Contractual Workstream Meeting 09-05-2023
MARTIN Let me share this thing that Iain shared earlier. George, you helped him with some of the questions for the survey, I don’t know if you got involved, Paul, I didn’t. He's asked for feedback if we’re interested. He said in his email that they’ve had 58 responses, so they’re starting to get some sort of data behind it. Chair of the Digital Construction working Group ISG. I’ll stick that on the chat. So, that’s come out since we last spoke. I’ve shared that thing from Cobuilder with you, George, is that of any interest?
GEORGE Yeah, I’m aware of it. Cobuilder are basically another player, but theirs is a charged service. They’re competing with MBS and Bim Object and all of these other people as well. If they make it free that would be fine. What they’re trying to do at the moment is get the templates together and quite frankly that’s what we’ve spent years doing with the CPA.
MARTIN Have you seen their webpage? Quite interesting in as much as I think slowly but surely we’re all coming to the same realisation that if we’ve got a backend library of sorts and we digitise our information exchange, a long list of what we want from everybody as and when using 19650, agree the dates, we should have a workable model that becomes a solution. In a nice way that says what we’ve been saying for the last 12 months, doesn’t it? In terms of a graphic.
GEORGE Yeah, and it’s what we’ve been doing for the last 7 or 8 years. What we’ve been trying to do is make sure that it’s not in a proprietary format and the difficulty is, t some extent, you mentioned 19650, it doesn’t do anything about defining data standards, it’s really processors. It’s the data libraries itself that is the big challenge as to how we get those defined.
MARTIN Yeah, 19650 allows you to put up a framework though to say this is what we’re going to do, this is our information exchange requirements at all of these stages from you and if you can get the designers and the CDP packages to sign up to that, you’ve got something to manage then in terms of information flow.
GEORGE Definitely, I agree with it. But the underlying data needs to be in a neutral standard format, in my view, not a proprietary format. So, that’s why we created the Templater as a basis for doing that. That’s quite a mature technology that can be used.
MARTIN They must know about that. You know them, they must know of you guys and what you’re doing. They must know it’s going to fall over if they’re gonna charge for it when there is a free service elsewhere.
GEORGE Yeah, they’ve done everything they can to undermine the free service because they don’t want it to be free. Where this originally started, Martin, is that the CPA came out with a desire to create a standard free service for manufacturers. And BRE and Cobuilder were competing for that particular contract, and BRE won it using the Templater and Cobuilder then said can we partner with you. So BRE, we agree that we should partner with them, so we went into a collaboration with them whereby we provided them with the Templater as a means to hold their information, because what they’d got was data that they’d been compiling on behalf of manufacturers. But when they realised it was also going to be available to other people like MBS and Bim Object, first of all it became clear that they didn’t have a lot of the data that they said they had and secondly they decided they were going to produce a competitive product. That emerged over probably 2 year and in the meantime they just spent quite a lot of time with CPA agitating and preventing a free service because they didn’t want it to be free. And that has undermined the whole Lexicon initiative, which is very frustrating.
MARTIN What does the future hold, then? Is it still full steam ahead? You’re talking about the Lexicon project’s future?
GEORGE I think the CPA have just made it too complicated, and also a lot of the product manufacturers are reluctant to engage with standardisation anyway. It’s a challenge, I don’t know the answer, to be honest.
PAUL Did you see the CIAV? RIBA Guide that came out recently? Safety critical elements. It’s interesting, what have all the guides got in common? None of them actually tell you how to do it. Apart from the FIS Guide on how you do best practices on penetration seals, which is a part 3 guide only, it gets you to a point where you can possibly select some penetration seals and it gives you a bit of an education around it. None of them tell you how to do them as individual components and put them together as a system. That’s why all of these different things that are coming out aren’t really solving the issue, it’s telling you how to look at the dance-floor rather than dance around on it. No one’s on the dance-floor, they’re just sitting around watching it and that’s half the problem, no one understands how you jig it all together.
MARTIN I think you’re right, I’m seeing more and ore stuff coming through, thinking it’s just adding to the layers of complexity here, there’s more to take onboard. No one’s produced a solution. I think we’re heading towards it, though, I think most people are clear in their heads now in terms of if you can get your list of deliverables sorted before you start digging holes and captured, and it’s streamlined in terms of 19650 or information exchange. Then, if you do have to wait that bit longer for the design to be resolved then you’re gonna have to wait that bit longer for the design to be resolved, not 6 years (as we alluded to earlier) but maybe 2 1/2 to get it right. Then fine, so be it, it’s just gonna mean more upfront costs for developers and central government and local government and I don’t know if they thought that through when they introduced all the legislation that’s forcing us all down this road of we can’t start until the design is done.
