171 - Exploring Water Mist Systems with Max Lakkonen

In our everyday job , modeling is one of the main tools Computer modeling or numerical modeling , cfd modeling these are the tools that we are mainly using in our performance-based design , and there are still some topics in the world of modeling fires that I believe are pretty rough , or I just approach with caution and one of such topics is modeling any water-based

fixed fire fighting systems , such as sprinklers , dilute systems or water mist . In this podcast episode , we will focus on the third one , the water mist systems , no matter how you distinct them high pressure , low pressure , water mist .

We're just going to talk about the water mist and we're going to talk about the challenges in how to model performance of such systems . But you know me , it's not just going to be how .

The question why will occur many times in the interview , because for me , and the most important thing , is why we need to model , what we get from the modeling , what is the objective that we're trying to achieve and , if we are clear on that , how do we reach that objective ? For this episode , I have invited Max LekaLnen .

He's the managing director of IFEB and literally a few weeks ago , or even a few days ago , max has been chosen as the president of the International Water Mist Association , so congratulations , max . It's a very good person to talk to . I've heard him speak about modeling water mist multiple times and I'm sure you will enjoy him talking as well .

One more thing IWMA the Water Mist Association has also released a position paper on modeling water mist with CFD , and that position paper is linked in the show notes and is a very good companion to this podcast episode , as it captures most of the stuff that we wanted to convey through the episode in a written form .

So , without further ado , let's spin the intro and jump into the episode . Without further ado , let's spin the intro and jump into the episode . Welcome to the Firesize Show . My name is Vojtěch Vyjgřínský and I will be your host . This podcast is brought to you in collaboration with OFR Consultants . Ofr is the UK's leading fire risk consultancy .

Its globally established team has developed a reputation for preeminent fire engineering expertise , with colleagues working across the world to help protect people , property and environment .

Established in the UK in 2016 as a startup business of two highly experienced fire engineering consultants , the business has grown phenomenally in just seven years , with offices across the country in seven locations , from Edinburgh to Bath and now employing more than a hundred professionals .

Colleagues are on a mission to continually explore the challenges that fire creates for clients and society , applying the best research experience and diligence for effective , tailored fire safety solutions . In 2024 , ofr will grow its team once more and is always keen to hear from industry professionals .

Who would like to collaborate on fire safety futures this year , get in touch at OFRConsultantscom . Hello everybody , I am here joined by Max Lakkonen from IFAB , managing Director of IFAB . Hey , max , good to see you in the podcast . How about you ? So I've invited .

Max Lakkonen , the Head of Scientific Council at the IWMA and who came into the show is the president of the IWMA . Congratulations on becoming the next president of the IWMA . Thank you very much . Any particular goals you have for this presidency ?

We have all the chrono , so definitely some changes are going to happen during the coming season I'm looking forward for that , and also the scientific council .

I think you had a presentation about modeling in iw may in warsaw , which I even covered in the fire science show . I know that you just hadn't talked about modeling water mist in antwerp in the iw may conference that was held literally in the last week .

So I thought , yeah , you're , you seem like a good guy to talk about modeling , and modeling is something my audience loves , so perhaps let's let's talk a little bit on modeling the water miss . How about that ?

But obviously there's many decades of work behind and also experimental tests , and I think that would be a good starting point , starting from not day one , but having the background and then ending up to the modeling which is nowadays too .

So how about we start with the definition , or where the water mist even came from ?

But obviously it has some historical reason that some other technology did not really appreciate too much smaller droplets so they expanded that to quite wide range .

Also other technologies , but in the earliest , at the very first standards , there was different categories , also due to droplets , but that was actually cancelled and now it's a flat rate one millimeter . But in practice the mis-IS technologies , they are far smaller droplets than that .

And traditionally high pressure systems they utilize the least water but they are having the highest pressure .

And you can imagine that , regardless what is the pressure level , if you're putting a mist to a data center or inside the electric cabinet , it's a very fine . But if you were would put mist . Then the road tunnels . So obviously there the droplets that they are much larger because they have to have much more momentum to it , the ventilation .

So that's that , let's say the introduction , but I would say that typically they are the range of between 50 and 250 , 300 microns .

