SYNTHESIS OF STEADY-STATE CONTROL STRUCTURES FOR COMPLETE CHEMICAL PLANTS (OPTIMIZATION, MODELING, DESIGN, OPERABILITY)

Traditional chemical engineering practice considers the synthesis of a process design and its control system to be separate and independent activities. That is, the design engineer assumes responsibility for selecting the optimum process flowsheet, operating conditions, and equipment design. Based on these decisions, the control engineer must address the following topics: What are the economic and non-economic control objectives? What state variables should be measured to monitor the control objectives? What manipulated variables should be adjusted to drive these variables to their desired values? What control laws should be implemented to interconnect the measured and manipulated variables?; This thesis outlines a systematic procedure to answer each of these questions. A steady-state analysis, based on short-cut equipment performance and economic models, allows rapid screening of a wide range of control structure alternatives. The resulting preliminary control structures and engineering insight provide a suitable starting point for further dynamic control studies.; This thesis also addresses the importance of the interaction between the process design and control activities. A steady-state controllability analysis is recommended at the preliminary design stage. Flowsheet structure or equipment design modifications are proposed to avoid the most significant sources of uncontrollability in chemical processes.; Because chemical plants are generally designed and optimized for a single set of operating conditions, additional equipment capacity is required to satisfy the process constraints in the face of expected disturbances. A flowsheet decomposition scheme is used to identify "bottlenecking" equipment which require overdesign. Approximately optimum overdesign factors for non-bottlenecking equipment are also calculated using short-cut performance and economic models.; This analysis is demonstrated in three case studies of complete chemical plants. In each case, simple control policies are derived to approach optimum steady-state control, providing an excellent starting point for dynamic simulation and advanced control studies.