PAUL It’s interesting, because you look at it and you go, look, everyone’s arguing around costs. People are arguing around different manufacturers for dampers and ducts etc. When you go, let’s just take it back a step and say, these are your words, George, what is the risk of the space you’re trying to address? And they’ll go don’t know. So how do you know what you’re putting in it? They’ll go I’m just looking for the best price. I think there is a bit of traction to get there, but at the moment I think people are getting there too late. It’s about trying to get it to come forward a little bit. My gut feel is it will take a good few years before the BIM world picks it up and fixes it. The reason I think that is, when you look at those sheets I put together for George a while ago, you go fire rated duct, there’s something like 47 bits of asset information you need to record in the model. And that doesn’t include the next possible 60 or 70 odd that you need depending on where you’re putting it. So I think it’s gonna take a while to get that…
GEORGE Paul’s absolutely right on that.
MARTIN I’m seeing it, in terms of the BIM world catching up, I get that. But I think in terms of the designer’s world, which is where I was, and the developers world, I think those people are beginning to realise that instead f comparing apples with apples, which is what we tend to do, stepping back from it and saying what size of orchard do we need. They’re coming away from…
PAUL Hard questions. They’re still not pinning the tail on the donkey in the right spot, they’re getting closer but are still that far away.
MARTIN It depends on the customer, it really does depend on the customer.
PAUL I totally agree with that, it depends on the knowledge of that particular customer you’re dealing with. That’s a behavioural thing, if everyone had the habitualness to listen to other peoples views you’ll probably get there a lot quicker. But it’s that bit, I’m not paying for that, I want something cheaper. Well, this thing does the function that you need, change the risk of your space, I’m not doing that. Well, you can’t have it both ways, can you? When you get to that dynamic argument of actually there is the safety case, there’s the thread that goes with that safety case for that wall with that stuff in it. It’s getting to that point where we’re almost there but probably not actually anywhere near it, we’re quite far away, depending on the customer.
MARTIN In the last month we’ve had two interesting developments. We’ve turned a customer down because he wasn’t ready and he was shocked to his boots, not a big job. The more we said you’re not ready, you need to go away and get your designers further on, you need to do a lot more work on the design before you can engage with us, we’re not prepared to start until you get more done. Good customer, absolutely shocked. Our team were talking to the architect and structural engineer, they were in a different world. What’s all this, why do you need all this information? It’s the law now, guys.
PAUL Someone raised a really good point the other day, they said we just need help because we’ve got one contractor who’s saying we need 5-600 mil of space in the riser to do all the seals and all the connections and the other one is saying they can just do it tight and flush. And I said the one that’s saying they can do it tight and flush is obviously going to be at fault here because they’re gonna learn the lesson at the installation phase that you can’t do it. And that’s the difficulty that some architects are having and I can fully appreciate it, conflicting information. Martin, do you speak to Richard Ashton a lot?
MARTIN Afraid so, yeah.
PAUL Obviously I’m only in the process side of it and I’ve said before in the last couple of groups, I’m getting to a phase where I need to bring this and that in to meet. We’re getting to the point where if you look at it, a lot of Tier 1 contractors still think that the problem lies with the builders work, and it’s not. It’s about getting the right product for the right risk and associated systems of the wall put together before you even get to try and coordinate it. I think this whole thing on the BIM side of it, I think it’s gonna take a lot longer to get that to an order where the Templaters and things will link because you’ve got to get it to the point where the manufacturer is onboard, the client’s onboard the architects onboard and that common language is being used.
GEORGE OK, just to go back to what Martin was asking earlier, were are we going with all of this with Lexicon and the like, the biggest challenge that I’ve found is to actually try and get from people what attribute information is needed to be added to a product to be able to be used. That’s the whole point about having the templates. One thing that’s become very clear, and it has taken me years to get here, is that depending on how that product is going to be used the information is going to be different. And it comes back to what Paul was saying earlier, the context in which the product is going to be used is the determinant in terms of what attributes you should be asking for. So, there is some attribute information, in fact a lot of it (probably more that 80%) is fixed, and therefore it’s common to every time that pump is going to be used. 20% of it is variable and some of that comes from the context in which it’s being used and the other is just finally its serial number and things like that. So what we need to be doing is looking at that in that sort of evolving detail.
