Large-scale integration of wind power generation on power system planning

Abstract

Environment concerns and energy security yields no more time to further delay taking actions against mankind’s largest challenge in the 21st century: man-made climate change. The recently rapid development in power electronics and energy conservation under the currently undergoing financial tsunami provide just the opportunity to develop wind generation, which is the most promising and booming one among all renewable energy. The integration of large amounts of wind generation into the existing power system brings in intensive researches into its impact on system reliability, stability and power quality. As a sustainable and renewable energy source, wind is the most promising one, the proportion of which is supposed to become much larger in the future as the era of burning cheap and abundant fossil fuel energy passed. But its variability and average predictability have constrained its full utilization. Research into the impacts of increasing wind penetration on system planning and operation has drawn a major attention recently. The most obvious obstacle of wind utilization is its variability and average predictability. This unique characteristic must be paid attention to so as to deal with the emerging reliability and operation problems. A higher requirement for wind forecasting accuracy is raised because of the dramatically increased wind power generation capacity all over the world. Therefore, to achieve a high level of wind penetration, an accurate wind speed forecasting tool needs to be built. This thesis proposes a novel short term wind forecasting model based on Ensemble Empirical Mode Decomposition (EEMD) and combination of Support vector machines (Svm). The forecasting results have been compared with the results obtained from the reference models. Extensive tests with historical wind data obtained from meteorological stations in Hong Kong and UK verified that the proposed model is indeed able to produce forecasting results with the highest accuracy among all the reference models. With the assistant of preliminary applications, the crucial roles that wind speed forecasting tool plays are specified and analyzed, i.e. saving system reserve and improving wind trading price. It has been shown that with a greatly improved forecasting accuracy, a robust wind forecasting tool would improve the wind power integration in both economic and technical aspects. Increasing amounts of grid-connected wind farms would also be a big challenge in system operation and security since high wind penetration level exert a great influence on system plan and operation. Thus with increasingly installed wind generation into the present electricity networks, it is a big concern to evaluate the wind penetration limit of an existing electric power system since high level of wind penetration will cause various problems. Under the definition of instantaneous wind power penetration, wind penetration level is the ratio of the wind power output to the total load demand at some specific moment. This type of penetration is also called output penetration. This thesis presents a method of estimating the wind penetration maximum by analyzing the self-organized criticality (SOC) of power system, based on the complexity system theory. SOC is based on the idea that complex behavior can develop spontaneously in certain many-body systems whose dynamics vary abruptly, i.e. the nonlinear dynamics of a complex system under disturbances organized the global system state near to the state that is marginal to major disruptions, often as cascades. In a modern power system, interconnections of power system do not only provide convenience, like enhancing the capability of the electricity network to absorb wind power, but also cause problems. Power systems with strong transmission networks and robust interconnections to neighboring systems are expected to show a superior capability of absorbing large amount of wind power. With the assistant of the slightly modified IEEE 30-bus system and 118-bus system, case studies have revealed the self-organized criticality of the power system as being companied by the increased ratio of the wind power output to the total load demand, i.e. the instantaneous wind power penetration. Furthermore, under the definition of capacity penetration, wind penetration level is the installed wind generation capacity normalized by the total generation capacity on the system. From the viewpoint of utilities, wind power fluctuations could be considered as a negative stochastic load disturbance sources, which would result in more complicated and uncertain load variations, i.e. lead to a great influence on the frequency control of power systems. Automatic Generation Control (AGC) is one of the key control systems required for a successful operation of interconnected power systems. The primary objective of AGC is to maintain the frequency of each control area and to keep the tie-line power close to the scheduled values by regulating the power outputs of AGC generators to accommodate fluctuating load demands. AGC performances in the normal interconnected power system operation are usually monitored and assessed by interchange power flow, system frequency and other guideline standards all the time. This thesis also focuses on the impacts of wind penetration on the AGC system and estimation of the wind penetration as limited by NERC's new AGC performance standards, CPS1 and CPS2. With the realistic simulations based on the representative and historical system data from the China Southern Power Grid power system, it has been revealed that both CPS1 and CPS2 have been deteriorated as being companied by the increasing installed wind generation capacity normalized by the total generation capacity on the system, i.e. the installed capacity penetration.