How to Manage FEA Projects? 1

[FEAChamp]

I'm the lead analyst in a small company. I like doing FEA, and figuring things out, but they want me to be more of a manager, and know how to schedule tasks, and work towards making everything more efficient. Its driving me crazy. When I first started out of college I was basically given a project, and I ran with it. Admittedly some of them took longer than they should, and the company as a whole had some project management issues, but in the end I usually produced results. We've added schedule processes, and basic stuff like that in the past couple years. But now we're getting to the point where they want me (us, the FEA team) to have meetings to review the FEA setup before running anything, and lay out every load case individually, every review, every refinement step in the schedule so they can better track hours and its starting to feel like extreme micromanagement. But since I really just want to be an engineer, solve FEA problems, and maybe review setups and projects with other FEA team members, and I've never had a knack for scheduling/managing I don't know how to push back without just sounding like I just don't want to improve, or do what they want. I just feel like FEA is a much more unknown than the typical drawing set, or design, and i feel like I've been gently reprimanded for doing something out of order or without a review or telling everyone my plan, even though I feel like it produced actionable results, and all they get focused on is that I didn't follow the exact procedure they want, or do things how they envisioned it.

Does anyone have any good resources or insight on this topic? I'd love to discuss this with someone with experience.

 

[CWB1]

If you have no desire to be an engineering lead or manager then I would recommend sitting down with your supervisor asap to discuss your role vs expectations. There is nothing wrong or uncommon with that desire, but hopefully you recognize that it can be career-limiting, management may think less of you for it, and its likely someone who was previously beneath you may become your supervisor. If you're the FOG then that attitude is somewhat expected, however if you're fairly young you may catch some flak for it.

Regardless, the ability to accurately plan any portion of a project only comes through experience and competence. If you understand the work then you should be able to break down your portion of large projects into the smallest individual tasks, estimate the resources and time necessary for each small task's completion, and provide a fairly accurate estimate/total to the project planning folks. If it seems overwhelming or you would like useful strategies/hints, spend some time reviewing lean or agile project planning principles and techniques.

I just feel like FEA is a much more unknown than the typical drawing set, or design, and i feel like I've been gently reprimanded for doing something out of order or without a review or telling everyone my plan, even though I feel like it produced actionable results, and all they get focused on is that I didn't follow the exact procedure they want, or do things how they envisioned it.

Having worked in numerous (purely science, not product dev) research depts I have heard similar sentiments occasionally. The reality of the situation is that the right solution at the wrong time is still the wrong solution. Management needs to know when work will be completed for business planning purposes. In the grand scheme of things, your efficiency isn't as important to them as your work being done by a specific date. If you're having problems consistently hitting those dates then you need to review your work processes, ensure you're breaking projects into manageable chunks, and have reasonable expectations of staff. Quite honestly, this shouldn't even be a lead or managerial task. Any competent, individual engineer should be able to handle their own portion of the project planning and at many companies, even junior engineers do.

 

[FEAChamp]

The company organization is extremely flat. The owner of the company, and the engineering manager are my only supervisors.

When I first started 7 years ago, I didn't schedule things (pretty much a company wide issue at the time), didn't update schedules, etc. I'd get a project and work on it for a week, or two and then report back to the customer. This was resulting in lots of overages from the estimates, and obviously wasn't tenable. Now for the past couple years, its standard practice to estimate the hours for each task. Some things would still go over on hours, but it was better.

Now the push is to quantize every step of every task. Now a study is one push of the run button. Run an initial study to get a feel for how the assembly behaves? Thats its own task with X hours. Need to refine the mesh further? That's another study, and should have its own task with X hours, and possibly be broken further down into "meshing time" and "running time". Need to test the sensitivity of how the load is applied to a part/structure/assembly, or want to just get a feel for how the part/structure/assembly responds to different restraints? Every single one of those is a different study, with a different task, with hour estimates and possible sub tasks for the remeshing, changing the loads, restraints, etc. Every setup needs to be reviewed before ever hitting run, so that gets its own task, and meetings need to be scheduled way ahead of time because they pack their schedules full of meetings every day. All results should be reviewed with them before telling the customer, so the same thing. We're talking about sub-day sized tasks that everything needs to be broken out into, with my/the team member's time split up in to each bucket when we log our time. We bill our customers hourly, so not only do I have to worry about the schedule, but I also have to worry about every single hour used.


I guess I feel like if we were talking about days and week long tasks, this might make more sense to me. But to break everything down this far just sets my head spinning, and I feel like I'm never doing it right. And I feel like it totally breaks what I think of the FEA process where I test different setups, test how sensitive to setup certain things are, run some low quality mesh runs to get a feel for the assembly, then add in more mesh detail and other details. Now it just feels more like an assembly line type process instead of being an analyst.

It just feels like if I was a mechanic, I'd have to prepare a schedule and log my time into different buckets that look like:
Diagnose issue 1
Diagnose issue 2
Diagnose issue 3 (that may or may not even be an issue)
Review findings with management and get consensus on what parts/service its needs
Preparing car for work
Removing the old part 1
Putting in new part 1
Removing the old part 2
Putting in the new part 2
Removing the old part 3
Putting in the new part 3
Review the fixed car with management
Reinstalling covers and panels
etc.

That's why I'm hoping to find some resources to see how other companies manage this stuff.

 

[jari001]

Here's how I've seen managers in the past react to similar situations:
[ul]
[li]Learn some strategies from other groups that seem to be handling the transition well.[/li]
[li]Explain to upper management why their department's work doesn't fit exactly into the project planning paradigm as it currently exists and then propose the tweaks that will incorporate their work.[/li]
[li]Drag their feet and by force of personality bend everyone else to their will.[/li]
[li]Quit.[/li]
[/ul]
This increased scrutiny to every quanta of work may be an over reach on the success of the hours estimate on the larger blocks of work and you need to ride it out until the ebb comes back to something more to your liking. I hope you can proceed with option 2 above with the support of the engineering manager. Just remember, others will have no problem telling you how long your work should take so that their work will be easier, so better fend for your group!

 

[CWB1]

Now the push is to quantize every step of every task.

What you're describing is a bit absurd, rather blindly taking lean to the extreme and missing the point. Ultimately your project plan should be placing resource and time estimates on the task - complete FEA analysis of X part, Y assembly, etc, NOT each step of those tasks unless they're complex enough to require a significant amount of time. That said, each analyst should be efficiently working via the scientific method, using the software to establish and study trends. They should have a fairly good idea of how to approach the study, how many runs or how much computing time they'll need, etc before they start. Crap happens and some tasks will run long, others short, but the goal overall is to have promises line up with your delivered dates. Most engineers tend to plan everything a bit conservative and I cannot argue with that, better underpromise and overdeliver than vice-versa IMHO. One sticking point for me however is when the "business" folks try to adjust your timeline to meet their need, IME the team's original estimate is always the best and the most accurate estimates are done by the low-level team doing the actual work, NOT their managers.