You mentioned in the panel applications you would have to have momentum to penetrate . Now , that's also something I didn't always get in WaterMist , because I assumed that this is a distinguishing line between the Sprinkler technology and WaterMist technology . Sprinklers would have those large droppers that penetrate .

They go down to the fuel , whereas in water mist you have something that basically acts in the gas phase itself , like wetting . This surface action on the fuel is much more limited in water mist . Can such simplification be made ? Or really , depending on what the most technology you can achieve , like the best of both worlds , the reality is somewhere between .

In same way also mist systems I mean , as I said , they are having relatively large droplets in tunnel applications and they do also wet it in insane sense than deluge system would do . So germans would say that the answer is Jain . So it's somewhere between .

But you know that , like 15 years ago , we had a long discussions with some other known people that whether the droplets would fly out from the tunnel with mist systems . And obviously I think that the mist concept was not water mist but real mistcept was that the droplet sizes . They were considered way too fine . What are utilized in internals ?

Traditionally mist systems , they're working three-dimensional , so they will fill the protected volume and then obviously the influence in the gas phase is having the major role , but a portion of the water will also collect on the surfaces , so it's also influencing there .

And then obviously one of the key things is obviously this extended or let's say , the very superior cooling effect by having high number of small droplets , the traditional big surface area evaporation and the way absorbing the heat .

And then , the last but not the least , because there are also some applications where this is really tuned and for example , with flammable liquids , it's also the prevention of oxygen by the large evaporation , the local evaporation , so that will block the entrance of the fresh oxygen .

So , yes , the technology works in different ways and , depending on the applicator , different characteristic are more emphasized than the others .

I wonder how many of those evaporate actually , or again , it depends on the fire .

So measuring also the water which is collected on the surfaces and trying to evaluate that how much water was consumed , and obviously larger the fire is , more it will then evaporate . So it's kind of proportional to the fire size .

But I would say that if we are having 50% , if we are having 50% , then we are really at the high end and typically you are spraying much larger area than where the fire is .

You are not very seldom , you are just targeting and it's more than the research where you are utilizing and you keep the things in control that you are trying to measure the evaporation rates that you are trying to measure the evaporation rates .

In smoke control it's all you know mathematical equations that tell you how much smoke you produce . Or CFD modeling , where you just simulate your smoke control system and you see if it works In the water mist . It's all based on experimental insight .

Is there also any particular reason why the industry is so uh , I wouldn't like I want to use the word obsessed , but it's not the correct word perhaps why it's so based on on physical testing ? And does , in such a world of such strong emphasis on physical testing , does performance based design or cfdD even have place for ?

two questions . I mean , if we are looking , to the history of water mist .

And obviously then water mist started to conquer the marine business and nowadays , for example , cruise liners , they are having almost 100% market share and also the engine ships or the machinery rooms . I mean they are having very , very high portion of this system . So that's where it all started .

But since day one , watermistist was considered something new and it was careful . I mean , everybody had to be careful because higher pressures and new approach and then a lot of experimental tests were required . Therefore , basically every new application requires some tests and also because the technologies are different .

You know , in different uh makes they are based on performance base design . They are not , let's say , prescriptive than that sprinklers are , for example , that you do a certain application rate and everything is good .

But obviously two companies might go to a very same protection objective or the performance with the two different nozzle spacing , two different floor rates , two different nozzle types , and that's been the reason why there has been so much testing in the past . But nowadays the situation is good .

There are many test standards and there are many companies and they are now covering a really huge number of approvals to various different kinds of applications .

But I'm like really curious , like let me rephrase the question Is there any way you could like talk about generic water mist , you know , and generic flow rates or generic pressures and just don't go through this expensive , you know pathway of testing ?

But coming now to your technology design , so the objective was given and it was given by the standardization bodies , by the authorities , that , okay , this is the test you need to do , this is accepted criteria , no matter how you get there .

And after you have passed the test , you have to do that , the component test , and after that you have the approval to install the system , so kind of like pbd , but powered by experimental research for a very narrow spectrum of use , right it's a narrow in that sense that you , for different risk category , you have to do on fire test series following the standard

. so obviously if you have to do on-fire test series following the standard , so obviously if you have a building , so in the building you might have an office space or accommodation , you might have a car park , you might have some electric room and obviously cable tunnel .

