Autonomous manufacturing: Planning and control in laser beam welding

Abstract

Due to workpiece varitety, short time limits and small batch lots, requiring production to be able to react to a given situation, conventional manufacturing control has to be replaced by closed loop control manufacturing. Even in laser beam welding the high complexities of the plant and the process require a highly skilled expert to operate the system. The operator is then automatically exposed to a high level of mental stress. The introduction of autonomous operating manufacturing cells becomes necessary to relieve the operator. A new approach was therefore taken, to consider planning and manufacturing as a single holistic unit of operation. Central to the autonomous manufacturing cell, is the supervision of the welding process by the simultaneous action of multiple sensors, linked to a process computer. On the one hand, this computer is responsible for coordinating and initializing the sensors. On the other hand, default or fixed-value process parameters and the workpiece design for example, can be supplied through links with the planning level. Intelligent, fault-tolerant algorithms constitute the superimposition of multiple control loops. According to this cascaded control structure, minor faults are automatically repaired by the system. Planning errors are fed back to the respective level of their origin. Only in the case of fatal errors is the operator informed and adequate action requested. The entire knowledge of the process is stored in data banks, for future application, to prevent the occurence of known failures. With respect to quality planning and assurance, it then becomes possible to predict expected processing results by extrapolation of actual data.Due to workpiece varitety, short time limits and small batch lots, requiring production to be able to react to a given situation, conventional manufacturing control has to be replaced by closed loop control manufacturing. Even in laser beam welding the high complexities of the plant and the process require a highly skilled expert to operate the system. The operator is then automatically exposed to a high level of mental stress. The introduction of autonomous operating manufacturing cells becomes necessary to relieve the operator. A new approach was therefore taken, to consider planning and manufacturing as a single holistic unit of operation. Central to the autonomous manufacturing cell, is the supervision of the welding process by the simultaneous action of multiple sensors, linked to a process computer. On the one hand, this computer is responsible for coordinating and initializing the sensors. On the other hand, default or fixed-value process parameters and the workpiece design for example, can be supp...