Navier’s James Linfield, one half of the dynamic duo behind Navier Flow Consultants was recently interviewed by Robin Knowles of CFD Engine for his Podcast Talking CFD.
James gave the benefit of his recent experience & shares his thoughts on:
- The origins of Navier
- Choosing a niche to focus on
- Advice for those looking to start a CFD consultancy in today’s market
- Navier’s approach to the cloud & why they have invested heavily in in-house compute power
- The importance of networking, building trust and credentialising yourself to potential clients.
Listen to the podcast here:
Or if you would rather read the transcript, scroll down.
Transcript:
– Robin
Today I’m joined by fellow Brit James Linfield co-founder of UK CFD consultancy Navier Flow Consultants. Welcome to the show James
– James
Hi Robin, thanks for having me. It’s a pleasure to be here.
– Robin
No problem. So as to Brits this show should be 40 minutes of complaining about the weather or politics as they are both completely bannas at the moment. I think we’ll stick to the more sedate talking about CFD, what do you reckon?
– James
Definitely.
– Robin
So as I understand it starting your own business is a relatively recent development compared to how long you’ve been doing CFD. So I’m quite interested to dig into what it’s like to make that jump from staffer to boss because I think there’s probably one or two listeners out there who might be in that position.
– James
Yeah
– Robin
Actually one or two listeners might be the total number but let’s make sure they don’t start a consultancy and compete with us.
– James
Definitely not.
– Robin
So can you tell me a bit more about Navier the company, not the old dude who was good at maths, and then we’ll dig into into how it all came about and that sort of thing.
– James
Yep yep sure. So Navier is a UK based engineering analysis and simulation consultancy specialising in the field of computational fluid dynamics or CFD.
– James
We are about two years old now. The company was formed in 2017. By myself and my colleague and business partner Mark Roberts. Basically both Mark and I had spent probably about ten years working in larger consultancies within different industries and it had always been a sort of a dream of ask to one day start are own business and about two years ago we finally sort of took the plunge. We saw the opportunity to effectively provide high end analysis capability to small to medium sized enterprises, which we thought was a good foundation to start a business on.
– Robin
Yeah dead easy.
– James
Exactly. Really simple just just do that.
– James
But yeah I mean yeah basically the kind of the goal really of Navier was to provide the sort of level of analysis capability that you might find in a larger OEM to SME’s basically that have a sort of an intermittent requirement for that level of analysis but it doesn’t warrant a full time analysis department.
– Robin
So were you and Mark both from a CFD background it wasn’t a structural, it was both doing CFD?
– James
Yes, it was both doing CFD. My background was working within a larger sort of R&D consultancy chiefly working in the aerospace sector. Mark’s background was oil and gas and renewables but we were both applying our CFD skills to a range of problems.
– James
But it’s interesting speaking, you know I’m sure you’ve found it Robin, but speaking to people in in different industries because although you’re applying it to a different problem at some core level the flow physics are the same. So whether you’re tackling a multi phase problem or whatever the application is slightly different but the flow physics are the same and some of the niggles are exactly the same.
– James
So actually having that kind of core competency in different types of flow physics. So I spent a lot of time working on projects involving low speed, high speed aerodynamics, multi phase flows, reacting flows with chemical kinetics, conjugate heat transfer, rotating flows you know all the usual suspects!
– Robin
All the easy stuff.
– James
Yeah exactly! Really simple stuff. Yeah definitely I think I’ve probably got a patch of thinning hair purely associated to a few of those.
– Robin
I think I’ve only got a patch of hair never mind a patch missing.
– Robin
So just to just to get this straight. You mentioned an R&D consultancy, did it do other things rather than just CFD, were you kind of like a business unit.
– James
Yeah. Absolutely correct. So it wasn’t a pure CFD consultancy. The company got involved in in general R&D activities spanning from design and manufacture of test rigs and test stands, experimental fluid dynamics, optical flow measurement techniques, software, data acquisition. You know, turnkey solution I think is the buzz word, but I was it was interesting actually because I got to to support sort of internal design activities from a CFD perspective so derisking designs and making suggestions and so on, but it was a combination of that and actually then also providing CFD services to external clients on a on a consultancy basis.
– James
So it was a bit of a mix, but yeah it was kind of an arm of the business if you like.
– Robin
So you had kind of a client facing aspect it wasn’t just that you were providing a resource to an internal team.
