Process simulation and computational fluid dynamics (CFD) are widely used technologies in computer-aided process engineering (CAPE).  Process simulation is used to perform material and energy balances for plant design and analysis.  CFD is used to calculate fluid flow, heat and mass transfer, and chemical reactions within individual items of process equipment, such as stirred tanks, reactors, fluidized beds, bubble columns, combustion systems, fuel cells, spray dryers, and others. 

Integrating process simulation and CFD software tools improves workflow between the process-engineering phase and detailed-engineering phase of the CAPE lifecycle.  Such integration also enables companies in the process industries to better understand and optimize the fluid mechanics that drive overall process performance and efficiency.  The optimization of individual equipment items using CFD can be performed within the context of the whole flowsheet, so that a global improvement is achieved, not a local one at the expense of another part of the process.  In addition, the process engineer can ensure that the impact of overall process performance on critical equipment items is not overlooked; for example, reactor catalyst temperature remains below sintering temperature or the proton exchange membrane in a fuel cell does not dry up.  Another advantage in combining the two technologies is that they are then based on the same physical properties and reaction kinetics.

Since October 2000, a contractor team consisting of Fluent, ALSTOM Power, Aspen Technology, Intergraph, and West Virginia University has been collaborating on a U.S. Department of Energy (DOE) Vision 21 project to couple process simulation and CFD software.  One of the authors (Zitney) was with AspenTech during this work.  The purpose of the Vision 21 program is to enable the integration of advanced power generation and chemical production technologies into 21st century systems that improve performance, energy efficiency, and environmental protection (NETL, 2001).  To achieve the program goals, the DOE identified the need for an integrated suite of simulation software tools that seamlessly integrates process simulation models together with CFD equipment models and other custom equipment models.  The Fluent-led project team developed such a system based on the Aspen Plus® process simulator, FLUENT® CFD models, and proprietary equipment models from ALSTOM Power.  The software integration was accomplished using the process-industry standard CAPE-OPEN (CO) interfaces for unit operations, physical properties, and reaction kinetics.

In this article, we describe the CO-based software framework and integrated workflow for configuring and running coupled Aspen Plus and FLUENT simulations.  We also highlight several chemical process and power generation applications simulated using this integrated CFD and process simulation capability. in the areas

Based on the CAPE-OPEN (CO) interfaces, the software framework for integrating process simulation and CFD is described by Osawe et al. 2002.   In this framework, the CO-compliant process simulation executive is Aspen Plus (Process Simulator ).  The Process Simulator can access any CO-compliant unit operation models including FLUENT (CFD), Custom Equipment Models , and Reduced Order Models .  Aspen Plus and FLUENT are general-purpose modeling tools.  Custom Equipment Models are CO-wrapped engineering models specially developed to describe an equipment item.  Reduced Order Models are CO-wrapped fast models based on previously computed CFD results.  The information exchange between all these elements is mediated by the Controller software component.

The Controller consists of an implementation of the CO interfaces, a COM-CORBA Bridge , graphical user interfaces (GUI), and a CFD Viewer .   We exploit the CO unit operation interface to use (e.g., create, edit, solve) FLUENT CFD models in an Aspen Plus flowsheet.  This interface also facilitates the bi-directional exchange of stream information (flow rate, temperature, pressure and composition) between Aspen Plus and FLUENT.  The CO thermodynamics and reaction kinetics interfaces are used to send Aspen Plus physical property (specific heat, density, viscosity, thermal conductivity, and molecular weight) and reaction data (Arrhenius constants) to FLUENT.

The COM-CORBA Bridge , which is the enabling software design pattern, allows a CO-compliant process simulator running under the Windows operating system to exchange information with CO-compliant models running under a different operating system.  COM implementations of CO interfaces are available in Aspen Plus.  CORBA implementations of CO interfaces are included in FLUENT.  Thus, for example, Aspen Plus running under the Windows 2000 operating system can communicate with FLUENT running under the LINUX operating system via the COM-CORBA Bridge .

The Controller has the ability to present the process engineer with a Model Selection GUI and a Model Edit GUI .  The Model Selection GUI allows the process engineer to select a CFD model from the CFD Model Database .  The Model Edit GUI allows the process engineer to edit parameters for the selected CFD model.  For example, a typical CFD parameter for a stirred tank reactor is the impeller speed.  For a fuel cell, voltage and current are common parameters.  The CFD Viewer allows the process engineer to view, within the process simulator, the results of a CFD simulation conducted as a part of an integrated simulation.  The CFD results include contours of velocity, temperature, pressure, and species mass fractions for a specified surface in the equipment item.

The CFD analyst uses the Configuration Wizard to convert a FLUENT CFD model into a CO-compliant model for use in a CO-compliant process simulator.  The CO-compliant CFD models are stored in a CFD Model Database .  The process engineer can browse and select a suitable CFD model by using the Model Selection GUI .  The user selects a CFD model by browsing a summary of the model features, rather than by guessing the features from a file name.

