Opening a CFD consulting firm?

[ramzerimar]

I've been thinking about entrepreneurship in the engineering field. Unlike other fields like computer science, where small companies with small investments often end up becoming huge business, engineering is usually a very expensive thing. So consulting is a way for those who don't want to work as employees.

I'm a student of MechE, very interested in FEA and CFD. I would like to hear some opinions on what are the odds of opening a CFD consulting firm. I mean, is there a market for that? And how difficult would it be (initial investment, people, etc...)? I would like to hear some opinions.

 

[Vanadium 50]

Are you being paid to do CFD now? And if not, what is your plan to get someone to pay you? If so, do you think they will become your first client?

 

[ramzerimar]

Are you being paid to do CFD now? And if not, what is your plan to get someone to pay you? If so, do you think they will become your first client?

No, I'm not being paid to do CFD right now. I'm just an undergraduate very interested in this field of study, who wants to work with that in the future. Opening a CFD firm is just something I've been thinking about for a distante future, after I have enough experience and decent contacts within the engineering field. I imagine that, as most consulting firms in engineering, people would look after CFD specialists to solve problems that a)they don't know how to solve b)they don't have the time to solve 3)they don't have the resources to solve.

Didn't understand the last question, about them becoming my first client.

 

[berkeman]

No, I'm not being paid to do CFD right now. I'm just an undergraduate very interested in this field of study, who wants to work with that in the future. Opening a CFD firm is just something I've been thinking about for a distante future, after I have enough experience and decent contacts within the engineering field. I imagine that, as most consulting firms in engineering, people would look after CFD specialists to solve problems that a)they don't know how to solve b)they don't have the time to solve 3)they don't have the resources to solve.

Usually you would start such a consulting firm (or work at one) after a bit of experience in the field. You should have a good solid resume/portolio that you can show to potential clients.

Didn't understand the last question, about them becoming my first client.

He's just saying that if you have a lot of experience already and have a client who really likes your work, you might approach them as being a client at the new firm you are starting. That could cause legal problems, though, so be careful going down that route. Especially if the contract you have signed with your current employer has a "Non-Competition Clause".

 

[ramzerimar]

But how is the market for CFD consulting firms? I mean, I know the kind of work you would do in there, but is there a demand for that in the market nowadays? I think that one of the most importants things you have to know to open a consulting firm is where to look for clients. Do big companies (like Boeing, for instance) hire consulting firms, or just small enterprises?

 

[phinds]

He's just saying that if you have a lot of experience already and have a client who really likes your work, you might approach them as being a client at the new firm you are starting. That could cause legal problems, though, so be careful going down that route. Especially if the contract you have signed with your current employer has a "Non-Competition Clause".

I think you have misinterpreted Vanadium's point, which I interpret as meaning very realistically "it's going to REALLY hard to get ANYONE to become your first client since you will have zero track record as a consultant (and if you have zero track record period then it will be impossible)".

 

[berkeman]

I think you have misinterpreted Vanadium's point, which I interpret as meaning very realistically "it's going to REALLY hard to get ANYONE to become your first client since you will have zero track record as a consultant (and if you have zero track record period then it will be impossible)".

Ah, I could very well have misinterpreted it. Especially the legal part -- I'm sure V50 is well aware of that. Thanks phinds for the clarification.

 

[Vanadium 50]

You need to get that first client from somewhere. That can be where you are working now (hypothetically), and yes there are legal pitfalls. Some companies like this transition more than others.

 

[gmax137]

But how is the market for CFD consulting firms?

I don't know. Where I work, we have engineers who specialize in doing the CFD work. Maybe you should do some searching online, as if you wanted to hire a CFD consultant, and see what you find. That would at least give you some ideas about how widespread such consultants are, and what kind of projects they work on.

As a side note, I have the impression that worthwhile CFD analyses are quite complex and time consuming (= expensive) so it is often an analysis of last resort. By "worthwhile" I mean analysis that really provides solid results and insights into the problem being investigated, as compared to the splashy "colorful fluid dynamics" cartoons produced by neophytes. So maybe there is a market for very talented consultants who can actually solve problems for their customers in a timely manner.

 

[BvU]

I imagine that, as most consulting firms in engineering, people would look after CFD specialists to solve problems that a)they don't know how to solve b)they don't have the time to solve 3)they don't have the resources to solve.