– James
Yeah exactly, so I did the client facing side of it which is good. You know it’s a very good experience and actually that’s in all honesty that’s probably the aspect I really enjoyed. You know it’s fun tackling people’s problems and trying to help people.
– Robin
So was it one of those scenarios where after ten years of doing it for somebody else you thought hang on a minute, I’m going to do this for me or was it more complicated than that.
– James
Yeah, so I think it’s a little bit perhaps a little bit more nuanced. So a lot of the time I was working with some some big OEM’s.
– James
You know some some large clients some small clients and so on and obviously seeing the benefits that pragmatic, well applied CFD can give and actually also I think the sort of supporting some of the internal design resource you get real good sense of, sometimes what you want to do and present as a CFD engineer is not necessarily what a design engineer is looking for or is interested in. So it’s always good to get that reality check of, right, this is the salient information that they care about.
– James
They don’t care that you’ve used a precondition byconguate matrix solver. They don’t care. So yeah I think I think that was an important lesson to learn. But, yeah sorry, coming back to the to the to the question.
– James
We really felt that this sort of capability could be of real benefit to small to medium sized enterprises. So we wanted to form Navier to really focus on delivering that benefit to those clients.
– Robin
I mean you mentioned problem solving and that being one of the fun bits I think that is kind of a a lot of what we do day to day but I just wondered if there’s a chance to kind of draw a distinction between the kind of problem solving that goes on a sort of design stage and then a kind of problem that goes on at panic. This isn’t working stage.
– Robin
Do you find yourselves drawn to one or other end of that spectrum.
– James
Yeah that’s a really good question. Obviously I would say the ideal situation is to do the analysis upfront and de-risk the product, but obviously things happen and you know we do get involved with actually trying to fault find and figure out why an existing product or system isn’t working.
– James
So that can be that can be quite fun actually, abit stressful, but actually quite fun.
– James
We did have one, so yeah I’ll share this story with you, I remember one meeting with a potential client that was actually in the pharmaceutical sector. They effectively had a product that they had designed, manufactured and had sort of got to the commissioning and sign off stage of it. It heavily involved fluid mechanics, but they hadn’t obviously engaged anyone with fluid mechanics expertise during the design phase.
– Robin
I can see where this is going.
– James
Yes exactly. So you know I remember sitting down with a few of the design engineers and you know one of the questions I always ask is it has anyone done any CFD on this before. And they all looked a little bit sheepish and looked at each other and sort of one of them admitted that yes he had basically had a bit of a go with a CAD embedded CFD tool, but you know confessed that he basically had no idea what he was doing and the results didn’t look believable. So they sort of just discounted it and moved on. So it’s like yeah okay. You know fair enough it’s good that at least you sort of recognize that that’s probably not a fair representation of what CFD can do for you.
– James
But in this instance I literally sort of took a look at their at design and basically said look you know I I could cost up doing some CFD for you and this and that and these are the benefits, but I can literally look at your design and tell you why it’s not working. Just from a fundamental fluid mechnaics perspective.
– Robin
Use your psychic powers. Did you close your eyes and wave your hands over it a little bit.
Yep – mystic ball.
No, and this is not even me saying oh great, you know look we’re wonderful. It was literally, there was such a fundamental fluid mechanics issues that it was not even worth doing the CFD it was a case of look, straight away here are some quick wins. Don’t do this. Definitely do this.
– James
You know afew really obvious things, well I say obvious, they’re obvious to someone from a fluids background. And they were like Okay yeah no that’s great. That’s really good. Well we’ll try that. And basically those modifications fixed their problem and we never got any work from them. But, I like to think that the moral of the story is actually, hopefully, by kind of being open and transparent and saying look this is how we think you can improve things.
– James
I’d like to think that in that situation if they do have a problem in the future or they do have a concept they’re not sure of. I’d like to think that they would come back to us because they know that effectively we’re going to be honest with them and if there are any kind of easy spots or low hanging fruit if you like that we can just say, yep this is the way to go with it. We try to always tackle problems from a kind of fundamental fluid mechanics approach. We don’t always jump into let’s do lots of CFD just because we can. Which I think is, well it is certainly I think something that you know some CFD practitioners have a bad reputation for, it’s kind of tunnel tunnel vision.
– Robin
Yeah, Maslow’s hammer. You’ve got your CFD hammer and you want to hit everything with it.
