Scheduling, optimization and control of multipurpose batch plants

Multipurpose batch plants are often designed to concurrently produce a number of high-value products using a common collection of batch and semicontinuous processes. Because of their increased use in a number of emerging, small scale chemical industries, the design and planning of batch plants are receiving increased attention.; The objective of this work is to develop a general model for the optimization and control of complex multipurpose batch plants.; The system is modelled as a discrete event system by using Minimax Algebra which provides a framework to write linear equations that describe the performance of the plant. The state variables in these equations quantify the time at which designated events occur.; A multipurpose batch plant is viewed as a network of activities. The sequencing graph of the system helps determining an effective set of variables that denote the nodes and the arcs of the graph. The advantage of this representation is that the resulting model is very general and can describe many process features such as storage, equipment setup times, parallel processing at the same stage, and alternative production routes. The model of the system is formulated as a mixed-integer linear program. Binary variables denote the arcs of the graph and represent the processing sequence in each unit. Continuous variables denote the nodes of the graphs and represent the starting times of the activities.; The Model Predictive Control (MPC) of multipurpose batch plants is also discussed. MPC uses an explicit and separately identifiable model and optimizes an open-loop system to implement closed-loop control. The on-line calculation involves the solution of a mixed-integer linear program. The computed control consists of selecting the sequence, timing, and processing paths required to bring the system from a measured state to a final state.