HEC-RAS 6.0, what you need to know: Part Two

Video Statistics and Information

Video
Captions Word Cloud
Reddit Comments
Captions
[Music] welcome to the australian water school the home of demand driven industry design training for the global water sector hi i'm craig price with surface water solutions welcome to part two in our webinar series on what's new in hec-ras 6.0 today we've got two expert presenters on for you probably the best that you can imagine getting for this particular topic would be stanford gibson and alex sanchez former the u.s army corps of engineers so this is an exciting topic that has worldwide implications when you look around at this map it it always blows me away and i always mention this every time but it's amazing to see the spread of attendees and the interest out there in our particular case today in how water moves sediments so i wanted to introduce uh with no further ado our two uh expert presenters and maybe we'll get marty to come on as well stanford gibson and alex sanchez you can read off what their qualifications are here but really they speak for themselves in having developed these modules we've got others on board as well who will answer your questions chris goodell is in the background and gary brunner and others who will answer q a questions but i wanted to just initially um as part of the introduction see if anybody had any guesses um as to that poll question which said how many of the attendees today have used hec 6. now this is really an old school question for you how many of you out there are old school and have used it before and what i'd like to see is uh how accurate have we can we simulate and estimate uh what the response is going to be so i'm certainly one of those who learned in fortran and dos systems of hec6 some of these uh you know several hundred attendees that we've got on the call um over to you alex and stanford how many would you estimate have uh would have used hec six in the past well i know three of us showing right now have so um that's put that puts a floor on my estimate but i'm gonna go with six percent all right alex i'm gonna go with five percent five percent now last week okay let's see if you did better than gary yes last week gary brunner um or a couple of weeks ago when we had this i was surprised at the 25 of the attendees uh of the almost a thousand over a thousand attendees had used hcc2 so we felt like we were skewing a little in the older direction there so he i don't know what his estimate would have been but he was way off so let's see if you guys did better than them um joel if you can put that uh the the results up here we'll see how we did and i think we were uh both uh wrong and and i would have guessed but i i would have guessed but but i saw the results coming in as they were coming in and so i kind of realized that uh we were all off a little bit in our estimates 25 so a big surprise marty what do you think of that would you have expected 25 no i was going to go a little bit higher than these guys i was thinking maybe 10 but um yeah 25 i'm surprised yes because it's been probably 30 years right well 20 it's 20 since it got built in um so this means that there's quite a few of us that uh on board who are showing the gray hairs a little bit so um in any other comments on that before uh before we dive into things get uh gary and chris feel free to weigh in on that as well because the other the other questions are whether you've used any of these uh hydraulic models for sediment transport and what's interesting um is that only seven 19 percent of the attendees have used hydraulic software for sediment so what were you doing in hec6 i don't get that difference how come 25 of the attendees have used hc6 but only 19 percent have used hydraulic modeling software for sediment transport there's some chatted in the chat box people are uh saying that they were probably confused by the question thinking it was raz 6o okay okay so let's make that distinction right up um have you downloaded 6o and x6 i see that's that's a that's a good point um we uh that we and just as a background um the original dos fortran versions of hec one two three four five and six um turned into some of the other softwares that the htc has out there right now so htc six was the predecessor to all this so i think you're on top of it gary i think that's exactly the reason why these results are skewed a little bit high so we'll probably go with an actual real number maybe closer to what you guys estimate at alex and stanford so um our content today um i'll just uh maybe introduce the topic briefly and um i know we've um chris i think i got this idea from some of your uh webinars uh or from some of your courses i think i saw some of these uh videos in there i don't know if you can see this debris flow coming down the question is you know how does when water moves as water there's a lot of models that'll handle that and there's a lot of information out there but when water starts moving with sediment in it and some of these boulders are massive you know look at that flow i don't know if you can even see it it might be a little small on your screens but these massive boulders that are moving down and you know together with the turbidity in the water the mud flow the debris flow you know how does that affect um the the flow patterns and the hydraulics and the physics when you get non-newtonian type flow and when we model this looking down on it modeling it as cross sections in 1d versus modeling it as grids in 2d those are the topics that we're going to cover