You're missing something here: they first of all have to think CFD can possibly solve their problem. Only under that condition and only once they have that idea, they will look for in-house 'specialists', interns, hobbyists, dilletants or (worst) try to solve it themselves. And only if that fails miserably and solving the problem urgently is worth a great deal to them will your points come into play. A lot of ifs. Even then they will search for portfolios with examples of (almost) exactly their particular problem having been solved elsewhere.

My company has in-house top-specialists whose talents are mostly wasted on fairly dumb problems that often (with hindsight) can be solved by other means such as good engineering thinking. But it's the lively, colorful, animated, presentations that constitute the practical advantage of the CFD approach in convincing decisionmakers to go that route and allocate resources.

As an aside: a business plan for a CFD firm would need enough room for real artists to distinguish their lively, colorful, animated, presentations from the lively, colorful, animated, presentations of others. Plus at least half of the turnover has to be set aside for schooling: things develop rapidly in this area.

If all this sounds a bit bleak: I once considered a career in scientific visualization (being besotted with cartoons as a kid and later at least as besotted with computers and the beautiful pictures you could envision making with them). Decided against it and never regretted pursuing a more run-of the mill scientific career.​

 

[CalcNerd]

While you might be great at CFD, I suspect you will need a partner with some grey hair and real experience. And HE won't want to partner with anyone of less stature unless they are brilliant!
.
So you should start looking for firms that you desire to work for that will hire someone like yourself. Look for several and don't be afraid to intern or work hundreds of miles from your current school to get the most valued experience. You will need several years of experience and hopefully work with a PE or two (that way, you can add that credential to your resume as well). Then you will be ready to consider consulting although a PE probably isn't a necessity for CFD consulting. However, a PE would help or may be necessary for you to get other work as well.
.
Or perhaps you can team up with a business Major to help you with the consulting too. Either way, just coming out of college (unless the college is a BIG name ie MIT or such) and trying to start consulting with zero experience will likely fail.

 

[Dr. Courtney]

I've rubbed shoulders with a few CFD consulting guys in the fields of blast physics and ballistics. I've also been a peer reviewer on lots of CFD grant proposals and papers. In my experience, all the guys getting paid for CFD consulting work have PhDs. They enter the field through a variety of paths. Since most of the CFD work I know of has military applications, it is subject to the ebb and flow of dept of defense funding. Even work tends to be paid for by private companies, when dept of defense funding is greatly reduced (as it has been from 2012 to now), those companies tend to scale back costs and stop hiring out much work. The CFD companies I know of that survived the past 5 years have a lot of other available tools and are fairly well diversified both in terms of industries served as well as research capabilities.

The transition to consulting is tricky in any technical specialty. As mentioned above, you will likely have a non-compete agreement if working for a firm that does CFD, so you likely need a non-CFD job for the length of the non-compete term. In a period of really high demand, it may be possible to transition from an academic research group (as a PhD student or post doc) to consulting work, since there are not usually non-compete agreements in these positions. But in most cases, it will take time to build consulting work up to the level required to earn a living from it. I found that having a faculty job without restrictions on consulting work provides a good base pay and benefits while building a reputation, business, and body of work in consulting.

Finding business in consulting is tricky. Most of the work I get paid to do by private companies is under some type of non-disclosure agreement where I cannot reveal either what we did or for who we did it. This means I cannot use completion of a successful project for one company to recruit business from others. We establish our credibility and ability to solve problems by publishing papers. By comparison with income from private interests, our income from public sources that allow results to be published is small. But most of those big private contracts come from people in industry reading our published papers and realizing we can solve their problem. We don't do CFD, but in many ways we compete with many firms that do CFD, because we provide experimental approaches to many of the same problems a firm with CFD capabilities might otherwise be hired to solve. In most cases, we can solve the same problems much more cost effectively.

 

[russ_watters]

I don't know. Where I work, we have engineers who specialize in doing the CFD work. Maybe you should do some searching online, as if you wanted to hire a CFD consultant, and see what you find. That would at least give you some ideas about how widespread such consultants are, and what kind of projects they work on.

