With the increasing penetration of wind power, reliable and cost-effective wind energy production is of more and more importance. The doubly-fed induction generator based turbine system is widely used and dominates the wind market. In this paper, an analytical approach to assess reliability for power capacitors, both the DC-link capacitor bank and AC-side filter capacitor bank, is presented considering the annual mission profile. Based on the electrical behavior at various loading conditions, the lifecycle of the single power capacitor can be predicted through its electro-thermal stresses. This percentile lifetime can be translated to the Weibull lifetime distribution of the power capacitor by considering the parameter uncertainties and tolerance variations. Thereafter, a reliability block diagram is used to bridge the reliability curves from the component-level of the individual capacitor to the system-level of the capacitor bank. A case study of a 2 MW wind turbine shows that the lifecycle is significantly reduced from the individual capacitor to the capacitor bank, where the DC-link capacitor bank dominates the lifetime consumption. Furthermore, the electrical stresses of the power capacitors are experimentally verified at a down-scaled 7.5 kW prototype.

For the applications where a single capacitor is incapable to meet the needs, multiple capacitors are connected in series or in parallel as a bank to fulfill the capacitance and voltage rating requirements. Even though some commercial products have already existed in the market, most of the design use the same capacitors or combine different types of capacitors by experience, so that the volume, cost, reliability and power loss are not optimized. To the best knowledge, no quantitative design considering all these design aspects is available for capacitor banks. This paper proposes a model-based optimal design method for hybrid capacitor banks consisting of both electrolytic capacitors and film capacitors. Performance factors, such as impedance characteristics, lifetime, power loss, cost, and volume, are modeled and considered in the optimization process. The selection of the capacitance ratio between the two types of capacitors and the number of capacitors connected in parallel are analyzed based on specific design constraints. A case study of the DC-link capacitor bank design for a 5.5 kW inverter is presented to demonstrate the modeling and optimal design process.

When a single-switch open-circuit fault occurs in the series resonant dual-active bridge (SRDAB) converter, the output DC voltage will drop by a half or rise by twice. To maintain continuous power supply, a fault tolerant control method based on the voltage single-loop control is proposed, where the rectifier-side output square voltage is regulated. Nevertheless, it may excite the resonance between the resonant inductors and DC capacitors, leading to severe low-frequency oscillations, (appearing as the envelope of the high-frequency current). This may trigger the over-current protection and the SRDAB fails to ride through the fault. To address this issue, low-frequency equivalent models are proposed for the bidirectional power-flow of the SRDAB, enabling frequency-domain analysis of the single-loop control. The analysis reveals that the oscillation depends on the duty-cycle and control parameters, and it is more likely to occur when the converter operates in the boost mode. However, it is not possible to suppress the oscillations by the voltage single-loop control. Thus, a dual-loop fault tolerant method is developed. The proposed control strategy includes an outer-loop voltage control, an inner-loop current envelope control and a non-linear correction unit. Experimental tests on a 1-kW SRDAB have validated the effectiveness of the proposed method.

DC/DC converter is widely used in power electronic applications as a critical part of power supply. During operation, component parameters of the DC/DC converter gradually degrade subjected to various environmental stresses. When they degrade to deviate from the normal value, soft faults will occur and directly affect the functionality and reliability of the whole system. To identify these anomalies effectively, a novel soft fault detection and diagnosis method based on contrastive learning is proposed for a DC/DC Buck converter. First, the output voltage ripples of the converter under healthy and faulty conditions are collected. Then, a feature representation of the voltage ripple is obtained through hierarchical contrastive learning at temporal-level and instance-level. Based on the learned representation, a support vector machine classifier is established to detect and locate soft faults. The proposed method does not need extra hardware or signal injection, thereby reducing the cost and complexity. Simulation and physical experiment results verify the effectiveness and robustness of the proposed method.