Zipeng Liang

Reactive Power Optimization Under Interval Uncertainty by the Linear Approximation Method and Its Modified Method

Reactive power optimization is a special kind of optimal power flow for optimizing voltage profile and reactive power flow in the steady state based on deterministic sets of the demand load and generation values, thus minimizing the real power losses or improving the voltage quality of the power grid. However, the input data in power systems have a certain degree of uncertainty that requires the reactive power optimization be solved by means of uncertain nonlinear programming, as advocated in the literature. To address this problem, we represent the uncertain input data as intervals and establish a model of the reactive power optimization that incorporates the interval uncertainties to describe the problem. The linear approximation method is formulated using the interval Taylor extension to help solve this type of problem. To obtain more accurate intervals for the state variables, the affine arithmetic-based power flow calculation is used to solve the interval power flow equation instead of crude computation based on the interval arithmetic, and thus the modified linear approximation method is developed. The proposed methods are presented in detail and the numerical results are analyzed to demonstrate their effectiveness and applicability, especially in comparison to the previously proposed chance constrained programming method.

Wrong data identification and correction for WAMS

With the widely use of wide area measurement system (WAMS), data measured by PMU is becoming more and more important for real-time operation in power system. At present, error warnings and neglect of real alarms problems are produced in WAMS when considering the injection of wrong data. However, state estimation performs poorly in dealing with the wrong data in WAMS causing tremendous computation resources and time. In this paper, a rapid identification and recovery method based on the concept of pattern recognition is put forward. Without repeated iterations and the injection of parameters of power grid, all the wrong data can be easily and quickly identified by changing from feature space to pattern space. To support our conclusion, 90 samples in practical application of WAMS are processed, demonstrating the effectiveness of the proposed method.

Aging state assessment of 110kV XLPE cable based on IRC and AE

To evaluate the aging condition of 110kV XLPE cables accurately, this paper employs the isothermal relaxation current (IRC) method and the activation energy (AE) method, discussing the relationship among the aging factor, the activation energy and the aging condition of cable insulation. The results show that compared with unaged cables, the cables under accelerated thermal aging have 0.348 more on the aging factor and their activation energy is generally lower. Besides, compared with the cables running for 14 years, the cables running for 25 years have 0.327 more on the aging factor, 15.05kJ/mol less on the activation energy and 30.6 years less on the remaining life. All these show that increasing of the internal traps number, deepening of the internal traps depth and the disintegration of chemical bonds are major microeconomic performances on cable aging. The differences in batching system and production technology system of cables can change the chemical structure in XLPE molecules. And assessment results are in line with expectations. Microscopically, evaluating the aging condition by IRC and AE has the accuracy and validity, which provides an effective method for the evaluation of XLPE cables aging condition from electrical and chemical aspects, and guides the maintenance and planning of cables.

Solution of interval reactive power optimization using genetic algorithm

Reactive power optimization is generally used to design an optimal profile of voltage and reactive power of power systems in steady state for deterministic sets of demand load and generation values, and it is a significant procedure in voltage control. However, the input data of power system is actually uncertain in practice, which makes reactive power optimization an uncertain nonlinear programming, and it is not solved properly at present. To address this problem, the input data is considered as interval and reactive power optimization incorporating interval uncertainties is proposed to model this problem. In order to solve this model, genetic algorithm is employed as the solution algorithm, where reliable power flow calculation is used to judge the constraints of the model. The IEEE14 system is tested and analyzed to demonstrate the effectiveness of the proposed method, especially in comparison to previously proposed chance constrained programming.