(204g) New Approaches to Process Simulation in the 21st Century

Traditional process simulation programs have tended to be focused on specific simulation objectives like “process design”, “hydraulic analysis”, “dynamics” or “optimization”. Consequently the library of unit operation models or equipment models in the simulation program has then been closely coupled to the kind of solver to be employed, such as “sequential modular”, “equation oriented” or “DAE integrator”. As a result, the choice of a given simulation program has tended to constrain engineers to a given workflow that matches their assumed expectations from the program. For example, an engineer may spend time developing heat and material balances in a steady state flowsheeting program, but then find the result of this work is not readily reusable for heat exchanger network analysis since network analysis requires a different kind of solver than the sequential modular one provided.

This paper will examine some features in a new breed of simulation program that can increase the power and usefulness of simulation to today’s engineer. These features can improve productivity and data reuse by reducing the need to transfer data from one tool to another or to create interfaces between different programs. Among these features are:

Separation of model representation from solution methods.A model should declare what is to be represented, independently of how that model is to be solved. A given model should be solvable in different ways according to the current needs of the engineer and it should be easy to switch from one solution method to another. Changing from sequential flow driven to hydraulic analysis should just be a matter of switching solution modes in the application.

Interactive simulation building.In the early days of computer programming the development cycle was very batch oriented. Simulation programs can often seem the same way: create a simulation, try to run it, check the error messages, modify the simulation, try again. Software development has moved to integrated development environments (IDEs) and simulation programs should do likewise. You should be able to build simulations incrementally, see partial results as you enter data, get an immediate response if there is a configuration problem, and do look ahead calculations so data can be filled in automatically for you.

Open models.The equipment models in a standard library are commonly closed to inspection, forcing a reliance on external documentation to understand what equations they implement and any restrictions or assumptions imposed. In contrast declarative models are written in the language of engineers with recognizable equations and with schematic diagrams to assist with more complex configurations. Furthermore, if a given standard model is not quite right for a specific purpose it should be easy to customize or extend that model within the simulation program without needing to deal with computer code or programming languages.

Managing complexity.Large simulations can become difficult to inspect, review and validate in traditional simulation applications due to the lack of organizational structure provided by the simulation environment. Concepts such as hierarchical levels of detail, composition and reusable templates, user-defined libraries, filtering, searching and sorting, and visual cues such as badge icons can make navigating a large amount of complex information much more manageable.

Collaboration and team work.Organizations have project teams that need to work together and share and manage data. They need change logs, revision management, audit trails, compare and merge, and shared access with access controls. A modern simulation program should provide the features for documentation and data management that are expected in a typical engineering office.

This paper will review the above topics with specific details and illustrations of use cases. The presentation will highlight ways that Invensys is actively addressing these areas in response to feedback and requests from engineers in the markets we serve.