A typical problem studied in the robust control literature involves design and analysis of a controller for a system with given actuators, sensors, and uncertainty description, On the other hand, most process control applications are more complex and less precisely defined, involving several additional tasks including development of an uncertainty description, control structure selection, and design of constraint/failure handling schemes. This paper introduces tools to handle such issues in the context of a sass study involving a high parity distillation column. It is shown that the structured singular value theory is a convenient framework to develop an integrated control structure selection and design method. This is demonstrated on the sensor selection and inferential controller design problem for the distillation column. In addition, it is shown that on-line optimization can be incorporated into the designed controller for handling actuator constraints and failures efficiently. Simulation results based on the linearized model show the potential benefits of applying the robust control theories to these types of processes, and we highlight some of the problematic points that we encountered along the way.