– James
Yeah it’s exactly that.
– Robin
So I mean you had some exposure to the not so interesting bits of running your own consultancy. The writing proposals and doing the accounting and things like that or you know managing a P&L or whatever in your previous life.
– James
Yes.
– Robin
Has anything sort of surprised you having gone out on your own, either on the upside or the downside.
– James
I think obviously having, yeah as you say, having worked in a consultancy you kind of get that experience, the client facing side of it, the proposals the tendering, obviously the technical work as well. I felt very kind of confident if you like in that area, because I’d been doing it for so long. But I think the areas of starting your own business that are difficult and I think sometimes they’re the areas that people don’t really publicise or talk about. I think kind of in an era of tech startups and Silicon Valley and all of this, it’s all kind of I think definitely glorified and you tend to only really hear about the upsides of starting a business. It always appears to just run smoothly and everything works perfectly. Whereas I think the reality is that it’s you know a bit of a roller coaster. The highs are really high and I’d say the lows are quite low at times. I think something that perhaps underestimated a little bit is is how much effort and work goes into you know the sales kind of side of things the business development.
– Robin
Networking?
– James
Yeah absolutely. And I think even with an existing network to kind of call on and you know try to build on it’s still really really difficult.
– James
It takes a lot of work. And obviously. I think as engineers we tend to sometimes want to get stuck into the technical stuff and that’s the fun part. That’s the bit you know you can be doing at midnight or whatever and you don’t really mind because it’s what you enjoy. But I think as engineers the networking side and sale side of it probably doesn’t come naturally. I would say it’s something that we’ve built on and learned and tried to get better at. But yeah it’s certainly not easy.
– Robin
Do you think you and Mark could have done this off the back of a degree or PhD or something like that, who knows you could have.
– Robin
Yeah I think it’s an interesting question and I’ll try not to sound too negative about it but I think it’s so so important to to have some relevant industrial experience to A: kind of credentials yourself to potential clients, and B: to actually have a network of people who have worked with you, who know that you’re good. Do you know what I mean? I think that’s invaluable.
– Robin
Yeah yeah
– James
And you just you just simply don’t have that coming straight out of university.
– James
So, I guess my humble advice would be you know get out there do do some consultancy roles, A: see if you even enjoy doing consultancy because that’s the other thing you know it’s…
– Robin
Who wouldn’t enjoy the the roller coaster of work coming in and disappearing and all that sort of thing. Yeah the advice of perhaps being able to go and learn on someone else’s dime is is a good one.
– James
Yes and probably less stressful when you’ve got a guaranteed paycheck every month.
– Robin
Yeah absolutely.
– Robin
So as somebody who’s just started a simulation business in the last couple of years you are all in on the cloud….?
– James
Yeah I knew, I knew that you were going to go this way Robin. So this is controversial. This is very controversial.
– Robin
Why because you are all in the cloud.
– James
Absolutely. Erm no.
– Robin
Basically you’ve turned it off.
– James
Yeah we’ve switched off at the plug.
– Robin
What?!
– James
I’m going to say something controversial. I think that sometimes the cloud can hinder innovation………..
– Robin
Sorry James I can’t hear you. Must be a problem.
– James
Yeah, this going to be all your listeners banging their fists on the table and shouting.
– Robin
No, I think you’re in the majority there. I think that most people aren’t anywhere near being all in with the cloud. But it would be good to unpack it a little bit. So you guys have kind of invested fairly heavily in house compute, have you?
– James
We have correct. Yes it was a bit of an executive decision early on to invest heavily in internal compute resource.
– Robin
Was that your comfy place. Is that something you’ve done before or was that new?
– James
So that was new.
– James
So I have used AWS and stuff quite a lot. So kind of the Linux side you know, obviously also being an OpenFOAM user and whatever the Linux side of stuff didn’t really worry me too much, but actually building a cluster from scratch is much more complicated than you may think. If you’ve never done it before.
– Robin
But I guess it’s not something you regret then.
– James
No no definitely not. It was a steep learning curve. There were lots of late nights and lots of coffee drunk. But I think that it’s something that is now a resource we can sort of really reap the benefits of.
