【作者】 汪洋；

【导师】 李文沅； 卢继平；

【作者基本信息】 重庆大学 ， 电气工程， 2009， 博士

【摘要】 广域测量系统WAMS借助GPS等高精度全球同步时钟,可以得到电网全局统一时空坐标下的动态信息,有效增强系统的可观测性,进而实现系统动态过程的实时监控,提高电网的自动控制和安全稳定水平。但是,如同其它任何系统一样,WAMS本身也有可能发生故障失效,而且由于WAMS涉及到了整个电力系统,因此其失效后果往往更加严重。如果只关注系统引入WAMS后的有利方面,而忽略它的潜在风险,这样得到的安全控制措施将是一个不完全的“单腿”方案。作者广泛检索了现有文献资料,结果表明:虽然在电力系统可靠性的其它领域中已经有了诸多成果,但目前对WAMS的失效研究几乎是一个空白,还没有相关的文章。论文率先提出了评估WAMS可靠性的分析方法,并建立了与WAMS功能特点相适应的可靠性评估模型。鉴于WAMS系统是包含多设备(包括多台同步相量测量装置PMU)和专用通信网络在内的软硬件相结合的复杂系统,论文先采用了故障树分析方法将WAMS分解为PMUs、相量数据集中器PDC、WAMS局部通信网络、WAMS主干通信网络以及控制中心几个部分,并给出了WAMS整体可靠性的计算公式。然后,论文针对WAMS各子部分的特点建立了相应的可靠性模型,并运用多种可靠性评估方法推导了WAMS各子部分的两状态等效概率模型;同时还对各子部分进行了灵敏度分析,以识别各子部分中影响可靠性的关键模块和参数。最后,利用各子部分的两状态等效概率模型和整个WAMS的可靠性计算公式建立了WAMS整体的等效两状态Markov模型。所提出的模型可以广泛应用于评估各种基于WAMS的控制措施对系统风险的影响,包括本文提出的基于WAMS的电压稳定性控制方案的风险。该模型还可以进一步用于评估引入WAMS后由常规电力系统和通信、控制、测量系统组成的广义系统的整体风险。作为WAMS实际功能应用的一个非常合理的延伸,论文接着建立了一套基于WAMS的电力系统电压稳定性指标。所提出的指标可以识别系统的电压稳定薄弱环节(薄弱线路和节点),并准确预测薄弱环节的传输功率极限。这些指标不但能够克服传统基于系统潮流信息的电压稳定性指标(如基于最大功率法和雅可比矩阵奇异法的指标等)的缺点,不需要对整个系统进行潮流计算,因此计算速度非常快;而且克服了现有大多数局部性指标的不足,在模型中考虑了系统对局部网络的影响,具有较高的精度;该指标基于WAMS动态实时测量数据,因此能自动处理与电压相关或与频率相关的实际负荷特性,捕捉到负荷的“慢动作”变化过程,及时启动系统紧急保护控制以避免系统发生电压崩溃;同时由于所提出的指标依赖的WAMS信息量较少,因此对WAMS本身的可靠性要求相对较低,更便于实际实施。4个IEEE试验系统的仿真结果表明,所提出的指标可以精确地判断出系统发生电压崩溃的薄弱环节,并预测出系统电压崩溃裕度。利用所提出的新型电压稳定性指标,论文进一步建立了考虑电压稳定性约束条件的最优切负荷模型。该模型通过切除部分负荷、减少系统中电压稳定性薄弱环节上的传输功率,可以有效地降低系统发生电压失稳的风险。论文采用了预测-校正原对偶内点法求解该模型,并通过Q-V模态分析研究切负荷前后系统降阶雅可比矩阵的特征值变化情况,评估切负荷控制措施的有效性。2个IEEE标准试验系统的仿真结果表明,该模型直接针对系统薄弱环节,自动辩识切负荷节点和最小切负荷量,具有很好的效果。更多还原

【Abstract】 The WAMS can be used to conduct real time monitoring and control in dynamic system states and enhance the system security level because it utilizes the high precise synchronous clock on the earth such as GPS to build a unified space-time ordinate for the whole system. However, like any other physical system, the WAMS can fail while it can provide much better observability and controllability in power system operations. The consequence of WAMS failure is severe and could be a large blackout because of its wide-area impacts if its emergency control function does not work properly. It has been recognized that the reliability of WAMS must be quantitatively evaluated and assured while utilizing its beneficial features. In other words, the WAMS based protection and control scheme would be a“single leg”one if the risk of WAMS failure is ignored. The author has carried a wide range of literature reading, which indicates that the quantitative reliability evaluation of WAMS has not been addressed so far although considerable efforts have been devoted to other aspects of power system reliability assessment. For the first time, the dissertation proposed the WAMS reliability evaluation techinques and established relevant reliability models of WAMS.The WAMS itself is a complex system that is associated with both software and hardware and contains multiple devices (such as many PMUs) and special communication networks. The whole WAMS is first divided into several sub-systems including PMUs, PDC, local communication network, backbone communication network and control center. The Fault Tree Analysis (FTA) method is used to derive the general formula of evaluating WAMS system reliability. Then, the reliability models of all sub-systems are individually developed based on their features and their two-state equivalent Markov models are established using a variety of probabilistic reliability assessment techniques. Sensitivity analyses are also performed to identify key modules and parameters in each sub-system. Finally, the two-state equivalent Markov model of the whole WAMS is established using the two-state models of sub-systems and the general formula. This two-state model of the whole WAMS can be widely used to assess the effects on the system risk of various WAMS based control schemes, including the voltage stability control scheme proposed in this paper. It can also be applied to evaluate the risk of a generalized system that contains a regular power system, communication networks, control and measurement systems. As a natural and rational extension of WAMS application, the dissertation proposes a set of new voltage stability indices based on the measurements of WAMS, which can be used to identify weak buses or circuits in the system and accurately predict their transfer capability limits. The indices have four major advantages:①Compared to traditional power flow-based voltage stability indices (such as those based on the maximum power method or the Jacobian matrix singularity), the proposed indices do not require any computation associated with system-wide power flow and therefore is much faster.②Compared to majority of the voltage indices based on local measurements, the proposed indices include the effect of overall system outside the local network using a new equivalent model of the system. This assures accuracy of the indices in modeling.③The WAMS based indices could automatically include time-dependent frequency and voltage related load characteristics and start up the emergency protection and control scheme to prevent voltage collapse in time.④The proposed indices require a small amount of WAMS measurements and therefore has less reliance on the reliability performance of WAMS. This feature is very useful in practical application of the indices. The simulation results of 4 IEEE standard test systems indicate that the proposed indices can accurately recognize weak buses and circuits, and predict the margin from the voltage collapse point.Based on the proposed voltage stability indices, the dissertation establishes a minimum load shedding Optimal Power Flow (OPF) model with voltage stability constraints. The power level required to transfer on the weak buses and circuits is reduced by the OPF model through shedding partial loads. This can effectively reduce the risk of voltage instability in the system. The Predictor Corrector Primary Dual Interior Point Method (PCPDIPM) is used to solve the OPF model. The effectiveness of load shedding is confirmed using the Q-V mode analysis method, which can be applied to compute eigenvalues of a reduced Jacobian matrix. The numerical results of 2 IEEE standard test systems demonstrate that the proposed OPF model is very effective since it automatically identifies the weak buses and circuits and calculates minimum load shedding.更多还原