As a side note, I have the impression that worthwhile CFD analyses are quite complex and time consuming (= expensive) so it is often an analysis of last resort. By "worthwhile" I mean analysis that really provides solid results and insights into the problem being investigated, as compared to the splashy "colorful fluid dynamics" cartoons produced by neophytes. So maybe there is a market for very talented consultants who can actually solve problems for their customers in a timely manner.

What type of firm?

My company (HVAC) has dabbled in CFD, but it is very expensive vs our needs, costing somewhere around $70,000 per seat per year Since we can't keep someone fully employed, we outsource.

So yeah, I think there is a market for it, but yeah, the cost of entry into the business and overhead for a small company is high.

And yeah, in order for clients to be willing to trust you enough to hire you, you need a resume of proven success. So starting a CFD engineering consulting firm straight out of college seems unlikely to work.

My dad started an energy consulting firm 25 years ago. The start-up cost was a PC running Excel (Er...Lotus 123), which he already had. But he needed clients, lots of them, quickly to ramp up to a profitable business. It took him months to sign his first client and a year of almost exclusively sales before he got his first commission check for his work. To win clients, he could show his 20 year track record as an enginer/businessman (I'm sure his Standford masters and Harvard MBA didn't do much harm...). Point being, you need a track record before you are likely to be successful signing clients.

 

[ramzerimar]

My company (HVAC) has dabbled in CFD, but it is very expensive vs our needs, costing somewhere around $70,000 per seat per year Since we can't keep someone fully employed, we outsource.

Does the high cost of CFD comes from licensing software?

Finding business in consulting is tricky. Most of the work I get paid to do by private companies is under some type of non-disclosure agreement where I cannot reveal either what we did or for who we did it

I've already experienced that, in fact. I work with a small non profit company founded by students of the university and we do engineering consulting for small projects (for inventors, small business, etc...) and we have to sign terms of secrecy for almost every project we make, so it's really difficult to build a meaningful portfolio.

Looks like CFD consulting is not a great idea, after all. Would a more general engineering consulting firm be more feasible? Maybe working with other types of analysis, and not dealing exclusively with CFD.

 

[russ_watters]

Does the high cost of CFD comes from licensing software?

Yes.

Looks like CFD consulting is not a great idea, after all. Would a more general engineering consulting firm be more feasible? Maybe working with other types of analysis, and not dealing exclusively with CFD.

The other problem remains; people hire consultants for their expertise/experience: you won't have any.

 

[ramzerimar]

About licensing software, wouldn't using open-source software (like OpenFOAM) instead of ANSYS Fluent be an option? I mean, Fluent is more robust and so, but I think open-source software is taking a share of the market.

The other problem remains; people hire consultants for their expertise/experience: you won't have any.

I'm not talking about opening a consulting firm straight out of college. I'm talking about someone who has spent some decent years in the industry and has enough experience and knowledege about the market needs.

 

[russ_watters]

About licensing software, wouldn't using open-source software (like OpenFOAM) instead of ANSYS Fluent be an option? I mean, Fluent is more robust and so, but I think open-source software is taking a share of the market.

If it really is good, maybe, but if it is good and free people wouldn't need to outsource it.

I'm not talking about opening a consulting firm straight out of college. I'm talking about someone who has spent some decent years in the industry and has enough experience and knowledege about the market needs.

Then sure. Lots of people do it.

 

[DrClaude]

If it really is good, maybe, but if it is good and free people wouldn't need to outsource it.

You forgot to factor in ease of use. OpenFOAM is "research grade" software, but it doesn't have a GUI, and has quite a steep learning curve.

 

[elkement]

Does the high cost of CFD comes from licensing software?
Would a more general engineering consulting firm be more feasible? Maybe working with other types of analysis, and not dealing exclusively with CFD.

My experience as a consulting engineer has been that selling yourself as a specialist in a narrow niche works best. Clients hire a freelance expert to solve a very specific technical problem. The next time they hire you again because you have proved to them that you acutually cover much more 'general' aspects than just your niche.

 

[gmax137]

What type of firm?

Nuclear power plant vendor. Design, licensing, and operational support.

Does the high cost of CFD comes from licensing software?