– James
And looping back round to kind of the motivation behind it. We wanted that internal resource so that we could develop internal sort of processes and methods to tackle really complex flow physics the type of stuff that would be prohibitive to tackle with your 18 or 36 core workstation or whatever. And also the type of flow physics where you don’t want to spool up a few hundred cores in the cloud, set your simulation going and come back the next day to find it’s just collapsed in a big heap and you’ve got a nice invoice waiting for you.
– Robin
That never happens!
– James
Never, absolutely never happens.
– James
So yes that was certainly part of the motivation. I should say that’s one of our early projects was doing quite a large aero optimisation study and we were running on 80 to a 100 million elements cell meshes. That’s when you start thinking, oh we might want to do 50, 60, 70, different types of runs
– Robin
Cha-Ching
– James
Yeah. Yeah exactly. We ran the numbers and we were like hang on a sec, we’re going to be handing over a lot of money to a cloud provider for this. So we thought actually let’s bite the bullet.
– James
Let’s you know invest internally and have that as a resource. So we now have have our own 320 core compute cluster that we’ve built with blood sweat and tears
– Robin
but when it’s time to make that 500 cores….
– Robin
Is that something that you would do or would you then go to the cloud.
– James
Yeah. That’s really a question. So we deliberately scoped the architecture of the cluster such that we can actually, it is very scalable scalable in let’s say manageable financial chunks.
– Robin
OK
– James
Obviously, the plan is not to have 3000 cores. You know that’s not where we are going with it.
– Robin
Yeah.
– James
I think ideally the circa 500 cores one day would be would be a good spot to be in.
– James
But I think in the interim we would certainly be open to using cloud resource just to help manage some of the peaks. If we did find that our internal resource was saturated.
– Robin
Yeah.
– Robin
Because I mean although although I am a huge proponent of cloud, it’s mainly because it makes sense for me, my compute requirement is spikey. Sometimes it’s loads, sometimes it’s nothing. I don’t have a kind of base load that could take advantage of having a reasonable sized compute resource to myself. It sounds like you do have a bit of that base load where you’re kind of developing methods and you’ve got something that you could leave running for a chunk of time and see a benefit.
– Robin
Are we’re talking about developing solvers or are we talking developing around tools or methods and things.
– James
Yeah exactly. It’s the latter.
– James
So we’re heavy OpenFOAM users, we don’t do any kind of source code or software development, it’s usually combining the existing capabilities if you like and developing the workflow so that it’s robust so that we can actually deliver consultancy jobs in a timely manner. Alot of the time it’s about building bash scripts and stuff that sit around it
– Robin
Yeah the automation.
– James
Exactly. So automating the dictionary generation and so on and making the workflow a bit more kind of optimised, having python scripts to drive paraview in the background, that kind of stuff, the stuff that I think you need to make open source codes more viable in a consultancy environment.
– James
And that takes an investment, a time investment if you like.
– James
It’s been developed over time. It’s building all the time. And the other thing that we try to do is use it as a resource and constanly improve by you know running validation and verification studies on different types of flow physics. I mean the amount of common research models and stuff that are out there from research institutes is crazy and there’s a lot of really good rich experimental data.
– James
So we’re always you know we’re always trying to benchmark our methods and our processes basically.
– Robin
So you mentioned OpenFOAM a few times, and obviously opensource for this kind of business model makes a lot of sense. But I mean when I speak to consultants they tend to be, well CFD shops in general, they tend to be either in one of two camps they’re either code agnostic they’ll use whatever they can find that will get the job done in the best way if that’s if that’s a Star CCM+ seat. And that’s got the tools that they need for that job or whether it’s OpenFOAM or whatever.
– Robin
But the other camp is we don’t we only use open source tools. Are you on the spectrum, do you fall either end?
– James
I would say that we don’t fall at either end.
– James
I mean I am a big OpenFOAM fan and we do use it very very heavily.
– James
But I would say that Navier is code agnostic really. Both Mark myself have obviously had a lot of experience in the past with with commercial codes. You know your Fluents and Star CCM’s etc.
– James
And we very much try to use the right tool for the job.
– James
Obviously there are benefits in being able to take advantage of the open source codes, especially with a 320 core compute cluster. You don’t want to be shelling out for parallel licences.
– Robin
Now that’s a few HPC packs isn’t it.
– James
Exactly. So there’s a little bit of a bias to want to use OpenFOAM but we accept that there are some areas where it’s not strong enough and we will just go with with the tried and tested commercial codes.
