点非有效接地系统,发生单相接地后由于故障电流微弱,故障稳态特征不明显,导致小电流接地故障检测比较困难。近年来,小电流接地故障选线已基本得到了解决,但故障测距还没有很好的解决办法,需要开发研究适用于小电流接地故障的故障定位方法。本文研究利用暂态故障信号的小电流接地故障故障测距算法及应用技术。通过分析小电流接地故障的故障特征,利用测量端的暂态电压、电流信号建立小电流接地故障的测距方程,求解测距方程,计算故障距离。根据上述思路,本文研究并得出了适用于小电流接地故障的故障测距算法,通过ATP数字仿真及静态模拟试验证明了该方法的正确性和实用性。具体研究内容包括:1、总结分析小电流接地故障暂稳态特征,重点分析故障暂态过程,研究接地电流信号中几个主要分量的暂态特征。分析可知,小电流接地故障的稳态故障信号不适用于故障计算,特别是经消弧线圈补偿后,基于稳态故障信号的检测技术完全失效。而故障暂态信号幅值较大,变化明显,且暂态分析时可以忽略消弧线圈影响,更适合用于小电流接地故障故障距离的计算。2、分析研究电力线路的传统等值电路,提出适用于小电流接地故障计算的Γ形等值电路模型,并通过分析二端口外特性及阻抗.频率特性验证Γ形等值电路的有效性。与Π形、T形等值电路相比,采用Γ形等值电路,可以减少网络节点,简化故障方程,便于故障距离计算。通过分析对比Π形、T形、Γ形等值电路的二端口外特性,得出Γ形等值电路等效不同长度线路时的适用频率;通过对采用分布参数模型和Γ形等值电路模型的稳态电路进行阻抗-频率特性分析,得出在适用频率范围内Γ形等值电路可以用于电力线路等值的结论。Γ形等值电路的提出为建立小电流接地故障数学模型奠定了基础。3、采用卡伦堡尔(Karrenbauer)变换对小电流接地故障进行三相解耦,利用单相接地故障边界条件建立基于Γ形等值电路的小电流接地故障复合模网,列写小电流接地故障测距数学方程。该数学模型即描述母线端电压、电流与故障线路、接地电阻之间关系的接地故障方程。4、利用双线性z变换实现小电流接地故障拉普拉斯方程的频域离散化,建立小电流接地故障的时域差分方程。分析小电流接地故障微分方程时发现,由于数值微分误差较大,导致离散化后的差分方程用于测距计算精度较差。为此本文首先将时域故障方程转化为s域的拉普拉斯方程,再通过双线形z变换将s域连续的拉普拉斯方程转换为z域离散的z方程,利用z变换定义最终得到时域的差分方程。由此得到了包含电压、电流变量及故障线路参数(非金属性故障时还包含过渡电阻)的可数值求解的时域小电流接地故障差分方程。转换过程中避免了多值映射、频谱混叠等问题,提高了故障差分方程模型的精度,为下一步故障距离求解奠定了基础。5、利用参数识别技术求解小电流接地故障测距超定方程。针对利用采样信号联立方程,直接求解小电流接地故障方程时测距误差较大、结果不稳定的问题,本文采用参数识别中的最小二乘估计法求解由多量采样信号构成的故障超定方程,不需要任何先验知识,适合用于小电流接地故障暂态计算。6、简化小电流接地故障辨识方程,将方程中对线路参数的辨识转化为对故障距离的辨识。为克服线路参数辨识方程求解复杂、测距精度差的缺点,利用已知线路分布参数,将故障线路模参数全部转化为故障距离与单位长度线路模参数的乘积,即将线路模参数辨识方程转化为故障距离辨识方程。这一简化减轻了最小二乘算法的拟合负担,提高了辨识精度。7、利用基于电磁暂态软件包ATP的数字仿真和基于配电网物理模型的静模试验对所提出的小电流接地故障测距算法进行验证。仿真和试验结果证明了本文提出的小电流接地故障测距算法的正确性、有效性和实用性。8、通过对各辨识项分量分析比较以及结果校验,选取精度最高的辨识结果作为最终的故障距离。小电流接地故障距离辨识方程中包含两个故障距离辨识项,对比分析各分量的能量大小及抗干扰度,确定其中精度较高的辨识项作为测距结果。本文提出的小电流接地故障测距算法测距精度高,灵敏、可靠且不受中性点运行方式影响,对暂态信号有非常好的适用性,能在半个周波甚至更短的时间窗内完成准确的故障测距,能够满足现场工程要求,且该技术不需要附加任何设备,成本低,工程实现方便。算法的不足之处是高阻接地故障的测距精度较低,需要进一步改进。本文提出的小电流接地故障暂态测距算法,可以解决小电流接地故障定位这一长期没有得到解决的难题。该技术的推广应用对于提高电力系统供电可靠性以及推进小电流接地故障测距技术的发展具有十分重要的意义。更多还原

【Abstract】 Fast development of economy and improvement of living standards lead to requirements to higher power supply reliability. In the countries which the power market reform has been carried out, supply reliability is the important indicator of power company performance supervised by the State Electricity Regulatory Commission. In this situation, the development and application of power line fault location technique is paid more attention to by power utilities, which intend to shorten time of fault searching and repairing, and improve the supply reliability and customer satisfaction.In China, most MV distribution networks (6-66kV) employ non-effectively earthed neutrals. In such systems, when a single phase to earth (SPE) fault occurs, the faulted system can be permitted to keep service for 1-2 hours without removing the fault, which reduces the probability of more serious accidents. Unlike effectively earthed systems, the fault current caused by SPE fault in non-effectively earthed systems is very small, and the steady characteristics are inconspicuous. So the fault location issue for SPE fault in non-effectively earthed systems has not being well solved for a long time. The existing techniques have many shortages, such as low sensitivity, inaccurate mathematic model, especially in the aspect of inadequate utilization of transient signals. Therefore new fault location techniques need to be developed in order to meet the requirement of non-effectively earthed systems.The algorithm and application technique of fault distance estimation for SPE fault is researched in this paper. With analysis to the fault characteristics of SPE fault, the fault distance equation can be gotten based on the measured transient voltage and current. Then the fault distance can be gotten by solving the equation.According to the above key thought, a series of corresponding researches were carried out. At last a new fault location estimation algorithm is proposed, which is suitable to the SPE fault in non-effectively earthed neutral network. The result of ATP simulation and static simulation has proved its accuracy and practicality. The detailed research works include:(1)The transient and steady characteristics of SPE fault in non-effectively earthed neutral network are analyzed, and the transient signals are researched in detail. From the analysis, the steady fault signal is not suitable to fault estimation; especially for the compensated system, the methods based on the steady signals are invalid totally. But the amplitude of transient signals is bigger and has an obvious variety, and in the transient analysis, the Peterson coil can be neglected, which is convenience to the fault location estimation of non-effectively earthed neutral network.