Another component of the software system is a Reduced-Order Model (ROM) framework.  This refers to a class of models that are based on CFD solutions, but are much faster than CFD models.  The ROMs are based on CFD solutions stored in the CFD database for a range of parameter values.  For example, we provide a ROM based on (piecewise) multiple linear regression, to demonstrate the concept.  Using the solver grid in the Model Edit GUI , it is possible to define a solution strategy that uses a combination of CFD and ROMs to simulate an equipment item; for example, the initial flowsheet iterations are based on a ROM and the final iterations are based on a high-fidelity CFD model.

Another component of the software architecture is a class of models called Custom Equipment Models .  These models are legacy computer programs (largely developed by industry) based on empirical information obtained from many years of experience in designing and operating certain equipment items.  They are typically very fast and accurate within the confines of a certain parameter range.

The workflow used to integrate CFD and process models is described here.  A CFD Analysis sequence consists of a CFD analyst developing the CFD model, using the Configuration Wizard to make the model CO-compliant, and adding the model to the CFD model database.  

The process engineer sets up the process model and selects CFD models to represent one or more unit operation blocks in the process diagram.  After placing the CO-CFD model icon on the process flowsheet, the process engineer uses the Model Selection GUI to select the CFD model of interest from the database.  The process engineer changes parameters in the CFD model by using the Model Edit GUI .  (Those are the parameters that the CFD analyst specified as editable during the configuration step.)  The process engineer may also define a solution strategy using the Model Edit GUI .  The solution strategy consists of a choice of different models used to represent the operation block and may, for example, include CFD models (coarse grid/fine grid, 2D/3D), reduced-order models, or custom equipment models.  The purpose of the solution strategy is to obtain a desired level of accuracy and speed.  The process engineer then conducts the integrated simulation.  After obtaining a converged solution, the process engineer views CFD results using the CFD Viewer in the Model Edit GUI .  The stream information at the outlet of the CFD unit operation block is obtained as usual from the process simulator.

A recent review of industrial applications of the CAPE-OPEN standard, including a brief discussion of the integrated process simulation and CFD solution described here, can be found in Pons (2003).  In this section, we highlight six integrated Aspen Plus and FLUENT applications, including two industrial power generation applications from ALSTOM Power.

In a chemical process application, Zitney and Syamlal (2002) coupled a 2D FLUENT CFD model of a stirred tank reactor model into a reaction-separation-recycle flowsheet in Aspen Plus.  The information exchange between Aspen Plus and FLUENT includes two material streams and a CAPE-OPEN parameter for the shaft speed in the reactor, along with physical properties and reaction kinetics.  The integrated simulations are used to determine an optimum shaft speed (CFD model parameter) for maximizing the rate of production of one of the products. 

In a fuel cell application, O’Brien et al. (2003) recently coupled CFD and process simulations to analyze high-temperature, auxiliary power units (APUs) based on solid oxide fuel cells (SOFCs).  With fuel-to-electricity conversion efficiencies approaching 50%, fuel cell APUs can dramatically reduce fuel consumption, cost, and pollutant emissions for the idling of heavy-duty truck engines.  A 3D CFD model is used to simulate the fluid flow, heat and mass transfer, electrochemistry, and current distribution in the SOFC.  The information exchange between Aspen Plus and FLUENT includes four material streams (inlet/outlet for cathode and anode) and two CAPE-OPEN parameters, namely the current and voltage in the SOFC.  Process simulations are used to perform overall material and energy balances on the tightly integrated APU flowsheet consisting of equipment items such as a reformer, desulfurizer, fuel cell stack, combustor, and various heat exchange and rotating equipment items.  Using the FLUENT and Aspen Plus integration toolkit, coupled CFD and process simulations are performed over the current range to generate a voltage-current curve and analyze the effect of current on fuel utilization, current density, power density, and overall process efficiency. 

Syamlal et al. (2003) also presented several applications of the integrated process simulation environment to model fuel cell systems.  In one example a solid oxide fuel cell (SOFC) is modeled with a CFD-based SOFC model.  The Aspen Plus process flowsheet consists of a reformer, SOFC, post-stack combustor, and heat exchangers.  The SOFC model considers the detailed fluid flow, electrochemistry, and current distribution in the fuel cell.  In a second system, a natural gas-based, proton exchange membrane (PEM) fuel cell system is considered.  The Aspen Plus flowsheet used for the coupled simulations consists of a reformer, shift converter, fuel cell, anode exhaust combustor, and heat exchangers.  The reformer is modeled with a CFD model that calculates the 3D distribution of the flow field, temperature, pressure and concentration in the reactor.  The information exchange between Aspen Plus and FLUENT includes four material streams (inlet/outlet for the process side and hot gas side of the reformer), temperature-dependent physical properties, and reaction kinetics parameters.  When the reformer model is executed from within the fuel cell flowsheet, the CFD model benefits from the ability to account for the effect of recycle streams.  The fuel gas is heated with the products of combustion from the anode exhaust burner.  The conversion in the reformer is limited by the energy available from the hot gas, which in turn depends upon the conversion in the reformer.  Furthermore the feed stream to the reformer is preheated with the outlet stream from the shift converter and the shell outlet gas.  The CFD model accounts for the radial variation in the temperature in the catalyst bed (in the tube) and predicts conversions that account for the limitations imposed by the heat transfer to the bed.  Another advantage of the coupled CFD model is that the detailed calculations provide the process engineer with information that, although not required for the process simulation, is important for the overall system design.  In the case of the reformer model, the detailed temperature distribution in the catalyst bed is useful to ensure that the temperature anywhere in the catalyst bed does not exceed the sintering temperature. 

