Optimization of the mechatronic design for multi-axes drive systems

在多轴传动的设计过程中，传动系统工程师必须面对从市场上大量可用的传动部件中选择最优部件的挑战，而这些部件只有有限的参数规格。本课题的主要目标是设计一种控制带宽高的伺服系统。传统的系统通常采用由电流、速度和位置控制组成的嵌套控制，由于环路之间的带宽损失，这种控制方式并不是最优的。为了设计低振动定位系统，通常使用坚硬的部件，因此通常使用昂贵的部件。然而，不断增加的成本压力要求在优化的多变量控制和部件的刚性之间进行权衡。在设计阶段，这种新方法需要对系统本征模的可观性和可控性进行控制工程分析，以便估计带宽。位置可观测性的一个决定因素是(位置)测量系统的位置和数量，对于传统的嵌套控制回路，这些测量系统仅限于一个编码器和一个线性长度测量系统。由此产生的某些状态的不可观测性导致了不可控状态，从而降低了带宽。前人的项目已经开发出一种最优化方法，从具有模型不确定性的模型中估计和优化最大机械带宽。通过具有组件模型的数据库进行黑盒优化，可以快速设计预定义的价格范围。然而，这种优化是基于传统的带过滤器的单轴嵌套控制。后续项目的思路是扩展多轴多变量控制的传动设计程序。此外，除了增加带宽外，还可以考虑多轴效应(如游览台的偏航和俯仰)。因此，改进了各种驱动优化的控制工程方法：借助相对增益阵列的优化，传感器的位置和选择将得到显著改善，从而显著提高系统的可观测性和可控性。H∞或LQG控制等现代控制方法是多输入多输出控制系统的基础。验证在带有实时控制系统的测试机上进行，该系统评估附加的直接测量系统并控制逆变器的占空比。

During the design process of multi-axis, drive systems engineers have to cope with the challenge of selecting optimal components out of a large number of available drive components on the market with only limited parameter specifications. The main objective is the design of a servo system with a high bandwidth in the control. Conventional systems usually use a nested control consisting of current, speed and position control, which is not optimal due the bandwidth loss from loop to loop. In order to design low-vibrating positioning systems, stiff and thus mostly expensive components are often used. However, the increasing cost pressure requires a trade-off between optimized multi-variable control and rigidity of the components. During the design phase, this new approach needs a control engineering analysis with regard to observability and controllability of the system’s eigen modes in order to estimate the bandwidth. A determining factor for the observability of the position is the location and number of the (position) measuring systems, which are limited to one encoder and one linear length measuring system for the conventional nested control loop. The resulting non-observability of certain states leads to uncontrollable states and thus to lower bandwidths. The predecessor project has already developed an optimization method, which estimates and optimizes the maximum mechanical bandwidth from a model with modelled uncertainties. A black box optimization via a database with component models enables a fast design for a predefined price range. However, this optimization was based on the conventional uniaxial nested control with filters. The idea of the subsequent project is to extend the drive design procedure for a multi-axial multi-variable control. Moreover, besides increasing the bandwidth, multiaxial effects (such as yawing and pitching of travel stands) can also be taken into account. Therefore, various methods of control engineering for drive optimization are refined: With the help of an optimization of the relative gain array, the sensor positions and selections and thus the observability and controllability of systems shall be improved significantly. Modern control methods such as H∞ or LQG control are the basis for the multiple input and multiple output control system. The validation takes place on on a test machine with a real-time control system, which evaluates additional direct measuring systems and controls the duty cycles of the inverters.