Programming Strategies Of Sequential Incremental-Scale Subproblems For Large Scale Data Reconciliation And Parameter Estimation With Multi-Operational Conditions
Zhengjiang Zhang,Zhijiang Shao,Junghui Chen
DOI: https://doi.org/10.1021/ie504977k
2015-01-01
Industrial & Engineering Chemistry Research
Abstract:Data reconciliation and parameter estimation (DRPE) is a crucial issue in model-based applications, such as real-time optimization and process control. In order to obtain more reliable parameter estimates, a series of measurement data sets from different operational conditions will be used for DRPE problems. However, the dimensionality of DRPE problems increases directly with the number of measurement data sets. The number of degrees of freedom in DRPE problems is usually very large. Therefore, it is very difficult to solve the DRPE problem with multioperational conditions. On the basis of the characteristics of the DRPE problem, two directions, including the direction of incremental objectives of the DRPE problem and the direction of incremental parameters of the DRPE problem, are considered to decompose the original DRPE optimization problem into a series of incremental-scale subproblems. Three programming strategies are proposed to solve a series of incremental-scale subproblems one by one. The solutions to the current subproblem are used as a set of good initial guesses of the next optimum subproblem after including a new subproblem. By solving a series of subproblems, the optimum values of the original large scale DRPE optimization problem can be derived efficiently. The effectiveness of the proposed strategies can be demonstrated through two industrial processes, including free radical polymerization of styrene and the purified terephthalic acid oxidation process system.