Modularization of Large Process Flowsheeting Models | AIChE

Modularization of Large Process Flowsheeting Models

Authors 

The concept of digital twin(DT) as a virtual model of a physical process was originally proposed by Grieves[1]. DTs are aiming to simulate a process in a digital environment and compare it with the physical plant, targeting various benefits like process optimization and anomaly detection[1].

Chemical plants consist of process equipment and instrumentation, making the processes very complex[2]. Consequently, the corresponding plant-scale DTs are equally complex, resulting in three major issues. First, the simulation speed (real-time factor) of high-fidelity, plant-scale dynamic models is low. Second, the scope of plant-scale simulations creates operational challenges for the users. Third, the versatility of the tool is limited, when exploring scenarios that require changes in the process model, since this exercise is often time consuming and requires specific expertise.

To overcome these issues, we proposed a modular approach, where the main flowsheet (covering the full asset), is composed of sub-flowsheets, modelling the different areas of the plant (i. e. compressor area, reactor area, etc.). Within each sub-flowsheet, it is possible to design several modules, where each module shares the same inputs / outputs. However, the internal design of the each module has a different level of fidelity. For example, a compressor area sub-flowsheet might contain a full-fidelity module and a low-fidelity module. The full-fidelity module uses a rigorous centrifugal compressor model, including the ancillary equipment items. The low-fidelity module consists of a simplified compressor with a specified discharge pressure. The simulation operator activates either, the full-fidelity, or the low-fidelity module, based on the fidelity needed by his research objectives. This modularization provides a versatile selection of process scope, increasing the processing speed of the targeted research.

  1. Michael Grieves, white paper (2014), Michael W. Grieves, LLC
  2. Eng. Res. Des. 87 (2009) 1430–1437