Ansys Engineering Knowledge Manager 1.1

13 October 2008

The manufacturing world is used to managing workflows and managing data, but can the same be said of simulation? In the first of two articles, Al Dean takes a look at Ansys Inc’s brand new offering to help gain control of simulation data and processes

Product Engineering Knowledge Manager
Company name Ansys Inc
Price On Application

Simulation is a field that’s both mature and growing exponentially at the same time. The traditional realm of the finite element analyst or the computational fluid dynamic analyst is one that’s well established within many industry sectors, and it’s getting entrenched into more, with a new breed of engineer and design focused technology gaining traction in the mainstream.
As with the 3D CAD revolution before, adoption brings a whole new set of problems, and as with any technology which provides benefits, the quantity of data we’re generating is growing. As each design iteration is tested, validated, assessed and reported upon, data is created, but unlike many design-related processes, it’s done in an ad-hoc, often random manner.

The EKM workflow tools allow you to define all manner of process, batch processing and automation using the information you have available, and provides you with the key to capturing knowledge and reusing it.

Some organisations might use PDM/PLM systems to capture simulation data as part of the workflow, or allow the user to link it to the generating CAD data, but systems that support that explicitly are few and far between (think Windchill, Teamcenter and Enovia). The majority are using file structures, shared network drives and such, and the whole thing can get very messy indeed. The use of advanced simulation tools can generate massive datasets. We’re not talking gigabytes: we’re talking terabytes of data at a time in set-up and results data.

On top of this, consider the wealth of knowledge contained within this data and the work put into it. When that data is stored and perhaps lost and then the knowledge goes with it, there’s no method of capturing that expertise and reusing it. And with today’s skills shortage across the whole industry, the big picture adds up to the need for something a little more specialised. Enter industry heavyweight Ansys and its Engineering Knowledge Manager (EKM) system, released in April. Let’s take a look at the first release.

EKM is a hosted application, so it resides on a server, whether that’s a centrally located resource in your office, central and accessible over the web, or just installed on a single user machine (all of which are possible). The idea is that you access it through a standard web browser. What’s interesting with this, compared to other ‘web-based apps’, is that you can do so with any browser, on any platform, so PC, UNIX and Mac users are okay (which makes a nice change). Accessibility is key, so this is absolutely essential. You can remotely log in and access your simulation data and files, run scripts and visualisation tools, and even kick off simulation runs. Out of the box, the system is configured to integrate with all of Ansys’ applications, whether that’s Ansys Structural, Workbench, Fluent, CFX and now Ansoft, but of course, and it’ll also work (with a bit of effort) with other simulation codes, whether bought in or in-house.

Whatever system you’re working with, EKM works pretty much like any data management system. The data is kept in structured folders (which are customisable) within a vault, access is controlled by log-in by groups or individuals, either on a per folder or individual file basis. The UI is very spare, with the file property and interaction window taking up most of the window, with a series of navigation options to the left, namely, ‘applications’, ‘my processes’, ‘my work items’ and ‘repository’ – all of which will become clear.

The first step, as you might expect, is to add data. When you upload data, the system does two things. Obviously it transfers the data to the vault for storage purposes, but alongside this, the system strips out all of the metadata from it and stores it in a separate database. EKM is tuned for simulation data, so alongside the usual part and file name, date and author attributes, it also extracts all manner of simulation related data – whether that’s the solver code used, boundary conditions or materials.

The segregation of this data is key. Firstly, it makes searching based on simulation specific terminology and properties almost instantaneous. Secondly, it allows you to find the data you want more quickly, without having to load potentially huge datasets. Lastly, it means you can conduct very quick comparisons of files, again, without a data loading overhead. When you do want to look at the data, the system uses cut down viewing applications to allow you to make a final check that you have the correct data before you start loading it – if you’re accessing terabytes of data over a web connection, you’re going to want to make damn sure it’s the right data.

Once you have your data within this central location, you can make it accessible to whoever needs it. This is one of the big ground level benefits of EKM, the ability to centralise data and make it accessible. By doing so, you widen understanding and the ability to reuse the knowledge it embodies. But what if you want to push things further? Then we get into the realms of standardisation of workflows and processes – and that’s something EKM is working towards.

Alongside the storage, search/retrieval and accessibility factors inherent within EKM, it has a set of tools that allow you to start to capture knowledge about how simulation tasks are performed, formalise them into templates that contain that knowledge, then distribute those templates to a wider audience. Using a series of links to the various applications in a workflow charting type interface, you can very quickly take a signed off simulation process and create a closed workflow that steps the user through the set-up, definition of a problem, calculation and results interrogation. The set-up process is pretty easy assuming you have a solid knowledge of what you’re doing, and time invested in a thorough look at how your experts conduct common simulation tasks can pay real dividends, as that data is captured and reusable and the more you reuse it, the more value you derive from the whole process.

There’s a huge information bleed in every organisation unless you’re using tools to capture the many things that make up your intellectual knowledge. Within simulation, you have the perfect embodiment of both the digital model and how it correlates to the real world, with real world constraints and conditions and under real world loading conditions. When you look at the current move towards using real test data to further enhance the accuracy of simulation tasks, greater multi-physics capabilities coming on-stream, all of which reduce the use of assumption, then it’s clear that anything that relates to controlling and managing simulation is absolutely critical.

In conclusion

As a technology, EKM is interesting. At first sight, those with experience of PDM and PLM will see it as pretty rudimentary, with check in/check out, file storage, access control, and workflow management. And they’d be entirely correct. It is pretty basic on the surface, but underlying that simplicity is one hell of a lot of functionality and capability. If you look at the way the system extracts, stores and maintains all manner of metadata, way over and above what you’d get with a standard data management tool, you have the ability to find the information you want, to a very granular level. That granularity means that you can locate the simulation data you’re looking for more quickly, giving greater potential for re-use. The fact that it can do so when working with extremely large datasets makes it all the more impressive. This is the first area where the system should be of interest to those with large quantities of simulation data, from small organisations right through to large scale, multinational enterprises.

On top of this, you have the ability to capture knowledge and reuse it, to distribute simulation processes to a much wider audience in a safe, controlled manner, making the whole system look much more interesting. If there is a concern, it’s the simple fact that with the inherent link between 3D geometry and simulation data, it would make sense to have everything within a single, searchable database and available from the same interface. But when you look at what else can be done outside of pure management and control, then the argument diminishes. Ansys is not alone in trying to add some order to simulation data – other vendors are attempting the same thing, although each with slightly different approaches. The good news is that Ansys has added a lot of value that can, once you’ve captured that data, start to squeeze additional value from it.

At a glance: Ansys EKM

Storing simulation data, results and other information
Server based database that indexes based on simulation metadata, enabling quick searches for information
Browser and platform independent so anyone can access the system, from wherever they need to
Controlled but pervasive access to critical performance information about your products
Ability to formalise workflows and standard simulation routines
Distribute access to simulation routines in a controlled and central environment
Integration with ANSYS simulation codes including Ansys Structural, Workbench, Fluent and CFX, as well as potential to integrate with third party apps and in-house code

Comments on this article:

Leave a comment

Enter the word you see below: