Search

In this paper, a sparse machine learning technique is applied to predict the next-hour wind power. The hourly wind power prediction values within a few future hours can be obtained by meteorological/physical methods, and such values are often broadcast and available for many wind generators. Our model takes into consideration those available forecast values, together with the real-time observations of the past hours, as well as the values in all the power generators in nearby locations. Such a model is consisted of features of high dimensions, and is solved by the sparse technique. We demonstrate our method using the realistic wind power data that belongs to the IEEE 118-bus test system named NREL-118. The modeling result shows that our approach leads to better prediction accuracy comparing to several other competing methods, and our results improves from the broadcast values obtained by meteorological/physical methods. Apart from that, we apply a novel nonparametric density estimation approach to give the probabilistic band of prediction, which is demonstrated by the 25% and 75% confidence interval of the prediction. The coverage rate is compared with that yielded from quantile regression.

This paper utilizes modern statistical and machine learning methodology to predict the oscillation mode of interest in complex power engineering systems. The damping ratio of the electromechanical oscillation mode is formulated as a function of the power of the generators and loads as well as bus voltage magnitudes in the entire power system. The celebrated Lasso algorithm is implemented to solve this high-dimension modeling problem. By the nature of the L1 design, the Lasso algorithm can automatically render a sparse solution, and by eliminating redundant features, it provides desirable prediction power. The resultant model processes a simple structure, and it is easily interpretable. The precision of our sparse modeling framework is demonstrated in the context of an IEEE 50-Generator 145-Bus power network and an online learning framework for the power system oscillation mode prediction is also provided.

Renewable energy sources are a significant means of addressing dwindling conventional fuels and have been increasingly incorporated in transmission power grids on a large scale. However, the uncertainty of renewable energy source power outputs threatens the voltage security of power grids, which restricts the full integration of these sources to some degree. To address this problem, this work employs the reactive power optimization incorporating interval uncertainties (RPOIU) model for the transmission systems. In the RPOIU model, state variables, including load voltages, reactive power generation, and bus angles, are regarded as variables with interval values. Meanwhile, to account for the fluctuation of the load demand and the renewable energy sources in multiple periods of one day, a multi-time RPOIU is proposed. The model seeks to obtain a day-ahead optimal voltage control strategy for ensuring that interval state variables remain within their operating limit constraints in each period of one day, as well for maintaining a low level of daily real power losses. To solve the multi-time RPOIU model, we propose the interval sequential quadratic programming method (ISQPM), which employs a second-order interval Taylor expansion. In addition, an improved interval power flow calculation method, i.e., the optimizing-scenarios method, is employed to obtain the accurate ranges of the interval variables, thus improving the convergence performance of the ISQPM. The quadratic penalty function incorporating the rounding-off process is proposed to deal with the discrete variables in the multi-time RPOIU model. Three cases are tested by simulation to demonstrate the effectiveness, good applicability, and robustness of the proposed method, particularly in comparison to the previously proposed linear approximation method.

The Yunnan power grid has a large-scale hydropower station, whereby the power generated at the station is transmitted via HVdc lines under an asynchronous operation mode. In the system performance test of the Yunnan asynchronous operation, a 0.05-Hz ultralow-frequency oscillation occurred in the grid. This ultralow-frequency oscillation is a frequency stability problem caused by instability of hydraulic turbine governing system, which is quite common in the islanded system and very rare in the normal synchronous system. The relationship between the speed governing system stability of one unit and that of the system with parallel operation units was proposed to explain that the high proportion of units with an unstable governing system will lead to system frequency instability and the reduction in loads in the traditional sense was proven to further worsen the frequency stability of the Yunnan asynchronous system. By using related measures, the ultralow-frequency oscillation of the Yunnan asynchronous system has been curbed effectively. System performance testing verified the effectiveness of the analysis and the measures.

Renewable energy sources provide an important means of reducing reliance on conventional fuels. However, some renewable energy resources such as wind and solar energies are intermittent, and their uncertainty threatens the operating security of the power grid. To solve this problem, this study proposes the use of intervals to model the power output of renewable energy resources and the power load demand, and accordingly develops an interval voltage control model, i.e. interval reactive power optimisation model. The proposed model considers the control modes of renewable energy power generators and can safeguard the security of power grids by ensuring that the voltages reside within established limits. An adaptive genetic algorithm is employed to solve the proposed model, where a newly developed interval power-flow (IPF) calculation is used to solve the IPF equations, and penalty functions are applied to express inequality constraints. The proposed method is introduced in detail, and simulation results are presented to demonstrate its performance in comparison with a previously proposed interval voltage control method, as well as its applicability to large systems with various fluctuations of input data. The proposed approach provides robust convergence, obtains lower system power losses, and substantially reduces the computation time.

