Projects

Identifikation & Regelung

  • Online-Identification of control parameters in electric drives for parameter tracking II
    The main goal of the project is the identification of the main parameters for electric drives with elastically coupled mechanics without additional sensors. This is achieved firstly with dedicated identification trajectories and secondly during operation by observing the parameter identifiability with respect to the features of the trajectory. The basis is a mechatronic model of the overall mechatronic system.
    Leaders: Prof. Dr.-Ing. Tobias Ortmaier
    Team: M. Sc. Mathias Tantau
    Year: 2017
    Sponsors: FVA, AiF
    Lifespan: 01.07.2017 bis 31.01.2020
  • Automated commissioning of electrical drive trains - Use of temporary commissioning sensors
    In order to gain detailed models for the automated design of feedback and feedforward control, prior knowledge can be combined with structure and parameter identification in various ways. It is currently already possible to identify models of electric drives by means of experiments and to use them for control design, Kalman filtering, online state and parameter estimation, etc. The problem remains that the uniqueness of the structure identification is not always given. As a possible solution, temporary commissioning sensors could be used, such as load-side position sensors, accelerometers, strain gauges, yaw rate sensors, and so on. However, little research has been done on how the position, type and quality of the temporary sensors contribute to solving the ambiguity problem and to improving the accuracy of the models. In addition, there are hardly any approaches for the systematic assessment of structure distinguishability. Research objectives: 1.) Higher quality and robustness of the techniques from previous projects due to temporary commissioning sensors 2.) Reduction of commissioning time by identification of the mechanical plant 3.) Commissioning box (low-cost hardware, or smartphone app) 4.) Broad field of application through automatic selection of the model structure 5.) Evaluation of the benefits of temporary commissioning sensors The temporary sensor can be reused after commissioning on other testbeds, or otherwise the expenses are affordable, as low-cost hardware is used. The methods developed are to be validated on various testbeds and integrated into the e-Workbench (FVA 827)
    Leaders: Prof. Dr.-Ing. Tobias Ortmaier
    Team: M. Sc. Mathias Tantau
    Year: 2020
    Sponsors: FVA
    Lifespan: 01.02.2020 - 28.02.2021
  • Automated Control Design based on (partly) automatically generated, Control-otpimal Models
    Model-based control design enables objective, reproducible controller settings and, if required, online parameter adjustment. The controller is tuned according to simple calculation rules or by means of simulations without costly experiments. The difficulty, however, is to determine the appropriate model for a given system. When modeling, it is often not clear which physical effects must be taken into account in the model and thus in the model-based controller, or which effects can be neglected, e.g. elasticity, friction, dead time, slack. The decision on a suitable control concept (e.g. phase-lead network vs. notch filter, ...) also requires a lot of expert knowledge. In this project, (partially) automated models are to be selected and model-based controllers designed. Criteria such as maximum dynamics and minimization of the overshoot lead to different parameterizations. On the basis of these and other application-dependent criteria that can be specified by the user, an overall concept for the model and controller should be generated. As a result, a significantly reduced commissioning time can be expected with comparably good controller properties. The evaluation of the models for the model selection must be carried out based on control engineering criteria. This is in contrast to commonly used methods such as cross-validation and information criteria. Possible approaches arise from the area of ​​model predictive control.
    Leaders: Dr.-Ing. Mark Wielitzka
    Team: M. Sc. Mathias Tantau
    Year: 2021
    Sponsors: FVA
    Lifespan: 01.03.2021 bis 28.02.2022