Right , because , uh , it's so specific I I find it difficult to to give like a you know , hire a new person to my lab and tell them okay , from tomorrow you're designing Watermist . Here's like five books you have to read and you'll be good . That won't be enough . You need experience from lab , right ?

It's not that you would for every building that you would need to do a validation or fire test . I mean it's just that these companies they have done for different number of applications that they are good , regardless of what is the building .

But transferring the information then from the companies , from the industry to the fire engineer , that's the hurdle to make the link , and obviously IWMA is also playing there between trying to spread the knowledge .

Previously you've also mentioned there are different types of water mist that would result on with different water distributions . I know there's low pressure water mist . I know there's high pressure water mist . Would you like to comment on the types of water mist and perhaps how big of a difference does it really make that you have ?

different operating pressures , because that seems to distinguish them .

But if you want to combine small droplet sizes together with the momentum , so having the smaller droplets and shooting them into the protective volume with the high momentum , that requires energy , that requires higher pressure and some of the applications you really gain of having both small droplets and then filling very effectively the volume .

So that's probably one of the fundamental changes . And obviously then it correlates to the flow rate that typically , because of this effect , higher pressures they are having lower flow rates . But if we are looking to the standardized risks , so there is a test protocol placed by the standards and obviously you can pass this test with low pressure or high pressure .

It's then up to the manufacturer what they're going to do .

. Perhaps it's time to start moving to the challenges with CFD modeling . The second part of my question where do you see modeling water mists as a part of the design ? I mean , you obviously can't replace the protocol with CFD itself , but where would it be useful for you as an engineer , to put water mists into your modeling ?

It would apply also to any other technology . So that's kind of virtual pre-testing where it's been used . And then obviously then we are not there that we would be able to fully justify the designs of any kind just based on pure CFD . So that's still difficult .

But again , if we are having good experimental data , so then we can use in the kind of post-processing assessing that , for example that compared to the file test , if the geometry would have different dimensions , so probably we can mitigate this based on on the cfd works . Or we will have a different ventilation conditions .

So again we can use cfd to assess the influences . Or we want to studyD to assess the influences , or we want to study the effect to the structure , or you name it .

experiment into slightly different geometry , slightly different ventilation conditions , by simulating the original experiment and the extension and then working out like what's the difference ? Like is it still applicable ? Something like that ?

I won't break any NDAs if I tell you I've been challenged to those simulations a few times in my life and then a few times we had to do them because the client really needed them .

What I have in mind is simulating what it means in an office layout , to prove that it works , or to give a proof that it doesn't like destroy the operations of ventilation or some other , but basically as a proof that in the building setting it's going to operate how to even say it nicely To give a bunch of nice colorful pictures that it works correctly .

What's your take on those aspects of PVD ? Because , if I may , uh , one more sentence that that's what we do with smoke control . For smoke control we don't have protocols , we don't have extensive experimental proof for any different occupancy . We just do the system as it is on the building and based on that we make conclusions does it work or not ?

So , yes , we also , we are doing a lot of simulation related to smoke control in buildings , in tunnels , in metro stations , and it's standardized there . But if we are looking to the office building , so typically in the office building I mean office areas , it's typical or ordinary areas at one .

Normally there is not even question whether the MIS system or shrinker , with the regular spacing five millimeters per minute , whether they would be accepted or not . Typically it has to do then , like you said , with the certain ventilation condition that some people will come with a question .

They probably don't do a fixed question for sprinklers because they think it's a standard . Very few people actually remember where this five millimeters per minute actually came from and probably that time the furniture was different and the building materials were also different . But anyway , that's a side topic , but for these special cases .

So there's something with the ventilation or with the atrium or whatsoever , and it has to do a lot with the experience of the people using the CFDS .

And I've asked him are sprinkler experiments expected to be compared with water mist as reference ? And he made a very sad face . So it's kind of weird . You know that one technology is given completely different scrutiny as another , even though we are in the same fire safety market .

And the performance of that , and especially if you compare it to the smoke control , where you have literally no experimental proof and you base everything on on your engineering judgment fueled by numerical modeling .

You you've said something very important explaining me that that case is that you will have difficulties capturing the influence between the mist and the fire .

