Peiyuan Chen

ARIMA-Based Time Series Model of Stochastic Wind Power Generation

This paper proposes a stochastic wind power model based on an autoregressive integrated moving average (ARIMA) process. The model takes into account the nonstationarity and physical limits of stochastic wind power generation. The model is constructed based on wind power measurement of one year from the Nysted offshore wind farm in Denmark. The proposed limited-ARIMA (LARIMA) model introduces a limiter and characterizes the stochastic wind power generation by mean level, temporal correlation and driving noise. The model is validated against the measurement in terms of temporal correlation and probability distribution. The LARIMA model outperforms a first-order transition matrix based discrete Markov model in terms of temporal correlation, probability distribution and model parameter number. The proposed LARIMA model is further extended to include the monthly variation of the stochastic wind power generation.

Probabilistic load flow: A review

This paper reviews the development of the probabilistic load flow (PLF) techniques. Applications of the PLF techniques in different areas of power system steady-state analysis are also discussed. The purpose of the review is to identify different available PLF techniques and their corresponding suitable applications so that a relatively accurate and efficient PLF algorithm can be determined for the concerned system, e.g. a distribution system with large integration of renewable energy based dispersed generations.

Stochastic Optimization of Wind Turbine Power Factor Using Stochastic Model of Wind Power

This paper proposes a stochastic optimization algorithm that aims to minimize the expectation of the system power losses by controlling wind turbine (WT) power factors. This objective of the optimization is subject to the probability constraints of bus voltage and line current requirements. The optimization algorithm utilizes the stochastic models of wind power generation (WPG) and load demand to take into account their stochastic variation. The stochastic model of WPG is developed on the basis of a limited autoregressive integrated moving average (LARIMA) model by introducing a cross-correlation structure to the LARIMA model. The proposed stochastic optimization is carried out on a 69-bus distribution system. Simulation results confirm that, under various combinations of WPG and load demand, the system power losses are considerably reduced with the optimal setting of WT power factor as compared to the case with unity power factor. Furthermore, an economic evaluation is carried out to quantify the value of power loss reduction. It is demonstrated that not only network operators but also WT owners can benefit from the optimal power factor setting, as WT owners can pay a much lower energy transfer fee to the network operators.

On Coordinated Control of OLTC and Reactive Power Compensation for Voltage Regulation in Distribution Systems With Wind Power

Active management strategies such as coordinated on load tap changer (OLTC) voltage control and reactive power compensation (RPC) are frequently suggested for voltage regulation in a distribution system with a high level of distributed generation (DG). This paper proposes a control and coordination algorithm for these two active management strategies. Voltage control through OLTC is achieved by using state estimation (SE) to determine the voltage in the network. To lower the implementation cost of the proposed control strategy, pseudo-measurements are used together with real-time measurement data in the SE. Moreover, the deadband of the automatic voltage control (AVC) relay is relaxed so that the AVC relay acts on the networks maximum or minimum voltage obtained through the SE. This is found to be simpler to realize than adjusting the set point of the AVC relay. Voltage control through RPC is actualized by using integral controllers implemented locally at the wind turbine site. Furthermore, RPC from the local wind turbine is also used to mitigate an overvoltage at a remote bus on the same feeder when the remote wind turbine reaches its regulation limit. The applicability of the proposed voltage regulation algorithm is successfully demonstrated using a case study system.

Study of power fluctuation from dispersed generations and loads and its impact on a distribution network through a probabilistic approach

In order to assess the performance of distribution system under normal operating conditions with large integration of renewable energy based dispersed generation (DG) dispersed generation, probabilistic modeling of the distribution system is necessary in order to take into consideration the stochastic behavior of load demands and DG units such as wind generation and combined heat and power plant generation. This paper classifies probabilistic models of load demands and DG units into summer and winter period, weekday and weekend as well as in 24 hours a day. The voltage results from the probabilistic load flow based on Monte Carlo method and those from the statistical measurement data are compared. Simulation results from the probabilistic model show good agreement with those from the statistical measurement data.

