| In the competitive global market, manufacturing systems, which typically contain manufacturing cells, must cope with unpredictable market demands. To handle these unpredictable demands, the manufacturing systems and their cells require rapid and cost-effective changes of both their hardware and software. Depending on their functional changes, the cell controllers must be changed quickly and correctly. However, the creation and debugging of the cell controller is very time intensive and laborious; therefore, current design practice cannot cope with such quick and correct changes over the systems' lifetime. To reduce the time for the creation, reconfiguration, and debugging of the cell controller, a cell controller design method must efficiently generate a correct cell controller. This dissertation describes a new modular design method for a cell controller, which focuses on the control of normal automatic sequences and error handling. This work consists of three separate parts, which complement each other.; In the first part of this work, we developed a modular cell controller design method with an automatic code generation algorithm. We proposed a new structure of the cell controller, which was divided into two important control functions: resource allocation control and operation control. For the operation control, we developed three basic operation blocks integrated with error handling control, as well as an automatic code generation algorithm. In addition, we proved the correctness of the operation control.; In the second part of this work, we proposed an error recovery planning and execution system for automatic error recovery and developed an automatic error recovery method for its implementation. This automatic error recovery method includes the design of the cell controller and formulation of an error recovery planning problem. To efficiently design the cell controller, we developed a modular design method. To formulate the error recovery planning problem, we defined situations, goals, and actions; and presented a simple planning algorithm. This planning algorithm generates the necessary error recovery actions so that a human operator does not have to. The error recovery planning and execution system can improve current manual error recovery practice, which may cause unnecessary downtime due to human mistakes.; In the final part of this work, we investigated the design efficiency of resource allocation control method with respect to deadlock problems. We applied two existing deadlock-free resource allocation control methods to an example RMS with complex part routings. We addressed new deadlock problems (higher-level deadlocks) that were not considered in these two existing methods and developed a new method to avoid such higher-level deadlocks. This resource allocation control method can be used to guarantee the deadlock-freeness in the cell controller.; Together all these contributions provide a unique design method to quickly and correctly create important functions of the cell controller for reconfigurable and flexible manufacturing cells. In addition, the automatic error recovery method can help improve current manual error recovery practice, which can cause additional unproductive downtime. |