so there's a lot of excitement around the new capabilities in hecarius and the ability to simulate these types of flows um we are very excited to hear from both uh stanford and alex on this again marty chris and gary will be in the background answering your questions so with that then i will turn straight over stanford if you want to start sharing your screen over to you stanford excited to hear from you all right hello everyone my name is stanford gibson i'm the sediment transport specialist at hcc and uh today i'm presenting with alex sanchez who's sediment transport and numerical methods specialist at hcc and everything that we show you today he's had his hands on as well we're just going to take some turns presenting um we're going to talk about the brown side of raz when you want to plot your results not in blue we're going to talk about mud debris and sediment a lot of really exciting new stuff in 6.0 and so let's just get right into it this is what we're going to do we're going to start out talking about mud and debris flow which isn't sediment transport this is going to be non-newtonian fixed bed um flow and we'll show you what that means and then i'll turn it over to alex who's led the 2d sediment modeling in 6.0 and both of those are brand new those have not been available in previous versions of raz of course 1d sediment has been in razz for a long time but we've done a lot there's at least 35 new 1d sediment features in in raz so we're just going to hit a couple of those at the end but let's start with mud and debris so in january 2019 a pacific storm blew into southern california and dropped a bunch of rain on the mountains up gradient from santa barbara but the maximum intensity rain was actually 13 millimeters in five minutes generally the watershed can can handle that sort of rain but the watershed had just burned and in fact the watershed was still burning and that rain actually put that put the fire out and so you know the newly exposed hydrophobic soils you did not absorb that rain it caused a debris flow that ended up killing 23 people causing over 200 million dollars of damage and putting 63 people in the hospital now this is just this is a dramatic case of what's becoming more and more common we've had we've had some of our biggest fire years uh in recent history um in california obviously australia it has been in the same position and when you kind of mix that with these intense high precipitation events and climate variability modern debris flows they're becoming a bigger concern and so we want to do risk analysis on these events the problem is that the most common water resource software that's used for risk analysis is hc rest you know we've got over a hundred thousand downloads a year um in over 200 countries this is what a lot of people use to do risk analysis and if you were to say use just newtonian clearwater flow for this santa barbara event this is what it looks like so this is the watershed the lidar of the watershed and the blue lines are the inundation boundary that the u.s geological survey measured for this mud and debris flow in santa barbara and it's actually very easy to go out and you know find the inundation boundary of modern debris flow because they leak it leaves mud and debris everywhere and so if you were to model this with clear water newtonian physics in hcc ras this is what you get you underestimate the impact pretty dramatically and so we added these non-newtonian monetary flow capabilities to raz in coordination with our partners at the other core laboratory um urtic and uh so what i'm going to do briefly for this first section is as quickly as i can i'm going to introduce you to for those of you who want a refresher on non-newtonian physics and actually what's going on um in the in the equations and in the model then we'll show you some laboratory and meso scale validations that we did with the model to show you kind of how it's performing um at a really controlled scale and then we'll go back to these field applications and show you how it's performing on field applications so let's start out with just a little bit of non-newtonian physics so as you add solids to a fluid that mixture behavior starts to change now you have to have a lot of solids for that to be the case you know any alluvial flow including some you know concentrations as high as like a dam removal or a reservoir flush that's still not really enough sediment to change the physics of the material but for a muttered reflow like cray showed in the introduction when you get enough solids in the material the particle interactions start to influence the physics of the flow and so as you as you add concentration in these high concentration flows the you get different particle interactions that cause the fluid to diverge from what we call newtonian assumptions but it's not just as you increase the concentration this isn't a simple straight gradient these behaviors also change as you as you add coarser and coarser sediment and so as you have more and coarser sediment the behavior of your fluid changes and starts to diverge from what you would expect for a newtonian flow and so if you consider how we do clear water flow um when we when you compare you when you compute clear water hydraulics generally this isn't entirely the case but generally most of the force transfer is at the boundary it's between the clear fluid and the boundary of the fluid or the bed and we generally get most of that force on transfer from what we call the rough from the from the roughness but now when you have particles