Strategic Placement of Wind Turbines in Smart Grids

Abstract This paper is to propose an optimization method that searches for the optimal installed capacity of wind turbines (WTs) in smart grids from the perspectives of both WTs developers and distribution network operator (DNO). The proposed hybrid optimization method combines the genetic algorithm and the multi-period optimal power flow analysis. The objective of the optimization is to maximize the net present value of the profits obtained by the WTs developers as well as that of the savings obtained by the DNO from network loss reduction. A 69 bus 11 kV radial distribution network is used as a case study to demonstrate the proposed algorithm, with the implementation of different active management schemes. Simulation results show that the optimal WT capacity from the perspective of DNO differs from the optimal value deemed by the WT developers. The resulting economic benefits can be used by both DNO and WT developers to justify an increase in WT installation if desired. The two contrasting schemes see either the WTs developers or DNO benefiting at the expense of the other. As such, using tradeoff analysis it may be possible to find suitable compromises.

The Effect of Wind Power Integration on the Frequency of Tap Changes of a Substation Transformer

As the capacity of wind power installed in a radial distribution system (DS) increases, there is a concern that it may introduce more frequent tap change operations in substation transformers. The increase in the frequency of tap changes (FTC) can accelerate the wear and tear of the tap changers. As a result, the introduction of wind power to DSs may be hindered. Hence the aim of this paper is to investigate the effect of wind power integration on the FTC in a radial DS. A case study shows that the changes on the FTC in DSs connected to relatively strong external grid is negligible up to significant level of penetration. But in DSs connected to a relatively weak external grid, a significant increase in the FTC has been observed as wind power penetration increases. Hence a further investigation is carried out to limit the FTC by using reactive power from local wind turbines. The results have shown that the methodology is very effective.

Optimal Allocation of Power-Electronic Interfaced Wind Turbines Using a Genetic Algorithm – Monte Carlo Hybrid Optimization Method

The increasing amount of wind power integrated to power systems presents a number of challenges to the system operation. One issue related to wind power integration concerns the location and capacities of the wind turbines (WTs) in the network. Although the location of wind turbines is mainly determined by the wind resource and geographic conditions, the location of wind turbines in a power system network may significantly affect the distribution of power flow, power losses, etc. Furthermore, modern WTs with power-electronic interface have the capability of controlling reactive power output, which can enhance the power system security and improve the system steady-state performance by reducing network losses. This chapter presents a hybrid optimization method that minimizes the annual system power losses. The optimization considers a 95%-probability of fulfilling the voltage and current limit requirements. The method combines the Genetic Algorithm (GA), gradient-based constrained nonlinear optimization algorithm and sequential Monte Carlo simulation (MCS). The GA searches for the optimal locations and capacities of WTs. The gradient-based optimization finds the optimal power factor setting of WTs. The sequential MCS takes into account the stochastic behaviour of wind power generation and load. The proposed hybrid optimization method is demonstrated on an 11 kV 69-bus distribution system.

Probabilistic assessment of wind power production on voltage profile in distribution networks

Distribution power networks are accommodating more and more generation and most of them are based on renewable sources, like wind power. Then, this new generation, generally known as distributed generation (DG), has a stochastic power production that affects technically the network in different ways. However, availability and power quality must be kept within standard limits. Deterministic approach can not cope with this issue and a probabilistic assessment is needed. Wind power has a high level of integration around the world. Then, its impact of voltage profile in a distribution network is analyzed.

Probabilistic load models for simulating the impact of load management

This paper analyzes a distribution system load time series through autocorrelation coefficient, power spectral density, probabilistic distribution and quantile value. Two probabilistic load models, i.e. the joint-normal model and the autoregressive model of order 12 (AR(12)), are proposed to simulate the impact of load management. The joint-normal model is superior in modeling the tail region of the hourly load distribution and implementing the change of hourly standard deviation. Whereas the AR(12) model requires much less parameter and is superior in modeling the autocorrelation. It is concluded that the AR(12) model is favored with limited measurement data and that the joint-normal model may provide better results with a large data set. Both models can be applied in general to model load time series and used in time-sequential simulation of distribution system planning.