Search

In recent years, online condition monitoring methods have gained popularity for their ability to predict latent insulation failure in electrical machines. Most of the existing methods are based on leakage current measurement. While neutral-voltage-based methods were firstly introduced for transformer monitoring, recent research has explored their application to electrical machines. However, the process of selecting the most suitable method for specific condition monitoring requirements remains ambiguous. This aricle offers a comprehensive comparison of the advantages and disadvantages of these methods, grounded in a theoretical analysis of the stator winding model. Sensitivities to various insulation aging problems are also studied, considering signal-to-noise ratios (SNRs) to determine the most suitable features for monitoring different aging issues. Additionally, a novel phase-to-ground insulation aging monitoring method is introduced, which accounts for the aging position along the winding. This method is based on the common-mode impedance spectrum and offers the benefit of reducing sensor bandwidth requirements when compared to existing methods. This research contributes to the refinement of condition monitoring for electrical machines, enabling more precise and efficient early detection of insulation aging problems.

Interturn short-circuit is typical asymmetrical fault of stator windings. The online diagnosis is significant for preventing the slight fault from further development into serious failure. To realize accurate detection of incipient interturn fault, specific oscillation modes involving entire phase winding are designed, which is generated during the cut-off transient process of antiparallel diode. Based on the characteristic frequency, the ratio of inductances between three phases can be calculated, which reflects the severity of the asymmetry between three phase windings. In the meanwhile, the oscillation frequency is not affected by the operating condition, which can be regarded as an inherent property of stator windings. This provides a superiority for online condition monitoring. In addition, the asymmetric fault of permanent magnet, such as the partial demagnetization, can be accurately distinguished. Online experimental tests are conducted to verify the feasibility of the proposed strategy.

The parametric uncertainties have a great effect on the output voltage of grid-forming inverters when they are operated under conventional modulated model predictive control (MPC). To address this known issue, a novel modulated model-free predictive control (MFPC) is proposed in this article. First, a new look-up table (LUT) formed by capacitor voltage differences and inductor current differences of LC-filter is designed for voltage-current predictions. Then, the state-space equations of different voltage vectors are established and reconstructed to update each voltage-current difference in the LUT per control period. Next, the duration of each voltage vector and the voltage prediction are calculated based on the updated voltage differences, so as to remove the impact of parameter variation on duration calculation and prediction calculation, enabling their robustness. Finally, the practicality of the proposed approach is demonstrated experimentally under both ideal and mismatched parameters. In comparison with conventional modulated MPC, the proposed modulated MFPC not only achieves comparable voltage performance, but also exhibits greater robustness.

Doubly-fed induction generators (DFIGs) play a vital role in wind power generation, yet their rotor winding insulation is susceptible to failure due to high dv/dt voltage from rotor-side converters and mechanical stress under varying operating conditions. While existing methods primarily focus on fault diagnosis, capable of detecting short-circuit faults but lacking in prognostic capabilities, this article delves into the development of an online condition monitoring method based on common-mode leakage current measurement. The study begins with an introduction to the model and measurement technique for winding insulation, followed by an analysis of capacitive coupling, which gives a path for high-frequency leakage current in DFIG systems. Furthermore, the article examines the impact of bearing impedance on rotor-side monitoring outcomes through theoretical analysis and simulation. Finally, the proposed method is validated through online experimentation, demonstrating its efficacy in providing real-time condition monitoring not only for rotor winding insulation but also for the state of shaft grounding devices in wind power generation systems.

The multistring topology is widely used in photovoltaic (PV) power generation applications. These dc-link capacitors of multistring dc/dc converters suffer from different ageing severities, so realizing their independent online monitoring is critical. The equivalent capacitance (Cx) is a key ageing indicator. First, a novel method for Cx estimation is proposed independent of both voltage and current across dc-link capacitor. The voltage of common dc bus and the output current of dc string are used. Second, a high-efficiency online monitoring strategy is proposed. The procedure is carried out daily after the conventional exits of dc strings. A set of data are collected for the linear fitting. It is proven the fitting curve has the intersection with y-axis that turns out to be the reciprocal of Cx. Simulation results clarify that the Cx estimation has high accuracy and the monitoring performance shows robustness to environment changes and load variations. Online experiments show that the error of Cx estimation can be limited within 3.6%. The novelty of proposed strategy lies in its high-efficiency characteristics that the independent online monitoring of multistring dc-link capacitors in PV applications can be achieved without extra hardware and intrusive operation.

Online condition monitoring for stator insulation is important for reliable operation of electrical machines. Recent studies show that the method based on the common-mode (CM) impedance spectrum is effective in online monitoring of the overall insulation aging condition. However, it is difficult to detect the interturn aging condition since the groundwall insulation aging has a similar impact on the monitoring results. In this article, the relationship between the change in the CM impedance spectrum and the stator winding insulation aging is studied based on the per-turn-unit broadband CM stator winding model. The dataset of impedance spectrum for different insulation aging conditions is built using this model. Multiple points in the impedance spectrum are used to detect the change in the interturn and groundwall insulation parameters separately. The online experimental results prove the validity of the proposed method. The proposed method can detect the insulation aging condition with the error about 6%. There is no other method that can quantify the aging degree of interturn insulation in the presence of groundwall insulation aging. Compared with the black-box data-driven methods, the proposed method is based on a model with clear physical meaning, which makes the results more reliable.