So what if you go another simplified way just reduce the fire size and just model the cooling effects to what the mist has on the smoke and air ? Would that give you anything , any knowledge you could work with ? Or it's just too generic to ? find it useful in engineering .

So we are taking the design fire curve based on the experimental tests , that we use this , and then we can model that how water miss is influenced , the outcome , the output of the fire , how it's cooling and how is , for example , shielding , what are the radiation level or heat transfer to the structure ?

There the question is that , if you don't have the experimental data , so then the question is that where do you cut the design curve ? So is that when the first sprinkler will activate , you then take the flat If you have the data from the free burn test , and that's where there is , let's say , normally the main discussion .

And if you don't have an access to a good enough experiment with heat release rate measurements , which are not that common actually , you would basically go alpha T squared growth applicable for a specific design situation and then , just based on activation , figure out a reasonable time to cut off the curve and flatten it . Do I ?

And then obviously there are also some tests where there is a data how it's influencing . But then obviously this has to do also a lot with the manufacturers . They process this data , so you have to be close to them .

I mean that would probably end in some like worst case scenario where you fail to influence the fire inside the vehicle , but she got the cooling effects which . Because , like , what question are we answering with this ?

Like if we cut down the case where we want to showcase a colorful pictures to the firefighter who has to say whether the building is approved or not , if we just cut this use of cfd , which obviously would be like one third or maybe half of the market and you want to study something , you want to understand , what question can you get ?

And the idea that comes to my mind is that , okay , I'm going to have different temperatures , my smoke is going to have different buoyancy , that most likely influences my flows and layers and everything right .

So this very same question remains regardless . What is then the technology you are utilizing ?

You know , trigger the jockey pump , then trigger the main pump and also before all four hats start up . So let's say I would have the initial trigger in 100 seconds . I would probably add like at least 60 seconds to that and then I would cut my alpha T-square curve in that case , or maybe alpha T-thirds curve if I was doing design for Singapore .

But yeah , I would probably do that . I'm not sure how valid or common that would be .

But we actually , I mean we started with the model , the main input data which is the influence to the heat release rate , and I mean over this whole CFD water mist combination . We actually published as an IWMA position paper last week where we as a society , we wanted to give a kind of best practice how to tackle these topics .

Then they would run the same test with water mist . It reaches whatever 30 megawatts , 50 megawatts , whatever the number . And then , uh , they give an impression that the water mist is capable of fighting 200 megawatt fire , and for me it's like such a trigger . I mean , what water mist does is preventing the 200 megawatt fire from ever happening .

But when it is 200 megawatts , there's not much you can do . And why I'm annoyed ? There's not much you can do . And why I'm annoyed ? Because we've been forced to run simulations for cases where someone wanted 200 megawatt fire and the water mist at the same time . And prove me it works .

And I'm like it's just like you're asking me for a Star Trek and Star Wars universes merged together in one space . It just will not work . It's not how it works . It triggers me a lot .

But there is one occasion where it's been done and very , very scary Something like I don't even remember it must be 15 years ago , but it had nothing to do with the road tunnels though , so that's it ?

And , having said that , our intention was not to go to 200 megawatt , the trigger imposed was 150 megawatts , but the safety valve of the pump unit popped and we had to restart the pump and then very , very nervous people started to be around and then we started to lose . Camera here , camera there . Nobody was allowed inside the tunnel .

Yeah , I can imagine that and you could see that how it was in TSD , how the tunnel was leaking also between the joints , but anyways , everything went fine . But obviously you don't want to .

If you are planning a road tunnel , I mean in the channel tunnel , the train can travel quite a long time having the trucks burning on board , so that's why there's a lot of time . But in the road terminal you obviously want to activate as quick as possible . If you go to that level of the fires , I mean most of the technical equipment is already destroyed .

So I mean it doesn't make sense .

I I don't think people appreciate logic enough in in the fire community and uh , for me that that's , that's a very logical conclusions that if you don't have your pipes and nozzles there , or they bend it or they fell down from the ceiling which is most likely when you have a fire of 200 megawatts that there's not much they can do to reduce the size , perhaps

reduce the . Of course they would reduce the temperatures a little bit , but yeah , it's going to be challenging . You must have used a lot of water in that channel fire test to get that piece under control .