The manpower needed to develop a good model for a given problem is not insignificant. Especially when the CFD route is taken because the problem is too hard to solve with approximations or bounding cases. Plus, these big models take days or weeks to run. Not that the CPU time is costly anymore, but if there's something wonky in the model or the inputs, it can churn away for a week before you find out it needs to be fixed up and re-run. That slows down the analysis process, which raises your manhours. I think it's worth listening to Dr.Courtney who says above he can run a test program cheaper than doing the CFD. That's really something to think about.

 

[ramzerimar]

I think it's worth listening to Dr.Courtney who says above he can run a test program cheaper than doing the CFD. That's really something to think about.

So what's the use of CFD? I mean, there must be a reason why CFD is used and researched, and I know it's used a lot in the aerospace and automotive industries. When does one uses CFD in a design?

 

[ramzerimar]

Or CFD is primarily used in fields where fluid flow is a critical aspect of the project (like the already mentioned aerospace field). And I imagine those industries would already have their own CFD specialists. In aerospace, wind tunnels are used a lot to collect data and validate CFD models, but when is one used over the other?

 

[mpresic]

Consulting is hardly an avenue for engineers who do not want to work as employees. Your firm will have a better probability of success if you, or your employees and or co-workers are certified as professional engineers, (PE.), The PE exam is usually taken by engineers who have been working in the field for perhaps 5 years. Most likely, to get this idea to work, you are going to need to get used to working as an employee for some time.

I occasionally run into consulting firm leaders. The "chief" of the firm has mostly gray hair. You can safely say (s)he is not a young person. When you say in your initial post you are thinking of this in the distant future, you might have to take this time frame in mind. You may have to be twice your current age to be the chief.

On the other hand, it is good you are looking to be a trailblazer. Many others are interested in following well worn paths, usually in academia, which they may not be suited. CFD is a pretty broad and active research area, and it takes many years to establish yourself. Perhaps if you became an expert in a less traveled field like tribology, rheology, or powder mettalurgy, you can be an "expert" with less time investment, and coursework. Because there are fewer e.g. tribologists, the customers will need to come to you. I think CFD experts are much more common.

 

[ramzerimar]

On the other hand, it is good you are looking to be a trailblazer. Many others are interested in following well worn paths, usually in academia, which they may not be suited. CFD is a pretty broad and active research area, and it takes many years to establish yourself. Perhaps if you became an expert in a less traveled field like tribology, rheology, or powder mettalurgy, you can be an "expert" with less time investment, and coursework. Because there are fewer e.g. tribologists, the customers will need to come to you. I think CFD experts are much more common.

I'm thinking about working in the aerospace industry as an aerodynamicist and specializing in CFD after college. I don't know which skills would be transferable from aerospace to other applications that I could explore in a consulting firm. I've thought about wind energy, perhaps.

Does FEA have a larger market than CFD? Maybe for piping systems, modal analysis, structure optimization, etc?

 

[Dr Transport]

Nuclear power plant vendor. Design, licensing, and operational support.The manpower needed to develop a good model for a given problem is not insignificant. Especially when the CFD route is taken because the problem is too hard to solve with approximations or bounding cases. Plus, these big models take days or weeks to run. Not that the CPU time is costly anymore, but if there's something wonky in the model or the inputs, it can churn away for a week before you find out it needs to be fixed up and re-run. That slows down the analysis process, which raises your manhours. I think it's worth listening to Dr.Courtney who says above he can run a test program cheaper than doing the CFD. That's really something to think about.

I can only imagine how long a nuclear power plant model would take to run, for 15 years I worked with the CFD group at my former employer, a friend of mine ran a design case for a test fixture, took 3 months to run to completion satisfactorily and that was for one configuration, he tied up two of their number crunching computers for almost two years getting the complete set of test cases completed. A good CFD'er will look at how the solution is converging and make adjustments, that comes with many, many years of experience. He would routinely take on a younger engineer, but until they had a masters in CFD and about 10 years experience, he wouldn't let them go off on their own and work independently.

 

[gmax137]

I can only imagine how long a nuclear power plant model would take to run

Oh, no, they don't model the entire system! Just close up details. Like one fuel rod passing through a support grid, that kind of thing. Most of it is regular pipe network flow (like Crane 410 methods). Plus a bunch of flow testing done way back in the 1960s and 1970s.

 

[meandcfd]

I am curious how the CFD consulting startup worked out