Now, the reason I still think the Templater is the right solution to this is because it’s a data dictionary which allows you to put in there all of the variable attributes with their description in a common way. With synonyms that allow you to describe things in different ways, but they’re all referencing back to the same attribute. And you can then put the unit of measure in there as well because as Paul was saying earlier the problem is everybody is working at too high a level of detail here. What most people don’t appreciate is IFC, for example, which is the international standard for data, as far as the attributes are concerned it doesn’t include unit of measure. That’s because in the UK we’ll use millimetres, in different European countries they’ll use centimetres, or they’ll use an imperial in America, so therefore what you need to do is to be able to have that as a variable that can then be commonly accessible. It’s fundamentally what we’re trying to do with these data templates which then can be applied once you know where it’s going to be used.
The product library, the solution that we’ve built, OK, you might say what we’ve done is proprietary, but we’ve done it in such a way that it’s open and free and it uses all the standards. It allows you to be selective in terms of what you need, so you’re actually picking something for that particular context. And that’s why I’ve been following this route of making the product library free because you’ll notice in that collaboration you sent me through by email, Martin, you won’t see any of the of the other product manufacturers library solutions, MBS or any of those on there, because Cobuilder is competing with them. Now, that’s understandable from Cobuilder’s point of view, but it means that we don’t have a…when you, for example, as Balfour Beatty, want to get your data it’s just another silo. That’s why I think we need to have something that’s at a fundamental level that’s free.
Now, how we fund it, Martin, I genuinely don’t know. I’ve given my IP away, effectively, in terms of that process and we’re getting quite a lot of traction from some players on that, including people like Lloyds Register. And if we can then tie that back to maybe cost data and also environmental information as well, if we’ve got carbon and fire those are two key elements. So, that’s the approach that I think we should be taking.
MARTIN This is going to sound weird, George, but like Paul was saying earlier there is going to be a while before the BIM world catches up, as it were. I don’t think that’s a bad thing, I think it’s needed before we can go down the route of doing it properly, like we were laying out in terms of getting the design right before you start. I think it’s an improvement at the end of the process, and it will continually improve. I don’t think it’s a you have to sort this out before you can do it properly.
PAUL Do you know what’s interesting, everyday is a school day and I always suspected that pipework was the hardest one to select and everybody would go to me ‘you’re wrong’. Something about it wasn’t right and I’ll tell you what, it’s the bloody hardest one to select. Because what we always forget, a bit like in dampers you have to understand what blade type you want because it screws your calculations up if you get it wrong. In pipework the material has got to be appropriate for the function for which it was designed originally which is the fluid its carrying. So what happens is you’ve got to look at the information requirements for thermal, you’ve got to look at in relation to what joining system you use and all the rest of it, before you even start talking about the combustibility of the material because that’s affected by the period of time fo the resistance you need, whether it’s two or four hours. Some materials flip into combustible after two hours. When you look at how everyone goes to Hilti or Rockwall, they go here’s a load of pipes, that’s not combustible, can you give me a seal for it.
They’ll have a go, but they won’t get it right, because we’ve actually got to tell people in the non-combustible world the thickness of the bore of the pipe, its reaction to fire, to make sure the seal that is being used is compatible with the bore thickness of the polyethylene or the HDP or whatever they’re using, the reaction to fire was the same as the one that was tested. None of that is happening before we get onto the next bit. We’re still going out the market at the last minute. What I’m trying to do is, the first thing is to get a proper code of practice that everyone can dance around, so everyone knows how to behave. Like Code-M doesn’t tell you how to commission a building from CIBSE, it tells you how to behave around it. Then you’ve got Guide A does air systems, W does water, that tells you how to identify those systems in your building, and BSRIA tells you how to do it. And the same principle applies to this.
So when you look at it you go once you know the appropriateness of the material for what you’re using, then you can look at the pipework seals that you want to put in because there is continuously sustained insulation throughout. It could be plastic changing to metal which is uncapped to capped, all of this stuff has got to be considered first before you go out and at the moment we’re just not doing it. And it’s all in the FIS guide about part 3 penetration seals, but actually it’s so big no one’s really bothered to digest it. You’ve got to kind of get it in pictorial form for an architect and a building services engineer before it even comes down to the likes of us, because otherwise they’re just gonna keep going around in a bloody circle, they’re never going to get it right.