(2)The traditional equivalent circuits of power lines are analyzed and a newΓ-mode equivalent circuit is proposed. The accuracy ofΓ-mode circuit is proved by analyzing its two-port external characteristics and impedance-frequency characteristics. The employing ofΓ-mode equivalent circuit can reduce the net nodes and simplify the fault equation. The analyzing results proved thatΓ-mode equivalent circuit can equal medium length power lines in the selected frequency band. The adoption ofΓ-mode equivalent circuit lays a solid foundation for SPE fault mathematical model.(3)The SPE fault can be decomposed to aerial mode component and earth mode component by Karrenbauer transformation. And based onΓ-mode equivalent circuit, the modal network of SPE fault can be built with the boundary condition. Further the fault location estimating equation can be gotten. The equation shows the relationship of variables (voltage, current) at bus end, faulted line parameters and earth resistance.(4)The Laplace fault equation is discretized in frequency domain using bilinear transformation, and then the corresponding difference equation in domain can be gotten. If the fault differential equation is transformed to difference equation directly, the estimation accuracy is worse. Therefor the differential equation in s domain is transformed to Laplace equation in first, and then the Laplace equation is transformed to z equation by bilinear transformation. At last the difference equation in time domain is gotten with z transformation definition, which the difference equation is the fault estimation equation including voltage, current, faulted line parameters and earth resistance. During the transformations, frequency alias is avoided and the accuracy of fault difference equation is improved.(5)The overdetermined equations of SPE fault are solved by the least square method. If only few samplings is used to solve the fault equation, the calculation error is bigger and the result is instable. In this paper the least square estimation is used to solve the fault overdetermined equations, which is formed by plentiful samplings. The least square estimation needs not any prior knowledge, which is the best approach to solve the fault transient problems.(6)Simplify the fault estimation equation by the transforming from faulted line parameters estimation to fault location estimation. To overcome the shortage of complex and low-accuracy of faulted line estimation equation, the faulted line parameters is transformed to the product which multiplying the fault distance and line parameters per unit length using the distributed parameters. So the fault location equation can be gotten. This simplification reduces the calculation of the least square estimation and improves the estimation accuracy.(7)The results of ATP simulation and static simulation have proved the accuracy of the proposed fault location estimation algorithm. With building the simulation model and setting simulation parameters, different conditions of SPE fault is simulated. The results proved the accuracy and practicality of this fault location algorithm.(8)With comparison and examination to each estimation items of results, the high-accuracy item is chosen as the last distance result. The fault location estimation equation includes two distance items. From the comparison of energy and interference immunity, the high-accuracy item can be gotten as the result.The proposed algorithm has the high reliability, sensitivity and applicability to transient signals. And it is free from the Peterson coil and fault position, and can estimate the fault distance with half cycle or more short sampling length. This fault location estimation method is low cost without any additional devices and can be realized easily, which can satisfy the requirement of field engineering.But when the earth resistance is much higher, the estimation accuracy decreased relatively, which should be improved in future.The proposed distance estimation technical makes the important progress in fault location field of SPE fault, and its application has obviously significance for the improvement of supply reliability.更多还原