Yunnan power grid has a large-scale hydropower station whereby the power generated at the station is been transmitted via high-voltage direct current (HVDC) transmission lines under asynchronous operation mode. In the system performance test of Yunnan asynchronous operation, 0.05 Hz ultra-low frequency oscillation occurred in the grid. We analyzed the damping characteristics of different types of units and figured out that the negative damping characteristic of the hydropower units are mainly responsible for the ultra-low frequency oscillation. There is a tendency for ultra-low frequency oscillation in a regional power grid with high ratio of hydropower penetration which transmits electricity using HVDC. The water hammer effect of hydraulic turbine influences the damping characteristic of the hydropower unit while the governor control gain influences the strength of the damping. By setting reasonable governor control gain and optimizing the dead-zone of HVDC frequency limit controller the ultra-low frequency oscillation of the Yunnan asynchronous system can be effectively curbed. Simulation results and system performance test verified the effectiveness of the method.

To solve on-line practicability problem in transient stability assessment (TSA) and preventive control, a method for on-line TSA and preventive control is presented using static information. We first analyze the correlation between static information of power grid operation and transient stability and propose an on-line TSA algorithm based on dominant instability generator group (DIGG). Later, the quasi-linear relationship between critical clearing time (CCT) and active power of DIGG has been conducted. Thus, for dangerous operation modes, adjusting active power of DIGG can reach a new stable operation point whose CCT is adequately large to avoid power grid instability. One this basis, another algorithm of readjusting generator active power is suggested. This method can quickly and precisely assess the stability of power grid and offer adjustment scheme for power grid operators. The effectiveness and reasonability of the method is verified on IEEE 39 and CSG power grid systems.

Hierarchical and partitioned means of present voltage control results in the coordination problem among multiple coupling zones. From the perspective of game theory, split zones are regarded as game players in order to develop an individual rationality based cooperative game model, in which players search efficient alliance to improve their control effect. Players in the same alliance have a common optimization goal, while alliances play a noncooperative game, and the Nash equilibrium strategy would allow secondary voltage control area to consider the interaction and coupling from others thus avoid voltage oscillation and overshoot. A solution for nontransferable utility cooperative game is proposed, the existence of Nash equilibrium in the sub-problem is proved as well. Furthermore, steady-state simulation on IEEE 39-bus system demonstrates that the cooperative game model can effectively coordinate reactive power sources, thus improve the technical specifications of secondary voltage control.

Islanded operation of HVDC sending system can reduce the influence of HVDC fault on the parallel AC system, enhance the security of the system and increase the transmission capacity. Due to relatively small system inertia and ESCR, there are difficulties of the frequency and voltage control in the islanded sending system. The conventional analysis are focuses on the process after HVDC bipolar trip. An unexpected bipolar trip occurred after the monopolar trip test in the onsite commissioning of Yun-Guang HVDC islanded operation on May 26th, 2013. Process analysis shows that direct causes of the bipolar trip are the reverse-action of PSS, the low ESCR, and frequency adaptability of the HVDC overvoltage protection. An supplymentary test, concentrating on frequency adaptability of the HVDC overvoltage protection, was conducted on July 14th, 2013. But the temporary overvoltage risk still existed in the supplymentary test. Wide-area PSS and high-voltage-side voltage control are proposed to solve temporary overvoltage problem after HVDC trip. In the EMTDC/PSCAD simulation, overvoltage decreases about 60 kV to 80 kV by using these two measures simultaneously.

In order to solve the problem of voltage sag, Dynamic Voltage Restorer (DVR) is used to compensate load voltage instantly based on active hybrid energy storage system, and its equivalent circuit is presented. Both the mathematical model and the control system of the supercapacitor/battery hybrid system are set up, which improves whole working characteristic and economy of DVR. Simulation and analysis on DVR with active hybrid energy storage system are carried out, and the results show that this active hybrid energy storage system can effectively maintain the voltage level of sensitive load as to long-duration power interrupts and frequent voltage sag.

This paper utilizes modern statistical and machine learning methodology to identify the sensitivity of the power system oscillation mode. The deviation of the electromechanical oscillation mode is formulated as a function of power deviation of the generators and loads as well as voltage deviation of buses in the entire power system from the base case. The celebrated Lasso algorithm is implemented to solve this high-dimension modeling problem. By the nature of the L 1 design, the Lasso algorithm can automatically render a sparse solution, and by eliminating redundant features, it provides desirable prediction power. The resultant model processes a simple structure, and it is easily interpretable. The precision of our sparse modeling framework is demonstrated in the context of an IEEE 50-Generator 145-Bus power network.