URL study guide
https://tue.osiris-student.nl/onderwijscatalogus/extern/cursus?cursuscode=4DM40&collegejaar=2025&taal=enOmschrijving
This course provides a framework within which concepts from dynamics and control can be used to model and control manufacturing networks. We consider different kinds of models: stochastic models, ODE fluid models, PDE fluid models, and Hybrid models. We use MPC for controlling manufacturing networks, and LMIs for analysis and design of hybrid systems. Furthermore, we apply these techniques to the control of traffic lights.Topics
- Introduction to Manufacturing Systems
- Stochastic Models
- Effective Process Times
- Throughput oriented approximation models of manufacturing networks
- Model-based Predictive Control (MPC)
- State dependent switching
- State independent switching
- Switching networks
- Linear Matrix Inequalities (LMIs)
Doelstellingen
- You are able to describe a control framework in which it is possible to control manufacturing systems using methods from control theory
- You know what EPT’s are and you know how to determine them from real data.
- You can make a throughput oriented (linear) approximation model for a manufacturing system
- You can give a discrete time formulation for a sampled linear continuous time system that is controlled by a piecewise constant input. You can also do this for inputs with time delays.
- You can explain the principle of Model Based Predictive Control (MPC)
- You know that an MPC controller not necessarily results in a stable closed-loop system.
- You know how to guarantee stability of an MPC controller
- You know the three commonly used models for manufacturing systems: discrete event models, queueing models and fluid models (with discrete positions). For each model you can give the pro’s and con’s.
- You are able to use tooling for determining optimal periodic schedules for traffic lights, and you can tune periodic schedules for a network of intersections by determining proper phase shifts.
- You know the Kumar-Seidman example of a re-entrant system that is unstable despite a utilization less than 1.
- You are able to perform stabilization of linear systems via time-dependent switching, including stabilization of Kapitza pendulum via vibration control;
- You are able to perform stabilization of linear systems via state-dependent switching;
- You are able to model manufacturing machines and tandem of manufacturing machines as continuous-time systems, to solve the demand tracking problem addressing a possible integrator windup issue, to analyze the tracking performance in case of capacity saturation by the method of describing functions;
- You are able to stabilize queuing non-acyclic networks via buffer-regulators and to simulate those systems.