MARTIN I was telling you about that customer with his design team. Their reaction surprised us, but at the same time didn’t surprise us, but they were nonplused by our approach. At the same time we’ve been approached by another customer who you wouldn’t dream of us doing a development for because they are normally associated with more dynamic, possibly, construction companies, less risk averse, possibly. And it’s a market that we pulled out of, deliberately, 5 or 6 years ago which is massive resi schemes, multiple stories in city centres. But they’ve approached us because they’ve heard that we’re trying to do this one step at a time and they’ve said we want you to talk us through the process step-by-step so that we can manage our designers. We understand about being responsible and accountable for the designers and we need your input. And we’re talking to them, and you honestly would not associate us with this developer, cutting edge.
PAUL As you said earlier yourself, Martin, it all depends on the customer, whether they’re motivated. One of the biggest risks I see is getting in the student models, we had it before back in the late 2000’s where people carved them up and chopped them, flipped them off to someone else and disappeared. They’re one of the big ones because it’s got everything, it’s got the sleeping risk and everything you want. Actually when you start getting into the equipment you need to satisfy those risks and actually perform…the leakage rate, it starts to get a bit expensive to do it properly. The famous word kicks in, value engineering.
MARTIN They want the words but they don’t understand what it is that they want. So, they want it to be carbon neutral and then you try and explain that to them and they don’t really want it. They want it to be insulation mega-performing and when you explain that to them they don’t really want it, they don’t really want passive house, they just want the word. It’s the same pain when you have to go through, you’re gonna have to comply with the law now, chaps, when you’re above 18 metres and people are going to sleep in your building. Yeah, we want all that, well, this is what it means, you’re going to have to spend the money up front. Not so keen now.
PAUL Someone said it’s a bit like going to France, paying for the course to speak French, but only getting the accent, you’ve actually got to learn the language as well. That’s absolutely spot-on.
MARTIN I think customers generally, compared to 12 months ago, it’s beginning to dawn on them. They’re getting better advise that this is coming and they can’t do anything about it, it’s now in place. And there are a lot of contractors out there who will do whatever they say, like the good old days, but there are a growing number who say, no, hang on, you’ve got to be ready for this, because if you’re not we’re not going to go dancing with you, to use Paul’s analogy. Designers, by god does that vary. Some of them are getting it and some of them are still carrying on with blinkers.
GEORGE One of the challenges that I’ve got is we’ve had the solution to a lot of this stuff for quite a few years. The problem I’ve got, first of all we’re a small company and therefore we’ve not been able to market it, but the technology is very sound. And also we’ve been through the process and we’ve tested it on live jobs, so a lot of what we’re doing is not theoretical. Second thing is that in BIM4housing I’ve obviously been fairly careful about being overtly promoting my solution because what I’ve been trying to do is to be neutral and encourage other people to engage and I think I’ve been pretty honourable about that, otherwise people wouldn’t continue to participate in what we’re doing.
But the difficulty is at what stage to I actually step out there and say, look, we’ve actually got something that resolves this. On a call this morning we’ve been appointed on 6 PPP college projects across Ireland and the methodology we’re using there follows on from what e’ve been doing with Sedexo on a PFI scheme there and also with a construction company called Rattigans that’s delivered 9 schools on D&B. And they’ve followed the process of using the product library and the Templater to actually compile the information and glues together what they’re doing in the BIM process and also generates the O&Ms. But what we’ve been doing, and this has evolved from the work we’ve been doing with Bouygues, is we’re looking at the technical submittal or RDD stage as being the final stage at which you pick the product.
You and I have spoken about this, Martin, clearly it’s a lot better if it can be done much earlier than that. But at whatever stage the determination of what you’re actually going to use, in can be as early as you sensibly like, that’s something that can be done at that stage. And having standardised product data is a fundamental part of that process, especially if it’s in a library that then the manufacturers are willing to keep up-to-date. I’ve spoken with 40 or 50 manufacturers now, fire related products, and they’re all willing to provide their information in the library environment. And I’d quite like to…I think the only way of doing this would be to maybe have a 1-to-1 with each, with Balfour Beatty, because everybody has got a slightly different process, and do the same with Mace etc. We’ve also got Travis Perkins on board with the idea of them doing their side of things.
PETER I’m from Bulgaria and work for the construction company, Balfour Beatty, part of Martin’s team.
GEORGE So, I don’t know quite how to move the thing forward, it’s very frustrating. I don’t know if that’s something that we could run a session just to test out.
MARTIN I get they’ll be an interface period and an overlap period, there will be times when we need that information when we’re doing the install, but I do think if we set off down that road it’s the wrong road to set off. We should be getting that information before we start doing install, that should be in place and the install should be a quality control procedure not a quality assurance procedure.
