In order to use the Linear MPC implicit control, the model is first linearize d about the operational point V_ p0 and L_ 10, where V_ p0 is the voltage needed to maintain the level at L_ 10 = 15 cm. We have a fairly representative nonlinear model of the Coupled Tanks system.Given this, we can choose the limit and scaling for MO and MV. T he tank level is limited between 0 to 25 cm and the tank voltage from 0 to 22 V. We could also define the disturbance tab as an Unmeasured Output (UO). The input, or Manipulated Variable (MV), is the pump voltage and the output is the measured level, denoted as the Measured Ou t put (MO). In this section, the step-by-step MPC control design procedure to control the level of water in Tank 1 of the Coupled Tanks system is outlined. Figure 1: Plant model signals used in a MPC controller (source: ) Coupled Tanks MPC Design You still need to have a representative model, but because of its adaptive nature, less control design is required and it can compensate more for effects such as parameter uncertainties and disturbances. Figure 1 illustrates the plant model signals used in the MPC controller. The MPC design starts with defining the various input and output signals and the i r limits. The MathWorks Model Predictive Control Toolbox makes it easy to design and implement MPC controllers. Model Predictive Control (MPC) is an advanced adaptive type control method that requires less knowledge of the plant dynamics. For more information about the control challenges and design using PID+FF, please see my previous blog post. If the model does not accurately represent the system, then the controller may not perform adequately and as expected. While it does perform well, the PID+FF control follow s a model-based design approach. In the supplied courseware, we go through the design of a PID and feed-forward controller (PID+FF) to control the level of water in the tank. The Quanser Coupled Tanks, shown below, is a reconfigurable, nonlinear process control experiment that is not easy to control. In this blog MPC is implemented both on the QLabs Virtual Coupled Tanks system as well as the Quanser Coupled Tanks hardware using the QUARC Real-Time Control software. So i t only made sense to test this on the Quanser Coupled Tanks using the MathWorks® Model Predictive Control T oolbox for Simulink ®. MPC was originally developed for the process control industry. Model Predictive Control (MPC) has gained a lot of popularity over the last 15 years and for a good reason – it works well with relatively little design effort. Industrial Applications & Process Control.
0 kommentar(er)