interacting in your fluid now you've got internal forces you've got particle collision forces you've got viscous forces you've got particle friction forces all of those change the dynamics of the material slow it down increase the stage increase the momentum and so we have to include that into the equations and so this is true for both the 1d and the 2d equations i show the 1d equations because they're a little bit more intuitive and the math is really not hard here we solve two equations the conservation of mass and the conservation momentum equation and the way that we account for this force loss in just clear water flow is what we call the friction slope and the friction slope inverts the manning's equation in order to get that friction force into the momentum equation well all we're going to do here is we're going to now add a second slope called the internal viscous and inertial loss slope or just the internal loss slope that's going to account for all these other forces and so it's going to be linear with a friction slope so it's pretty intuitive we're just going to have this additional slope in the momentum equation but how do you get that slope well i would answer that question with another question what is the equation for shear stress well the equation for shear stress is gamma rs or the unit weight of water times the hydraulic radius which you know we can estimate as the depth times the friction slope so you can back out the friction slope in terms of the shear stress well that's where the magic happens in the non-newtonian computations because you can also do that for the internal loss slope the internal loss slope is some sort of internal loss shear divided by the unit weight of the water and the depth so all we need to do is you know be able to cast our internal losses in terms of some sort of shear stress in order to compute this internal loss and now what we can do is we can leverage stress-strain relationships from a field we call rheology um to which is just the the physics of how different materials strain and so these are some of the simplest stress-strained rheological models um the first is just a newtonian fluid a newtonian fluid um the stress strain relationship is linear and passes through zero and all that means is if you put a little bit of stress on water it's going to deform it's going to shear it ca water can't hold up under any stress so it strains immediately but other fluids have different stress strain relationships and the simplest alternate stress strain relationship is a bingham plastic and so the blue line is the bingham plastic now this has a this is still linear it still has a linear stress strain relationship but it has a higher slope because this the slope of the stress-strain relationship is the viscosity of the fluid and the fluid is more viscous but then it also has this additional term the yield stress and what that means is that it has a non-zero stress under which things start to strain so you can put a certain amount of stress on this kind of fluid and it won't move and what that means is that the opposite is also true that eventually the fluid can come to rest in a hydraulic model water will never come to rest it'll just keep running out until it reaches high ground but a non-newtonian fluid will come to rest because eventually the internal strength will be stronger than the strain it's under now there are many other more complicated non-newtonian models um all of these are included in raz but the same basic idea is true is that we'll use the stress strain relationship to compute an internal stress and then back out the internal loss friction slope um and put it in the momentum equation okay so that is the very simplest that uh that we can uh we can cover the the theory behind that but let's show you how it works um does this work you know this is what we call a single phase non-newtonian approach to to the non-newtonian physics um and so you know as as most of you know those of us who who develop models we don't get to just develop a model and then release it we have to go through a series of rigorous um validation and verification tests before we can release it and so we have applied this to a full suite of laboratory and mesoscale models i'm just going to show you two of them so this is a flume study done by persons at all um in the early 2000s and what parsons did is they just took a pipe half a pipe they could they took a pipe and they cut it in half and they ran a series of very high concentration slurries down this pipe and uh we're talking about like 68 to 84 um solids by volume but the the cool thing about this model is that they this plume is that they used materials of very different gradations they used very classic materials and very cohesive materials and ran you know 30 different experiments with six different materials so we ran these experiments in raz um to do that we created a one-dimensional cross-section of this half-pipe and then we converted it to a two-dimensional mesh a two-dimensional mesh with one millimeter cells now i don't know what the world record is for smallest cells in a 2d mesh but i think i'm in the game here with one millimeter um cells but then we could make a two-dimensional model of this one-dimensional phenomena um in raz and so let me show you the result this is one of the experiments and here i'm showing you the velocity grid after 10 seconds of the newtonian flow and here's the velocity result in the same scale of the newtonian result um after 10 seconds now that's pretty small let me load up so now this is just a