– Robin
Yeah. Makes perfect sense. I suppose that varies between industries perhaps or applications.
– Robin
Have you got a sector that you home in on or are you fairly broadly focused.
– James
Yes I’d say we’re fairly broadly focused.
– Robin
is that right, that’s probably not a very good, can you be broadly focused?
– James
Yeah that’s very true.
– Robin
I’ll take that out in the edit don’t worry.
– James
So I would say that we’re focused on, well how I like to think of it more as the types of problems, so the aerodynamics, the multi phase flows, you know the reacting flows. What we apply those to you know bridges aerospace, power generation, built environment type stuff, water treatment industry. Recently we’ve been doing a bit of work in the marine sector as well. That’s been quite interesting because that’s actually effectively applying a lot of the processes and techniques that we developed for civil sector.
– James
So doing wind comfort studies for pedestrians around buildings and so on we’re now applying that to cruise liners for open decks wind comfort. Ensuring that these weird and wonderful super structures that they build don’t induce a gust of wind that is gonna blow someone overboard.
– Robin
Which one of you gets to go for two weeks and make sure that it’s working?
– James
I have dropped that hint with our client a few times, it would be very nice to go.
– Robin
For some validation.
– James
Exactly. Me with my pitot tube walking around.
– Robin
And a cocktail.
– Robin
So I was speaking to another consultant as part of this series. He was suggested that they are seeing a trend towards customers being able to do a lot more themselves and what’s left behind for consultants is the trickier stuff, the low hanging fruit has gone and we’re left with the tricky physics the the harder to deliver projects. Is that a trend that you see or is CFD within small business or the smaller engineering companies perhaps not at that level yet.
– James
Yes so that’s interesting. I think with the continuous development of you know, sort of, well I don’t want to be too derogatory, black box CFD tools shall we say.
– Robin
It’s not a pejorative
– James
You know they are getting easier and easier to use and they’re being marketed as you don’t have to have any kind of CFD experience just click a few buttons and away you go.
– James
I think those types of products probably are filling up some of the space where it’s more sort of run of the mill analysis. That’s not necessarily particularly complex. Now whether they get the most from them. Let’s not go there. That’s probably a completely different podcast. But I think you’re right. I think where people are probably more inclined to reach out to consultants is when they’ve got a really sort of unfamiliar complex problem that’s well out of their comfort zone basically. And actually that’s kind of where I think we had a lot of value because we’ve seen a lot of weird and wonderful problems in lots of industries. Many of them fairly kind of cutting edge technology companies. And so you see a lot of opportunity to transfer technologies and approaches to solve problems and come up with innovative solutions. And obviously for us having the cluster as a resource aswell means not a lot of stuff scares us really…… famous last words.
– Robin
That’s a nice place to be. Sort of brings you round to that kind of domain expertise that we were talking about earlier. So I mean on the one hand you’ve got the argument of having a niche focus of being just oil & gas for example. But then you’ve got on the other side you’ve got like you’ve just mentioned the fact that you’ve seen things in other industries in other domains which may be innovative and novel in a new industry something that might not be super trick to you but to them is a revelation and being able to sort of facilitate that crossover is a kind of a interesting place to be.
– James
Yeah definitely, I mean I think if you’re deeply entrenched in a certain industry and you’ve been looking at the same problems and you’re a bit too close to it, I think sometimes it can be easy to be a little bit blinkered with what’s happening in other sectors.
– James
And there’s definitely, definitely opportunity for technology transfer and cross pollination of ideas.
– Robin
Yeah absolutely. I’ve been pretty surprised at the willingness of kind of small business owners and consultants to sort of share their experiences around this. I find it particularly interested, if somebody wants to get in touch with you or Mark or you guys about anything that they’ve heard or just because you’re one of the good guys. What’s the best way for them to do that?
– James
Yeah so I would say LinkedIN, I’m always up for connecting with like minded people. Just connect with me, drop me a message, happy to have a chat and answer questions.
– James
Obviously you can also find our website fairly easily and just drop us an email. We’re always keen to connect and network with people.
– Robin
Yes, the network is king.
– Robin
Cool, I want to thank you for taking your time to come on the show James and giving us an insight into what you’ve been up to and how you do it. No guests, no podcast.
– James
Absolutely, it’s been a realy pleasure being a guest, keep up the good work I’ve really enjoyed listening to the previous episodes