### Refine

#### Year of publication

- 2020 (10)

#### Document Type

- Conference Proceeding (6)
- Article (2)
- Other (2)

#### Institute

- Mechatronik und Elektrotechnik (10) (remove)

Application and machine learning methods for dynamic load point controls of electric vehicles (xEVs)
(2020)

From the customer's perspective, the appeal of electric vehicles depends on the simplicity and ease of their use, such as flexible access to electric power from the grid to recharge the batteries of their vehicles. Therefore, the expansion of charging infrastructure will be an important part of electric mobility. The related charging infrastructure is a big challenge for the load capacity of the grid connection without additional intelligent charge management: if the control of the charging process is not implemented, it is necessary to ensure the total of the maximum output of all xEVs at the grid connection point, which requires huge costs. This paper proposes to build a prediction module for forecasting dynamic charging load using machine learning (ML) techniques. The module will be integrated into a real charge management concept with optimization procedures for controlling the dynamic load point. The value of load forecasting through practical load data of a car park were taken to illustrate the proposed methods. The prediction performance of different ML methods under the same data condition (e.g., holiday data) are compared and evaluated.

The use of small-scale wind turbines (SSWTs) in private households not only allows for increased renewable energy generation, but also improved grid stability and resilience of individual regions. However, there are strict requirements regarding the efficiency, reliability and electrical safety of SSWTs because of their low power levels, long payback period and the fact that they are installed nearby residential areas. This paper proposes a novel yaw inductive power transfer system, based on multilevel inverter, which mitigates the main disadvantages of SSWTs - slip rings, low voltage energy generation etc. The system utilizes low voltage MOSFETs in multilevel boost-series resonant topology combined with ZVS techniques to maximize efficiency. The operating principle, switching waveforms and behavior of the inverter are described and analyzed. Cell voltage balancing algorithms are presented as well. Two different control techniques for power flow regulation have been introduced and compared. The effectiveness of the proposed solution has been verified by building a 3.3 kW prototype system and comprehensive measuring of its performance. The experimental results show that peak power transfer efficiency reaches 92.5% while converting generator voltage from 60 V DC to grid compatible 400 V DC.

In recent years, a number of bidirectional inductive power transfer systems (BD-IPT) suitable for wireless grid integration of electric vehicles have been developed. These developments have been fueled by the enhanced efficiency and spatial tolerance offered by BD-IPT systems. A typical BD-IPT system utilizes two synchronized full-bridge converters operating at fixed duty cycles to drive the primary and secondary magnetic couplers. However, in order to cater for a wide range of loading conditions, additional circuitry is employed at the expense of cost and power density. As an alternative solution, this paper proposes a novel power converter, named a boost active bridge (BAB), to replace the full-bridge converters. The BAB topology caters to a wide range of loading conditions without the need for any extra switching devices. A comprehensive mathematical model that predicts steady-state currents, voltages, and power transfer is presented to highlight the operating principles of the BAB technology. Experimental results obtained from a 3.5-kW prototype show a nearly constant efficiency under all loading conditions, validating the viability of the proposed BAB topology.

Industrie 4.0 und Big Data
(2020)

Coverage Probability of Methods for Steady-State Availability Inference with a Confidence Interval
(2020)

The quality of a repairable system can be described using its availability. Typically, a high degree of availability is demanded by the customer. To analyze the availability of a repairable system, the specification of reliability and maintainability are needed. Usually, they are demonstrated based on limited sample sizes, e.g., by analyzing the failure times of a life time test. The evaluation of the test results yields a mean-value distribution of reliability and maintainability as well as its confidence interval. Consequently, the calculated availability based on these inputs also need to be expressed including a confidence interval.
In this paper, firstly several approaches to calculate the confidence interval of steady-state availability based on reliability and maintainability are presented. Afterwards, a procedure to investigate and evaluate the quality and accuracy of the confidence intervals calculated with the presented methods is shown. Therefore, the coverage probability as the most common indicator is used. Based on an exemplary parameter study which is performed, the accuracy of the confidence intervals determined with the methods is investigated and evaluated in the case of exponentially distributed failure and repair times. Finally, several hints for an effective availability calculation with confidence intervals are given.

Reliability demonstration is performed before a product is released to the market. Often, this demonstration is based on accelerated life testing of samples with limited sample size. Accelerated life testing aims to parameterize a statistical life-stress model. Based on such a model, the reliability demonstration is performed for the stress a product is experiencing during operation. The reliability needs to be inferred with a confidence interval so that the uncertainty, which stems from limited sample data, can be considered. Typically, the load conditions of a field population show a significant variation of stress. A method to consider the comprehensive statistical uncertainty and distribution of stress and life-stress model was recently published. However, this method is limited to applications with constant stress over time. In this paper we present a first approach for a method that is able to consider the distribution of load spectra and statistical lifetime model as well as their uncertainties due to limited sample sizes and allows the consideration of non-constant, i.e. time-varying, stresses for the reliability demonstration. The presented method enables the reliability inference at use condition with confidence interval for cases in which the data consists of accelerated life testing results and a sample of load spectra. The result of an illustrational evaluation is shown and concepts for further extensions of the method are introduced.

This paper focuses on data-driven remaining useful life prediction using ensemble methods for prognostics and health management. An important factor for the performance of an ensemble method is the diversity within the ensemble. An effective neural network ensemble method that emphasizes the generation of diversity is negative correlation learning. It is argued that for both diagnosis and prognosis, the consideration of uncertainties has a substantial added benefit over a simple point estimate. For this reason, a prediction interval is derived for the ensemble method negative correlation learning using the delta method. In the delta method, the neural network is treated as a nonlinear regression model, which is approximated by a Taylor series. A look at the derived formula of the prediction interval, emphasizes that negative correlation learning behaves inversely to a regularization. Furthermore, the formula for a diversity parameter of zero is equal to the prediction interval of the regular multilayer perceptron.

For an efficient operation of a low voltage PMSM an optimized voltage usage is very important. Because of the relation between the low voltage and the high currents in this type of machine, a large voltage reserve is needed to compensate the influence of parameter mismatches and to guarantee a stable current control. As the power is limited by the low voltage in this type of hybrid drive systems, optimizing the voltage usage is also required to maximize the power and the torque availability. This paper describes a closed loop flux control to maximize the voltage usage. The controller feedback is used to estimate and maximize the available torque for each operating point.

For a low voltage IPMSM used in a hybrid drive system of a consumer car, it is of the highest importance to design a torque controller circuit that produces an accurate torque at the shaft. The accurate torque is needed to distribute the load between the combustion engine, or the manual break, and the electrical drive. As the capacitance of the batteries used in this type of car is usually very small, the control of the batteris state of charge and its output current is quite critical. Therefore, a precise torque control is elementary. Temperature changes have a large impact on the IPMSM internal parameters. Especially the permanent magnet flux and the stator resistance are affected by temperature changes. There are techniques to observe and calculate the temperature variation of these parameters. This contribution describes a method to handle the influence of temperature variation on the actual torque at the shaft, by correcting the current commands of the open loop controller.