zoom of the same thing and um you can see that here is the actual result um with the at the bottom with the non-newtonian result versus the newtonian result the model actually is you know is uh dem is replicating not only the velocity but also the shape of the of the plug and then you know that's kind of anecdotal so here's the general form um in the general form here we're plotting observed velocity versus computed velocity and up here the blue dots are what you would get if you just use newtonian physics in raz and just ran water through the flume and down here is just the simplest non-newtonian model we have in raz the the bingham plastic with measured parameters these are actual measured parameters and uh you can see that the results start to fall along the line of unity now the other thing that we could do because this is a two-dimensional model is parsons at all measured lateral velocities as well because the velocities differ across the flume and so they the here we the open circles are observed and the closed circles are computed and so we could get velocities at each cell across the flume and compare them to their lateral measured and if you use a good turbulence closure you end up getting a very close result not only not only bulk but laterally so we actually have a paper that just got published on this this is uh earth service process and land forms um and you can see that we didn't just use the bing and plastic we used all the other rheological models and you can uh you can go see that if you want or if you're not super into reading journal papers like i don't know most of the world um we've decided that we want to make this more accessible so uh we developed a um what i'm just calling a audio visual methods and materials and this is available on the hcc website at this point and uh we basically can walk you through the process of creating this flume from scratch and doing all the experiments we did doing the experience we did in the paper from scratch um with a series of youtube videos um that you can walk through okay so that is the laboratory scale but you know how does this work on the mesoscale well the mesoscale analysis kind of the gold standard of debris flow mesoscale analysis is this usgs geological survey debrief loom um this work was led by alan iris by allen iverson and i guess i'll i'll play that this it's in the pacific northwest and they have a giant flume up there um that they send these massive debris plumes down and then they take measurements as these plumes go down and so um you know the the this uh iverson and his team at the the usgs flume um have just done incredible work this is the best they've done dozens of these experiments and so this is what we tried to simulate as part of our validation verification again i just made a 1d model of this um and then turned it in it into a terrain and then we simulated it in raz with the non-newtonian result and what they did is they used these these stage gauges to look at the stage hydrograph of the plume as it passed through and you can see here is the computed and the measured of the plume as it passes through and the other thing that we get is we get the the general shape of the of the of the final plume um which is a you know kind of oval lobe now if this was water this would just run out until it reached you know the edge of washington or high ground but because this is non-newtonian fluid the particles eventually interlock and so you get you get a plume that's at rest which we also capture in now one of the nice things about the way we've done this is we've collaborated with the other lab at um in the core of engineers urtic um under the leadership of ian dr ian floyd down there and we did not actually put any of these methods in raz all of these methods are in a library called debris lab and this library exists so that multiple models can call it so the other core 2d model a model called adaptive hydraulics or adh they call the same library as ras and so raz is a 1d finite difference 2d finite volume sub grid with them tree fortran code and adaptive hydraulics is a 2d finite area adaptive mesh c plus code and we're both calling the same library and we're getting very similar results and so this is another paper that we have in review um in nhs where we've we've replicated several of these vmv experiments both with adh and hcc ras with very different hydraulic schemes but the same debris library and uh and this and what part of the goal is that this would be available for other models as well okay finally field applications okay so let's go back to santa barbara you recall kind of how poorly newtonian flow did with the santa barbara analysis and so we re-ran this analysis with the simplest possible assumptions because you know coming up with these parameters can be pretty hard so we asked the question what would happen if we ran bingham plastic the very simplest rheological model with the default parameters for general soil that's available in julian's textbook which would be basically the best guess you could make out of the box knowing nothing about the system and we ran this um this event with those and here's the result and so you can see that we are now getting both the depth and the extent of the of the debris flow with much more accuracy and we did this for all three of these watersheds and improved the the result um dramatically and then the other big application of non-newtonian models is mine tailing damn failures and of course you know the most famous mind-healing damn failures are the two that have happened in