Representing industrial power plant applications, ALSTOM Power modeled a conventional 30 MWe coal-fired steam plant for municipal electricity generation and an advanced 250 MW, natural gas-fired, combined cycle (NGCC) power plant (Sloan et al., 2002, 2003).  In the conventional steam plant, a FLUENT 3D CFD model represents the gas-side and steam-side of the boiler.  Information exchange between Aspen Plus and FLUENT includes 25 CAPE-OPEN material ports for the gas-side streams, 16 CAPE-OPEN parameters for the steam-side streams, and three CAPE-OPEN CFD parameters representing the damper position (bypass resistance).  An Aspen Plus design specification is used to adjust a FLUENT model parameter, namely the damper position, to control the steam temperature at 763 K.  In the NGCC plant, a FLUENT 3D CFD model is used for the heat recovery steam generator (HRSG), which consists of several nested heat exchangers and pollutant control devices.  Information exchange between Aspen Plus and FLUENT includes two CAPE-OPEN material stream connections on the gas side and multiple connections using 60 CAPE-OPEN parameters on the steam side.  An Aspen Plus design specification is used to adjust the high-pressure pump feed rate to achieve steam temperature of 838 K.

This article describes an integration framework that uses the CAPE-OPEN standard for combining process simulation (Aspen Plus) with detailed CFD-based equipment models (FLUENT), custom equipment models, and fast reduced-order models (ROMs) based on CFD results.   The integration capabilities have been applied to several chemical and power generation processes to demonstrate the feasibility of the approach.  The results show that combined CFD and process simulation offers new opportunities to analyze and optimize overall plant performance with respect to mixing and fluid flow behavior.

Acknowledgements

References

National Energy Technology Laboratory (NETL), Morgantown, WV, “Vision 21 Technology Roadmap,” Available online at www.netl.doe.gov/coalpower/vision21/pubs/v21rdmp.pdf, (2001).

O’Brien, T.J., S.E. Zitney, W.A. Rogers, and A. Cugini, “Simulation Tools for the Design of Vision 21 Energy Plants,” Presented at the 226th ACS National Meeting, September 7-11, New York, NY (2003).
Osawe, M.O, P. Felix, M. Syamlal, I. Lapshin, K.J. Cleetus, and S.E. Zitney, "An Integrated Process Simulation and CFD Environment Using the CAPE-OPEN Interface Specifications,” Presented at the AIChE 2002 Annual Meeting, November 3-8, Paper No. 250c, Indianapolis, IN (2002).

Pons, M., “Industrial implementations of the CAPE-OPEN standard,” Presented at the Sixth Italian Conference on Chemical and Process Engineering (IcheaP-6), June 8-11, Pisa, Italy (2003).

Sloan, D.G., W.A. Fiveland, S.E. Zitney, and M. Syamlal, "Software Integration for Power Plant Simulations,” Presented at The 27th International Technical Conference on Coal Utilization & Fuel Systems, March 4-7, Clearwater, Florida (2002).

Sloan, D.G. and W.A. Fiveland, “Development of Technologies and Analytical Capabilities for Vision 21 Energy Plants: Final Simulation Results for Demonstration Cases 1 and 2, under Cooperative Agreement No. DE-FC26-00NT40954,” U.S. Department of Energy, National Energy Technology Laboratory (NETL) Report (2003).

Syamlal, M., Jens I. Madsen, Stephen E. Zitney, and William A. Rogers, “Application of an Integrated Process Simulation and CFD Environment to Model Fuel Cell Systems,” Presented at the AIChE 2003 Spring National Meeting, March 30 - April 3, New Orleans, LA (2003).

Zitney, S.E. and M. Syamlal, "Integrated Process Simulation and CFD for Improved Process Engineering,” Proc. of the European Symposium on Computer Aided Process Engineering –12, ESCAPE-12, May 26-29, The Hague, The Netherlands, 2002; J. Grievink and J. van Schijndel, Eds.), pp. 397-402 (2002).