GEORGE No, I agree. What makes you think I’m not saying that? Because I used the term RDD?
MARTIN Yeah, or technical submittals, I think that’s way too late.
GEORGE So, could you just talk me through it. In terms of the product selection, what would the process be that you would use to do that?
MARTIN RIBA stages 2 and 3.
GEORGE Yeah, I understand what stage it is, but is there a process. Basically what you’re saying is you’ve got a designed product which is a descriptive specification.
MARTIN You’ve got a design and your looking for solutions and you want to specify, so it’s back to the old Iain McIlwee difference between a performance specification and a general specification.
GEORGE And if you look at it from Balfour Beatty’s point of view, that product selection process. Who does that and what’s it called?
MARTIN Design development. Overall design development, it’s done by the CDP package partner and the professional designer. Most of what you’ve talked about there would have been the M&E engineer and the M&E contractor. But they both need to be engaged and somebody paying for that service.
PAUL It’s a hard one because there is not a standard definition in JCT for design and development. Normally if you’re gonna put something in a contract it’s a pending in. Martin, this is a really hard question to answer, by the way…
MARTIN I think it’s important and George is right to raise it, how do we nutshell that particular phase.
GEORGE How is it different, my understanding of technical submittal is that you’ve got a performance specification and you’re then selecting a product that the technical submittal process is then getting an approval that the product that’s been selected is matching the performance specification. Is that right?
MARTIN Yeah, that’s what we do at the moment, generally speaking.
GEORGE So, you’ve got a performance specification. If that tech sub process is actually brought forward to pre-construction then it’s still a tech sub process, isn’t it?
MARTIN It’s a process. It still needs to be done, but I would call it development.
PAUL It’s a real hard one to answer. If you look at it from a contractural point of view, design development only exists if it’s defined in the contract. When you look at it in reality, the reality is that if you get a performance specification and someone does a technical submission against performance specification, I’m gonna give a classic one of a fire damper and they go away and they find something and give it to the building services guy to give an A to. You don’t know actually whether what they’ve selected is correct because the magic numbers of how you select it in the space risk just won’t exist in the conversation. If a building services consultant is saying A-status it doesn’t mean it’s actually appropriate to where you’re putting it and it’s actually not worth the paper it’s written on. That’s the current downside of the way that the industry behaves, you can’t guarantee…it’s kind of like that procurement thing for modes, poor tender specification where they go fire damper, CDP.
You look at it and you go well, what is it, what’s it protecting against. Everyone’s asking the question, it gets thrown down to a supplier who goes is this all right? Yeah, it’s fine, you can use that. The same thing is happening with pipework but in a different context because the original consultant has not really specified the pipework anyway and go contractors choice, the contractor goes away and…pipework later on as cheap as chips, the best they can get their hands on. Then it goes over to someone who buys the passive fire seal for it and doesn’t even know the material that’s been selected, doesn’t know if it’s non-combustible and this is where that disaster all kicks in, doesn’t know the size or the bore thickness of it. That’s the kind of issue.
MARTIN I think it’s tricky for George because he’s thinking even when we do the tech subs it’s done 100% perfectly and it often isn’t. There is stuff left to chance at times, everybody has approved it but it’s still not right. For me the whole premise of the Building Safety Act is you’re not doing that when you’re trying to put up a building and finding that the bloody thing doesn’t fit. What you’re doing is you’re doing that before you start building.
GEORGE The way I’m looking at it, whatever we call it I think we are going to need to find a term. Maybe stage 1 tech sub or whatever.
MARTIN I don’t see it as a submission and an approval, I see it as an iterative design process.
GEORGE But at some stage you’ve got to have a selection, haven’t you?
MARTIN Out of that process should come a bang, we’re having one of those, go away and price that up. But that informs the project. There’s no veering away from it, thats what’s going in.
GEORGE Unless there is real need to change because that product is no longer available.
MARTIN Exactly that.
PAUL The hard bit is, and Iain McIlwee is working on this with Joe Silia, is the whole walls as a system approach. When you look at the 5 QA points that exist, the space risk, building type. How it operates, pipework is static so it’s simple, smoke control you’re doing something with it like sprinklers, fire doors are transient because you’re going through them, they’ve got a different operational functionality. Then you can look at how you classify it against those two things, how it functions and sits within the space risk. Then you can look at the type of wall that it works individually in, then you can look at actually is the geometrics there so you can put it in a wall. Then you’ve got to look at all the other products in that wall…to get to the right type of fire wall to house the component. And that’s the bit that when you look fo 10 specifications of and people are uncoordinated and the rest of it, it’s because no one has done that exercise to get to that point of performance specification, which really I think should be the classification number and the appropriate wall it goes in and you can go and select a product against that number.
