Out-of-the-box thinking speeds FEA
An application written with the Femap API turns three weeks of modeling into an automated, three-minute, forms-driven operation
Automating rocket chamber analysis
Saratech, Inc. provides industry expertise in assessing, delivering and setting up product lifecycle management (PLM) technology. This includes training and mentoring. Services cover a broad range of industries – aerospace, automotive, consumer products, life sciences, government, energy and utilities, and machinery and industrial. In 2009, Saratech acquired Saber Design and Analysis Services, a specialist in finite element analysis (FEA) software, support, training and consulting, with a range of FEA services including structural, thermal, computational fluid dynamics (CFD) and explicit transient dynamic analyses. Saratech also represents Siemens PLM Software as a reseller of the Femap™ software FEA pre- and post-processor and the NX™ Nastran® software FEA solver.
When D&E Propulsion and Power, one of the long-term customers of Saratech (formerly Saber) contacted John Stewart, now chief operating officer at Saratech, about doing some structural analysis related to the design of rocket thrust chambers, Stewart saw an opportunity to take advantage of Femap in a way that people often overlook. Stewart explains, “What is the fastest, most efficient way to get the answers needed for making engineering decisions? Sometimes the fastest way is to build a model in the traditional sense. But in this case, due to the repetitive nature of the work, it made more sense to build an application that would perform the work automatically. This involves using a capability of Femap that many people don’t take advantage of – the application programming interface (API). It’s more typical to think of Femap as a pre- and post-processor, but it’s also a very strong development environment.”
Eight hours of work done in seconds
The task that Stewart decided to automate was that of building finite element models for evaluating plastic strains in the hot gas walls of regenerative liquid rocket thrust chambers. “This kind of analysis involves highly transient thermal and pressure loading, and we’re trying to evaluate plastic strains in various areas of the cross sections,” he explains. “You have to look at a number of different cross sections, which were traditionally built manually.”
Doing this work by hand typically required at least three weeks to construct all the necessary models. (The run time was relatively short since these were 2D sections.) “Because the modeling took so long, our client was limited in how many stations he could evaluate and the number of design permutations he could run. The result was that analysis could not be used to optimize the design,” Stewart adds.
Using the Femap API, engineers at Saratech wrote a forms-driven Femap application that builds the models automatically. The application is named “RTE File Reader by Saratech.” “RTE” refers to a batch-style Fortran program that is used to perform a single steady-state thermal analysis of the subject thrust chamber. To use the application the user selects an RTE output file, specifies mesh size (for either a structural or a thermal model), defines materials and then chooses which stations to model. There’s also the option of selecting all stations, which could be as many as 60. The application then reads in the Fortran data and creates each model one by one, showing its progress at the bottom of the screen. Each model requires only a second or two, “which represents about eight hours of work by hand,” Stewart says. The application also verifies that temperature interpolations were done correctly, loads the models and creates local constraints. An additional step the application completes is the construction of a 3D model for transient modeling. “That takes about two weeks to do manually, so it just wasn’t done,” he adds.
Standalone program too
This application runs in its own graphical user interface inside Femap, but Saratech’s engineers also compiled it to run as a standalone program that calls the Femap library. “This is an advantage of Femap,” Stewart explains. “There are other FEA programs with scripting languages but those are proprietary and there’s no capability to easily integrate the applications you create into other programs. Because Femap uses Visual Basic, you can use it as a component, essentially running Femap in the background, in any (Microsoft) Windows application you develop or any Windows application that is compatible with Visual Basic.”
Using Femap and its API, Saratech gave its customer the ability to perform almost an unlimited number of structural analyses, ultimately enabling the company to use FEA to evaluate and optimize its designs. “We’ve taken the intelligence of an experienced analyst and made it so that any engineer on the program can do the fairly complex nonlinear structural analysis,” Stewart says.
This kind of programming capability greatly expands the possibilities for automating rocket engine design. As David Mohr, president of D&E Propulsion and Power, explains, “Traditionally, the engine system energy balance, the combustion chamber’s thermal analysis, and the combustion chamber’s structural analysis have been very different and distinct jobs. While a certain amount of ‘optimization’ could be performed if enough money were available, patience usually ran out pretty quickly. Our short-term vision is an integrated environment where these tasks are carried out simultaneously.” The RTE Reader will be a component inside that environment. “An overall program would drive Femap transparently and would pass data from one application to the next, allowing many months of engineering in rocket engine design to be done very quickly,” Stewart adds.
Customers in other industries who have seen demonstrations of the RTE Reader quickly grasped the possibilities for automating their own repetitive analysis tasks or creating parametric-based FEA modeling systems within Femap. There are substantial business drivers – including engineering costs and time-to-market – that make this type of development attractive, Stewart notes.