The rapid development of wind energy in the power sectors raises the question about the reliability of wind turbines for power system planning and operation. The electrical subsystem of wind turbines (ESWT), which is one of the most vulnerable parts of the wind turbine, is investigated in this paper. The hygrothermal aging of power electronic devices (PEDs) is modeled for the first time in the comprehensive reliability evaluation of ESWT, by using a novel stationary “circuit-like'' approach. First, the failure mechanism of the hygrothermal aging, which includes the solder layer fatigue damage and packaging material performance degradation, is explained. Then, a moisture diffusion resistance concept and a hygrothermal equivalent circuit are proposed to quantitate the hygrothermal aging behavior. A conditional probability function is developed to calculate the time-varying failure rate of PEDs. At last, the stochastic renewal process is simulated to evaluate the reliability for ESWT through the sequential Monte Carlo simulation, in which failure, repair, and replacement states of devices are all included. The effectiveness of our proposed reliability evaluation method is verified on an ESWT in a 2 MW wind turbine use time series data collected from a wind farm in China.

An unreliability tracing model for power systems assigns responsibility for unreliability to each component by quantifying its impact on the power system’s reliability. This paper extends the concept of unreliability tracing to power systems with high penetration of wind power, to realize unified allocation of load shedding to conventional generating units and wind power. First, a quantitative evaluation model for load shedding is established, which considers unreliable factors, such as failures of components, fluctuation, and intermittence of wind power. Then, load shedding at a certain time is decomposed into the part due to remaining unreliable factors that happened before the last time and the part due to unreliable factors that happened between the last time and the certain time, by decomposing the load shedding formula. Next, for parameters of the second part of the load shedding formula, unreliable factors that caused the value of these parameters to be located in regions leading to load shedding are decomposed into two parts in time sequence. The second part of load shedding is also decomposed into two parts corresponding to the decomposed unreliable factors. According to the decomposition of unreliable factors, two criteria for judging whether the component is responsible are extracted. Finally, the last two parts of load shedding are allocated to each component based on assumed baseline values of these criteria for each component, and the first part of load shedding is allocated, recursively. Several systems, including a modified provincial power system, are used to validate the rationality and validity of the model.

The outputs of renewable energy sources (RESs) are inherently variable and uncertain, such as wind power (WP) and photovoltaic (PV). However, the outputs of various types of RESs in different regions are complementary. If the capacity of RESs could be properly allocated during system planning, variability of the total output could be reduced. Consequently, system reliability and renewable energy (RE) consumption could be improved. This paper proposes an analytical model for optimal complementary capacity allocation of RESs to decrease variability of the total output. The model considers the capacity ratio of RESs as decision variables and the coefficient of variation (CV) of the total output as the objective function. The proposed approach transforms the single-level optimization model into a bilevel optimization model and derives an analytical equation that can directly calculate the optimal complementary capacity ratio (OCCR) of system RESs. Case studies on wind and solar farms in Xinjiang and Qinghai, China, are performed to verify the effectiveness of the proposed analytical allocation method.