GEORGE My attempt with this is to try and put some sort of standardised structure behind it. At some stage we want to deploy AI in terms of doing this, but before we get to that we’ve got to get human intelligence to be systematised so that we’ve got the rules and the relationships, like in the decision tree that Paul’s been putting together.
PAUL (shares screen). Just to add to it. The fire damper one has been done and has been presented. This is the bit where I need to get you involved, it’s all about how this works as a system. Because that’s the complex version and that’s the simplified version. That’s your five QA quality assurance to get to your magic number and all your installation. And then it’s around how you do the dampers, it’s all about getting this bit right as well, the blade types for pressure loss, and how you select your frame types against the wall. And the same sort of thing for installation. This is the kind of thing that you need for that machine learning because this filters it all out.
But pipework is the absolute git because you’ve got to consider all of this…it’s just not the same thing. But I’m trying to get them done for all the different things…It’s got to be such a way that you get back to that magic thing of actually what you’re looking for. I’ve got to put a note in here about where the joints are as well. It’s not easy to get it right, and even when it comes down to pipework, if you’re going to a pipework penetration seal manufacturer this is what you’ve got to give them in scheduling form for them to give you a seal. It’s all the same stuff, now on Excel as well. I think you’ll end up where you get the BIM world spot-on with things recorded digitally like this before you get there. So you know what it was, but you know what you’ve selected against it. This is ductwork, but trying to get all of this into the BIM world straight away is going to take the pre-effort to get people to embrace doing that part.
GEORGE To some extent that’s why the Templater is so important on this because it filters just what you need to see for a particular purpose.
MARTIN This might sound weird, but it’s just dawned on me whilst Paul was going there, what we’re trying to do here…the solution to all those things that Paul threw up there, currently that comes from the people doing the design. The designer who’s got the knowledge, the skills, the experience and the expertise in order to do that and to know what the solution is. What we’re trying to do here is unpick all of those things that he’s learnt and been taught and put it into a series of components, products, simple bits that make up this thing that he’s designing, with all this information that Paul was saying effects it. He’s using his skill and expertise to come up with a solution whereas what we’re trying to do here is find a way and means of picking out each and every part of that solution and getting something that will pull it all out of a database somewhere and put it together and say there’s the solution. It’s bloody hard, that is, George.
PAUL When you look at this Tier 1 group, obviously Balfour is a part of it, getting that group fully aligned, and I’m getting there with it. The worst thing you can do is sit there and think part 3, part 2, part 1, 5 and the rest of it, that they’ve got any relationship to each other because they haven’t. And you’ve got to remember that the ventilation system, before the fire damper is inserted it’s there to perform a function which is pat-L and delivery of air…and everything has got to be right first because if you set the wrong component you screw that part up, that’s the consequence of it. In pipework terms, you get the wrong type of pipework and you start doing wrong types of jointing and going from different materials, you’re creating a penetration seal nightmare which the people who select the seals can’t do because they’re not in their level of expertise to do it.
So you kind of have to workbook it out where people get to go through what they’re doing at the principal stage because architects, MEP consultants, fire engineers, they don’t really understand the product classification. And you can’t throw under the bus for it, you’ve got to have a method to support them to get closer to the mark. The only way I think you can do it is by doing that. When it comes down to getting this BIM bit right, I think by doing that and having those it will make the AI easier to program later on.
GEORGE Can I just clarify, Martin, I think we’re several years off being able to do tis using AI.
MARTIN Is there a comparable industry where they’re doing this? Outside of construction where they’re further down the journey?
GEORGE I don’t know.
PAUL I can possibly answer this. I’ve had some conversations with the MTC about how they do aviation, they do turbines and things like that, because they can’t test everything, the same way we can’t test everything.
MARTIN Yeah, they go for a model or they go for a prototype.
PAUL They’ve got modelling which does the majority of the work, so the AI is kind of there already imbedded. But what they’ve managed to do, George, and it would be worth you speaking with Benjamin at the MTC, what they’ve ended up doing is they’ve got certain things that they test. They use that data about materials and what they’re doing and the process is to then model it so they limit the amount of testing that they have to do, because you can imaging taking one of the biggest planes into the world and smashing it into the ground isn’t such a good idea. You look at the Musk example, what he’s doing at the moment, he’s not sending them up to succeed, eh’s sending them up to spot the failures, that goes back into his model and then they move on. So for every crash there’s probably like a thousand iterations of model that are going on with AI before it gets to that point. So, they treat it completely differently.