I know guys at FM are really good at doing stuff like that , like modeling before they burn , and perhaps I'll get someone from FM to talk on that one day . How about the gas phase ? So the phenomena that are happening in gas phase , for me two are important . One is three are important .

One is evaporation , that is , how much heat is taken out in the gas phase . The second is the change in oxygen concentration with the evaporation , which you've mentioned , important for the pool fires .

And third one , uh , for me an important , the uh , water induced airflow , like like the momentum pushed onto the air and how it changes , because for me it's important to consider the smoke layer disruption , for example . So how confident are you with our capability of modeling those phenomena with our CFD packages ?

And perhaps there's also tricks of the trade that people should be aware of .

So I I think these are . If we would then combine these , these three things also with the droplets and the interaction of droplets together , droplets together , these would be , I mean , big stickly .

Then further , but if we are looking from the , the technology as it is I'm now referring cfd technology as it is so what I've seen is that I think we can predict with a fairly good confidence the , the influences when it comes to the cooling . Normally we are not so much of interest about the , the air or , let's say , the oxygen , okay , itself .

I think that that's more the academic research , but we are more interested that what will happen for the fire . So , and obviously that's more the , the question and the , the interaction between the , for example , with the stratification , which truly is of your interest .

So I think that's also quite well known and there's a lot of data , also from experimental data , from the test .

One more that we didn't touch modeling the water mist release . That I find really challenging , because if you want to model water mist it doesn't magically appear in your model you have to spray it from something right .

And to model a real nozzle and model the physics that happens at the nozzle itself , that's like nozzle itself , that's a PhD , like that's a simulation on its own right , but I don't want to do a PhD every time I sell my CFD to someone . I need a reliable model . So there are simplifications In FDS .

There is this simple model you put some angles , you put the pressures , you put the flows and you get some outcome . How you feel about that ? Where are we heading with this ? Is it going to be more simple , like I'll get the catalog of nozzles from manufacturer one , two , three , I'll just put them in my model and they will work ?

Or there is a lot of stuff that I should pay attention to when I try to model that no , I don't think so that it would be so complicated .

But typically the mist nozzles , they are having the optimum performance , they are having the multiple orifices , so multiple kind of jets per head , and this is like you said . You have the model , the geometry , you need to know the droplet distribution and then you need to know the velocity and that's pretty much all what you need to do .

So we've been quite at least satisfied with the way we have done and we have done also comparison just with the application rate , experiential versus CFD , and with the results .

Then we have to do much more analytical work , we have to look to some application which being close , or it might be also that we have to think about the interaction between water and the fire itself and how it's going to influence . We don't want to go with the overly conservative approach .

So there are give a kind of guidance that what is best practice , the kind of boundaries , what is what is allowed , what is not . I think that's the the important message that I would say max , thank you .

And then obviously it's then to talk with the organizations who have done it in the past , to talk also with the manufacturers if experimental data is needed , and I think that would be the way forward with this topic .

Of course there are standards , there's more widespread knowledge , but for the modeling tasks you seem to need some very specific details that you probably will get from the manufacturers . But it may take some effort . Max said it's quite straightforward . However , in between the interview and today I had a phone call with a colleague . He's a designer .

It was completely unrelated to the podcast episode but we were discussing Watomis . He was asking questions about my opinions on Watomist systems and he told me that it's kind of rough because it's kind of black box , you know , with the manufacturers knowing everything and designers not knowing that much . And you know it made me think . It kind of made me resonate .

Even though we were highly optimistic with Max about this cooperation between the designer and the water mist companies in the podcast interview , I think in the real world there would still be examples where this knowledge is not that easily accessible , or perhaps it is and people just don't know it is .

In that case there's a communication issue , and you know I'm crazy about good communication between different silos in fire safety engineering . Otherwise we can't solve the problems of fire safety engineering . So , yeah , I would like to bring your attention to this topic .

I wonder what are your experiences working with water mist suppliers and obtaining the data that was necessary for you to succeed with your modeling tasks , and I wish you all the best in getting that knowledge easily , in obtaining the data and being able to do some good modeling that has clear objectives , that answers important questions and that solves real problems .

Thanks for listening to the Fire Science Show and more fire science coming your next week on Wednesday , as always . Cheers , bye you .