Abstract
The inverter is one of the key components of wind turbine, and it is a complex circuit composed of a series of components such as a variety of electronic components and power devices. Therefore, it is difficult to accurately identify the operation states of inverter and some problems regarding its own circuit, especially in the early stages of failure. However, if the inverter temperature prediction model can be established, the early states can be identified through the judgment of the output temperature. Accordingly, considering whether the inverter heats up normally from the perspective of heat dissipation, a method for the early operation state identification of the inverter is provided in this paper. A variable selection method based on fusion analysis of correlation and physical relationship is adopted to extract variables as input variables, which have high correlation with inverter temperature. Then multi-input and multi-output temperature prediction model of inverter is established based on a non-linear autoregressive exogenous model (NARX) network, and the prediction temperature residual is used as the real-time standard to evaluate the inverter states. For validating this, the validity and reliability of the established temperature prediction model are verified through case analysis, and the performance comparison with various models demonstrates that the proposed method has higher accuracy. The construction method of the prediction model can be used for reference to other aspects of wind turbine. All these bring huge benefits to wind energy industry.
Highlights
The inverter is the hub connecting the generator and the power grid, which is mainly composed of two inverters with the same structure: the machine-side inverter realizes three-phase rectification and converts the AC voltage on the machine side into the DC voltage on the DC coupling capacitor; the inverter realizes the inversion and converts the DC voltage into AC voltage with the same frequency as the grid voltage [1]
This paper aims at large direct-drive wind turbines, and provides a method for early operation state identification of the inverter by concerning whether the inverter generates heat normally from the perspective of heat dissipation
Combining with wind farm SCADA data, this paper establishes a multi-input and multi-output temperature prediction model based on nonlinear autoregressive exogenous model (NARX) neural network, in order to study the multi-step-ahead temperature prediction for wind turbine (WT) inverter
Summary
The inverter is the hub connecting the generator and the power grid, which is mainly composed of two inverters with the same structure: the machine-side inverter realizes three-phase rectification and converts the AC voltage on the machine side into the DC voltage on the DC coupling capacitor; the inverter realizes the inversion and converts the DC voltage into AC voltage with the same frequency as the grid voltage [1]. The related researches mainly focus on the state prediction of key components such as wind turbine gearbox [17], generator [18], pitch system [19] by using SCADA data, mainly using long-short term memory (LSTM) [20], support vector machine [21], artificial neural network [22] and other methods, rarely involving complex electrical equipment such as inverter. The inverter state prediction and SCADA system are organically combined, in this paper, to establish a temperature prediction model based on NARX network, so as to identify the status of inverter, which can more effectively provide the operation and maintenance personnel with planned maintenance decision information, and provide a theoretical basis for the thermal management of WT inverter
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