GEORGE Yeah, I think what they also do is that they learn from what they’ve done and try not to do it again. Whereas in construction we don’t learn…
MARTIN You’re right, each project is a new rocket, we start again from scratch.
PAUL If you get a really tragic plane crash like the Air France one, really sad and they couldn’t work it out, it’s a bit like saying, well, obviously the steering wheel is at fault because it’s crashed into the ground. But it’s not, it’s all the other bits that go behind it, the causation, that led to that error. There’s a clinical comparison of previous work to look at the work coming up to compare what they’ve got, what they probably do is take out what’s left and go and look at that. Completely different, way more advanced than construction.
GEORGE Absolutely. That’s why I come back to the need for us to have standardised libraries of things that therefore become repeatable. And by doing that we’ve got things that you can QA. You can QA them at a template level, at a product level, at a system level, at a building type level, so you’re able them to have fully QA’d solutions. And by doing that it means that information becomes interoperable. If you don’t do that you basically do it fresh each time with people, for example, just putting attributes into BIM models, then you’ll never get to where we need to get to. And you’ll also always have to be checking things manually. Whereas what we can already do is we can also make the checking and the validation process.
MARTIN But can we do that with the design process? That’s where my head is at the moment. Can you design from a list of products?
GEORGE You’re not designing from a list of products, you’re designing from a list of generic items that have got attributes that satisfy the generic requirements. So, by doing that it means that you go through 3 steps: You define your information requirements in a machine-readable format. So, there’s a requirement, often they’re performance requirement, you’ve then got a product which is probably a generic product that is your design solution to that particular requirement. So, you’ve therefore got something that is clearly auditable and you can do comparisons between what was required and what the design solution is. And then you compare the design solution with an actual product. So those are the 3 steps that I would say.
PAUL It’s the same thing, George, it’s a case of if the product manufacturers display the information with all the essential characteristics like the 40 odd for smoke controlled ducts or fire rated ducts that I was talking about earlier, then you can use that as a generics for putting it in. Then you can compare that against the product you selected because it should be in the same language.
GEORGE Correct. That’s where the Templaters come in because you’ve got the language in a machine-readable comparison. But the reason I’ve backed off doing that is because that’s the area that seems to be the sticking point. What I’m saying is that the generic requirements for the generic product, we can have that in a machine-readable format, so you can test the requirement against the prescribed product, as it were, which isn’t necessarily a manufactured product but it’s a solution to that that you can then do all of your checks against and then as long as you’ve got a product data sheet for that unique product you’ve at least got the opportunity…And you can slowly see where it is, the context in which it’s going to be used, then the human can apply the real intelligence to make that assessment. At the moment the way tech subs work, whenever that is being done the product selection bit seems to be overwhelming for the design team to actually do that. Whereas if they had clearer information on product data it would make that job much easier.
PAUL Let me just share a diagram because I think this will make it a bit easier. When you think about it part of the problem at the moment is the behaviour around it. Then it’s about classification of all the different products, how you compare them and how your builders work is and how you record it. A this stage if they’ve got a better understanding of how you do that then you’re gonna get to that point where it’s going to be easier. What’s gonna happen is once you’ve got that you actually gonna do this initially into there and get it in the schedule. BWIC is Builders Work In Connection.
MARTIN I spoke to Richard on Wednesday, John Rawls is doing the same thinking, we’ve got quite a few exercises ongoing about timing of design and depth of design within the organisation at the moment. And it’s been brought about by the law changing, there being a requirement to do this and also because we’ve seen for a while now that we can’t carry on reinventing the wheel every time.
GEORGE Coming back to my question, the product selection is a combination, it’s in part the performance specification that is required, you’ve then got a prescriptive specification of what the solution to that performance specification is and then you’ve got the selected product. Is that right?
MARTIN Looking at it from a structural engineers point of view you always pick something, you pick a solution and you do all the maths to show that it works. You don’t do the maths and the answer pops out, you pick an answer, that’s based on your experience, and then you prove that it works. You do all the maths and all the calculations.
PAUL I think it starts with 9.81, Martin. I can see it with structure, you pick a solution that you want and you make sure that it works. The hard bit is with all these other ones like ventilation, the underlying criteria is that it must deliver enough air for you to breath and it must do it with an energy criteria that doesn’t exceed 2 watts a litre for a fan, then you work your way back form that criteria against all the space risks in the building to deliver it. So it’s the same principle, the problem is the complications of it involved multiple different space risks, whereas structure is just REI.
