Search

This paper presents an offline deep learning approach focused to model and identify a 270 V-to-28 V DC-DC step-down converter used in on-board distribution systems of more electric aircrafts (MEA). Manufacturers usually do not provide enough information of the converters. Thus, it is difficult to perform design and planning tasks and to check the behavior of the power distribution system without an accurate model. This work considers the converter as a black-box, and trains a wavelet convolutional neural network (WCNN) that is able of accurately reproducing the behavior of the DC-DC converter from a large set of experimental data. The methodology to design a WCNN based on the characteristics of the input and output signals of the converter is also described. The method is validated with experimental data obtained from a setup that replicates the 28 V on-board distribution system of an aircraft. The results presented in this paper show a high correlation between measured and estimated data, robustness and low computational burden. This paper also compares the proposed approach against other techniques presented in the literature. It is possible to extend this method to other DC-DC converters, depending on their requirements.

This paper proposes an approach to estimate the state of health of DC-DC converters that feed the electrical system of an electric vehicle. They have an important role in providing a smooth and rectified DC voltage to the electric machine. Thus, it is important to diagnose the actual status and predict the future performance of the converter and specifically of the electrolytic capacitors, in order to avoid malfunctioning and failures, since it is known they have the highest failure rates among power converter components. To this end, accelerated aging tests of the electrolytic capacitors are performed by applying an electrical overstress. The gathered data are used to train a CNN-LSTM model that is capable of predicting the future values of the capacitance and the equivalent series resistance (ESR) of the electrolytic capacitor. This model can be used to estimate the remaining useful life of the device, thus, increasing the reliability of the system and ensuring an adequate operating condition of the electric motor.

This paper proposes an approach to estimate the parameters of an AC-DC boost power factor corrector converter which includes an EMI filter. To this end, once the topology is known, measurements at the input and output terminals of the converter are done to identify the values of the passive elements. The proposed methodology is based on the trust-region nonlinear least squares algorithm to identify the parameters of the converter. The steady-state and the transient signals of the converter at the input/output terminals are acquired non-intrusively without any internal modification of the circuitry. The accuracy of the parameter identification carried out is determined by comparing the estimated values with the actual values provided by the manufacturer, and by contrasting the measured signals with the ones obtained with a simulation model with the estimated values of the parameters. The results presented in this paper prove the accuracy of the proposed approach, which can be extended to other power converters and filters.

This paper presents a non-intrusive approach for modeling a bidirectional DC-DC converter used in mild hybrid electric vehicles. A black-box identification methodology is proposed to find a model based on the data acquired from the input/output terminals. Measured data include the steady state and transient response, and different operating conditions of the DC-DC converter, including the buck and boost modes. A deep learning architecture based on a long-short-term memory neural network (LSTM-NN) is applied. The trained network is tested under a set of operating points different from those used during the training stage. The proposed method is compared with three black-box modeling techniques commonly used in power converters, proving its superior performance. Results presented in this paper indicate that the proposed model is able to replicate the behavior of the bidirectional converter without a priori knowledge of the converter circuitry. This approach can also be applied to other power devices.

This paper presents a fast and efficient methodology to extract the electromagnetic magnitudes of a PMSM from its torque-speed-efficiency map. Thus, having the efficiency distribution along the torque-speed region, the algorithm is able to determine the electromagnetic parameters of the PMSM with no any geometry analysis involved. A d-q electrical model is used to compute the PMSM performance. The electromagnetic magnitudes to be modelled are the stator winding resistance, iron losses and d-q linkage flux, which jointly with the d-q currents, allows determining the d-q inductances. This approach is well suited as a fast electromagnetic parameter identification method to determine the electromagnetic magnitudes with no prior knowledge of the PMSM geometry. The obtained data can be helpful to perform a customized control algorithm of the PMSM. If the motor is available, by testing different operating points and extracting the efficiency data, it is possible to extract full information about the electromagnetic characteristics of the PMSM while adjusting the control algorithm to improve the motor performance. Finally, the magneto-static magnitudes are validated by means of FEA, using the public experimental data available of Toyota Prius 2004 PMSM.

