Holistic modelling, control configuration, and design systematics for locally concentrated Multi-Motor Drive Systems - Follow-up application

2017年批准的项目主题是工业可重构多电机驱动系统，即通过变速箱连接的多个感应电机共享所需功率的驱动系统。该项目由帕德博恩大学的两个部门进行，目的是分析和建模电工和机械工程方面的问题。在电力电子和电气驱动（LEA）部门进行的电工任务研究工作中，出现了新的问题，这些问题将在一年的后续项目中得到解决。在之前的项目中，LEA开发了第一批逆变器模型，可以高精度地估计受控感应电机的相位电压。在此基础上，对异步电动机的多种非线性电磁效应进行建模，推导出一种准确估计转矩和损耗的结构。其中包含的模型参数通过记录在试验台的测量数据离线识别。该方法与热模型相结合，首次证明了同时进行电-热建模和离线识别为精确的转矩控制和有效的操作策略提供了很大的潜力。先前项目的这些主要结果发表在几个会议和期刊论文上。到目前为止，电和热子模型只是轻微耦合。在后续项目中，这两个领域现在被整合到进一步的重要方面，如对电机饱和行为的热影响和电导体中的集肤效应，这样就可以对感应电机进行高度精确的扭矩和温度估计，从而充分利用该方法的潜力。从这个多域模型中，导出了一种自适应运行策略，以便能够以效率最优的方式在整个运行范围内运行给定的电机，并且对必要的控制硬件要求很低。在文献中，通常在转矩控制或运行策略中只考虑电机的个别非线性，导致显著的不准确性或可避免的损失。后续项目的进一步要点涉及离线识别的优化测量数据采集的结构化程序，以及对识别过程中可以忽略的单个测量变量（例如转子温度）的程度的详细评估。

The subject of the project approved in 2017 were industrial reconfigurable multi-motor drive systems, i.e., drive systems in which several induction motors interconnected via a gearbox share the required power. The project was conducted by two departments of the Paderborn University in order to analyze and model both the electrotechnical and the mechanical engineering aspects. During the research work on the electrotechnical tasks, performed by the department of Power Electronics and Electrical Drives (LEA), new questions arose which are to be treated within a one-year follow-up project. In the previous project, LEA developed at first models of the inverter with which the phase voltages of a controlled induction motor can be estimated with high accuracy. Based on this voltage estimation and the modeling of numerous nonlinear electromagnetic effects of an induction motor, a structure for accurate torque and loss estimation was derived. The model parameters contained therein are identified offline using measurement data recorded on the test bench. This method was combined with a thermal model and it was shown for the first time that simultaneous electrical-thermal modeling and offline identification offers high potential for both accurate torque control and efficient operating strategy. These main results of the previous project were published in several conference and journal papers. So far, the electrical and thermal submodels were only slightly coupled. In the follow-up project, both domains are now integrated with respect to further important aspects such as the thermal influence on the saturation behavior of the motor and on the skin effect in the electrical conductors so that a highly accurate torque and temperature estimation of an induction motor is possible, thus exploiting the full potential of the approach. From this multi-domain model, an adaptive operating strategy is derived in the next step, in order to be able to operate a given motor in the complete operating range in an efficiency-optimal way and still with low demands on the necessary control hardware. In the literature, usually only individual nonlinearities of the motor are considered in the torque control or in the operating strategy, resulting in significant inaccuracies or avoidable losses. Further points of the follow-up project concern the structured procedure for optimized measurement data acquisition for offline identification and a detailed evaluation of the extent to which individual measurement variables (e.g. rotor temperature) can be dispensed with during identification.