MARTIN You’re right, you start off with a solution and then you work and see whether it stands up to the various variables that you’ve outlined. You make sure it works in every scenario that you can imagine. That applies to any engineering solution, doesn’t it?
PAUL This is good because when you look at how I’m trying to deliver messages this kind of thing really helps. When you look at looking for an example to use it’s really good because the thing with pipework, you draw the pipework system out in the building, deliver X amount of kilograms per second whether it’s cold water or hot water or domestic, then you work your way to ensure your pumping power is not too high and the material is appropriate for all the circumstances for which it travels as a system. And that’s the bit where it goes wrong because nobody wants to do that grunt work, that’s why they CDP it. Dealing with pipework either side of the wall you need an 8539 fire resistant fixing to support it, but if that pipework goes across the top of life safety critical systems then you’ve probably got to put fire resistant fixings throughout whereas sprinklers are actually testing the fire resistant fixings and so is the smoke control duct, but general pipework is not. It’s look a bit mind blowing when you look at how far you have to go down the rabbit hole for things like pipework. It’s a better way of explaining it, mind, how you’ve done it there, this is why these groups are so powerful.
MARTIN I find these groups really useful as well because it does help me think through all the scenarios and what’s going on at the moment.
PAUL When you look at, George, what you’re saying, you’re absolutely right, someone designs a ventilation system to travel through a building, you then go back and you’ve now got your solution for how it’s delivering air, you go back and you pick your componentry to suit the space risk of a building. That’s when it becomes a bit tricky. For structural engineering, the one that always causes the issue is are you suing reactive or non-reactive coatings, that’s where it becomes a bit lairy. Because that’s again a summation of the space risk what it’s within and what interfaces against the steel. If it’s drywall you use an encapsulation, if it’s non-drywall you can use reactive coatings. That’s the trickier part of it. If you look at it and reverse engineer it it’s probably a better way for me to explain it.
GEORGE We could actually do this as an exercise, Martin, with Mace and Bouygues, and Waites or whoever, around the standardised information requirements for some fire critical items for a building and then look at what the product solutions for those are and then also the context. So, in other words, the spaces that they are supporting. Because I think that’s broadly what we’re saying, it’s a matter of the information requirements, the performance specification, the context in which that product is going to be used and then what product is being selected.
MARTIN Yeah. Deep. I just wanted to share something, George, that you put me onto from a conversation some time ago. I found this which you might find interesting (shares on screen ‘Energy and Carbon Reporting Framework’). I just found it very, very useful. Page 11, there is a lot on there, there is a lot to go on there. Started thinking about layering on Firey, 19650, government soft landings which is now a British standard. And that thing you referred to earlier, Paul, safety critical features. Then there’s the mother of all documents that came out of the Risk Authority. That’s the Ian Abley document from the insurers, it’s an astonishing piece of work.
PAUL I had Ian Abley speak for us at the Public Health Conference, the A-Z essentials.
MARTIN Yeah, he’s managed to take what is a quite complicated 8 stages and eh’s split it into 26, top man. Back to this, I think you mentioned MTC, there they are with the University of Cambridge.
PAUL I’ve got a contact for the MTC, I was doing some stuff on dampers and things, I’ll get a contact over to you, George.
GEORGE Has anybody done anything in terms of comparing the gateways across the different process maps.
MARTIN Yes, quite a few people have done that, mapped it on to the RIBA stages.
GEORGE I’ve got it against the RIBA stages, I wondered if it was against any of the other work stages, like structure etc.
MARTIN We’ll do it on the next one. We’re trying to map out the RIBA stages against the IStructE stages against the BG6 stages.
GEORGE That’s exactly what I’m looking for.
MARTIN So am I! But somebody’s got to do this. I’m looking for a resource from somewhere within the organisation that can help me otherwise I’m gonna have to stop sleeping at nights in order to do this.
GEORGE The design group was looking for it last week.
MARTIN I bet. It’s taking a long time, but I think we’re all coming to the same thinking. But, we did some work with CIC and CLC, Balfour Beatty, one of the managing directors got heavily involved in that, and they are not ready. And they’re reporting into DLUHC and frankly they’re not ready. And the fact that the regulator resigned a couple of weeks ago, before he got going, it doesn’t get any weirder, does it?
GEORGE Not at all, I agree. He’s almost certainly gone into private practice.
MARTIN Well, he’s said, hang on, I can’t do this on my own, I’m going to need some staff, a bigger van.