Switching mode power converters are being extensively applied in different power conversion systems. Parameter identification comprises a set of techniques focused on extracting the relevant parameters of the converters in order to generate accurate discrete simulation models or to design enhanced condition diagnosis schemes. This paper applies a non-invasive optimization approach based on the non-linear least squares algorithm to determine the model parameters of different commercially available DC-DC power converters (buck, boost and buck-boost) from experimental data, including the parameters related to passive, parasitic and control loop elements. The proposed approach is based on a non-invasive on-line acquisition of the input/output voltages and currents under both steady state and transient conditions. The proposed method can also be applied to many other applications requiring precise and efficient parameter identification, including rectifiers, filters, or power supplies among others.

In the past, there have been some practical experiments to study the effect of high current and high temperature on the contact resistance of the substation connectors. However, there is a scarcity of practical works to study the effect of Aeolian vibrations. This paper presents an experimental study to analyze the effect of wind-induced vibrations on the contact resistance of the substation connector for tubular aluminum bus bars. Substation connectors are usually subjected to high current and high temperature stresses, thus being difficult to analyze the effect on vibrations when combined with these stresses. This paper presents a comparative study when the substation connectors operate with and without the effects of Aeolian vibration. Experimental results presented in this work show that the effect of vibrations on the contact resistance of the substation connectors is greatly influenced by the position of the connector and the type of mechanical support on the bus bar.

Artificial neural networks allow the identification of black-box models. This paper proposes a method aimed at replicating the static and dynamic behavior of a DC-DC power converter based on a recurrent nonlinear autoregressive exogenous neural network. The method proposed in this work applies an algorithm that trains a neural network based on the inputs and outputs (currents and voltages) of a Buck converter. The approach is validated by means of simulated data of a realistic nonsynchronous Buck converter model programmed in Simulink and by means of experimental results. The predictions made by the neural network are compared to the actual outputs of the system, to determine the accuracy of the method, thus validating the proposed approach. Both simulation and experimental results show the feasibility and accuracy of the proposed black-box approach.

Six-pulse diode rectifiers are widely used in many power conversion systems. In this work a non-intrusive approach for parameter identification of such power device is presented and its accuracy and performance is evaluated from experimental data. This parameter identification approach is focused on generating accurate discrete simulation models of three-phase rectifiers, although it can also be applied for health condition monitoring and diagnosis purposes. The approach presented in this paper is based on applying the non-linear least squares (NLS) algorithm following an optimization strategy to identify the parameters of a white-box model of the analyzed power devices. This approach analyzes the currents and voltages measured at the input/output terminals, thus performing a non-invasive acquisition of such signals, which is compatible with sealed systems. The method proposed in this paper can also be applied to many other power devices needing accurate simulation models or applying health condition strategies, switched power supplies or power converters.

This paper presents a novel approach for black-box modelling of 270 V -to- 28 V DC-DC step-down converters used in more electric aircrafts (MEA). These converters normally feed constant power loads (CPL). The proposed deep learning approach, uses offline experimental data of the converter to find an accurate model that reproduces its behavior. It covers a broad range of loading conditions to build a model that replicates the whole behavior of the converter. This paper compares the performance of the proposed method, which requires a very low computational burden once the model is trained, with that of a conventional recurrent neural network (RNNs) topology. Results presented in this paper show the ability of the obtained solution to accurately emulate the behavior of the real step-down converter when the internal structure is unknown, with no knowledge of the internal parameters, thus preventing disclosure of manufacturers confidential data. The modeling strategy presented in this paper is validated with experimental data by using a step-down converter used in aircrafts. The approach is compared to existing modeling techniques to test its accuracy. This approach can also be applied to many power devices, including diverse types of power converters, power supplies, or filters among others.