A simultaneous‐modular approach to process flowsheeting and optimization. Part I: Theory and implementation

Abstract

Three basic problem formulations for the simultaneous-modular approach are discussed, as are three alternative methods for computing the flowsheet-level Jacobian. A new algorithm for partitioning and tearing when using the Simultaneous-modular approach is also presented. SIMMOD, our implementation of the simultaneous-modular approach is then outlined. In subsequent papers in this series we use SIMMOD to study the performance of the simultaneous-modular approach on process simulation and optimization problems, as well as to perform experiments on different numerical strategies for implementing the simultaneous-modular approach.

Parallel processing strategies for chemical process flowsheeting

“…In this case, however, there are further opportunities for parallelization. There are two variations of the SimM approach: a tear-stream formulation (for example, Biegler, 1985;Chen and Stadtherr, 1985) and a connecting-stream formulation (for example, Jirapongphan et al, 1980;Trevino-Lozano et al, 1985). Since the two formulations differ in their potential for parallelism, we will discuss them separately.…”

Analysis of parallelism in modular flowsheet calculations

“…Now consider the simultaneous modular approach. Chen and Stadtherr (1985a) give an excellent discussion of three basic formulations of this strategy which we briefly review. In formulation I all connecting streams in the flowsheet are torn and treated as two separate streams, one input and one output, resulting in an unnecessarily large system of nonlinear equations to be solved.…”

“…Using the description of Chen and Stadtherr (1985a), mathematically the flowsheet-level equations for formulation I1 can be partitioned as:…”

“…The purpose of this work is to illustrate possible advantages of parallelism in the modular process flowsheeting problem. The underlying numerical strategy is similar to the simultaneous 1based approach, while offering the potential for speed and flexibility, has acquired a poor reputation for reliability (Chen and Stadtherr, 1985a). In addition, from a parallel processing viewpoint the large storage requirements and complexity of the computing routines needed by this approach (such as the sparse matrix routines) will require considerable software engineering to implement.…”

Object-oriented simulation of hydrometallurgical processes: Part I. Requirements and implementation

“…[8] The differences between these programs are concentrated in the way that flowsheet equations are solved, and the facilities provided. [9,10] Many process simulators are designed to handle various problems concerning flowsheet analysis, [11,12,13] flowsheet optimization, [14][15][16][17][18] and simultaneous process optimization and heat integration. [19] The need for a detailed simulation of plants that use units and streams which differ greatly in structure from those of a conventional chemical plant leads to the construction of specific simulators which treat the particular processes in a more dedicated way.…”