(104c) Achieving Targeted Product Mix Using Web-Based Optimization Tool for Chemical Plants with Multiple Products

Many chemical manufacturing plants are designed to co-produce multiple products which may differ with respect to factors such as chain length or degree of substitution. Market demand will determine the actual production requirement for each individual product and may vary widely over the course of a month, year, or longer. Multiple controllable conditions can be used to achieve the desired production mix and suppress unwanted by-products. The sheer number of variables and process constraints make meeting multiple production requirements a challenging task. A tool that guides decision-making using near-real time optimization control can maximize the overall profit of the plant. Moreover, coordination with other plants is also very important to meet the strategy of the business. Currently, this multi-dimensional control process highly relies on operators’ experience aided by numerical calculation using tools like Excel which support what-if scenarios. Thus, to find the optimal operation conditions for each plant in real time is inefficient and may not yield the optimal result. The goal of this research work is to develop an inverse optimization tool, which uses desired production mix and other decision variables to determine the optimal control conditions, to help operators make the decision more efficiently and reliably. This web-based optimization tool will combine existing numerical models and safety constraints. This tool has helped production plants to increase production rate and achieve additional margin. Ongoing change management and improvement efforts for the tool will also be briefly highlighted.