brazil recently um but uh you know this is something that we're analyzing all over the world and so um the core of engineer has the core engineers has several brazilian um collaborators we work we we work closely with some folks in brasilia and so professor leonardo mora worked with dr calvin creech and hanato amoren um in order to model this event and so uh what you're seeing on the left is the actual event itself most of you probably seen this video um and what you're seeing on the right is the non-newtonian simulation of um of this event in hcc ras we get not only the extent relatively we get not only the extent correct but we also get the timing there are two places two locations along the way where you can measure the timing and we're getting that correct as well okay um so that's that's a very brief introduction to non-newtonian um we did all this under the the leadership of uh ian floyd who is um kind of the he's the the point guard or the quarterback or i would i guess i'd stay on the football field number six um that really distributes the ball for the uh for the post wildfire efforts in the corps and uh what we're going to do next is we're going to move to 2d sediment and i'm going to continue to move the slides but i'm going to mute myself and hand things over to dr sanchez so alex before we dive into that um just wanted to see on the upvoted questions if there's any that you helped to answer along the way that you wanted to highlight um just maybe cover a question or two right now before you dive into that um i i think there's one that's been upvoted the most um that we might want to just hit just before you dive into this and i'll just uh see that one i don't know if you've seen that one from helvisio here um and he's just asking he says does does hecara's 6-0 allow the modeling of both non-newtonian and newtonian fluids at the same time for example to consider the dilution effects of a dam break in the downstream water course um so anyway that's uh you know simultaneous uh non-newtonian and newtonian uh that that's one of the questions that got the most upvotes so i don't know if you're stanford one of the hit that before uh you dive back into your presentation sure i mean as as gary answered you have to choose right now in 6.0 you have to choose and not only do you have to choose either or but you can only run a single concentration the next step is we're very close to releasing a time series of concentration where you have a global concentration but a time series what's after that is we're going to allow you to define concentration boundary conditions we will route the sediment we'll have a very simple depositional model um and we'll actually change the concentration dynamically in the model to get the to get the sort of uh you know downstream um concentration attenuation that that you would see in one of these systems that latter one um will not be available in 6.0 look for it in like 6.1 sounds sounds good okay so um back to you alex all right thank you yeah my name is alex sanchez and today i'd like to talk to you about 2d simon transport and hc raz um this is something that we've been working on for several years and we're very excited to release it 2d sediment is going to be a beta feature within 6.0 which itself is in a beta release currently the first release was in mid-december and update number two was released at the beginning of this month um even though 2d settlement will be a beta feature it will have a full interface meaning that the user will be able to specify all of the necessary initial conditions boundary conditions and input parameters to run simulations and to visualize uh results um 2d sediment will be compatible wait uh can you go back 2d sediment will be compatible with 1d sediment in the sense that it will use a lot of the same formulations for example the sediment transport potential formulas and it will share a lot of the same user interface for input parameters and boundary conditions however there are some differences between 2d and 1d sentiment and braz and i'll try to point them out as i go through this presentation um you can run sediment transport with both the diffusion wave and shallow water equations in raz we have draft documentation that is available in pdf form when you download 6.0 we have a technical reference report a user's manual and the verification validation report is in progress that will be released uh along with the official release of uh 2d sediment transport and we're also moving uh to we're also migrating a lot of this documentation to uh confluence and online format which will be easier to update as the as the as 2d sentiment evolves okay usually beta features i mean beta features or or models um or even first releases they they're pretty feature limited they have limited capabilities uh that's not the case with 2d7 sentiment it has a lot of features this is a list of some of them i'm not going to describe them all in detail but i'll just mention a few that are noteworthy for example supports mixed cohesiveness cohesive sediments variable grain and densities it has hindered settling flocculation consolidation vertical varying cohesive properties splash erosion morphologic acceleration and then what really makes it unique is the subgrid approach as uh current users of hcc raz will know hc ras has a very useful subgrid approach which allows you to use coarser grids and still get a very accurate result well the sediment model the 2d sediment model was designed for to be used in conjunction with the the hydraulic model and it also has the subgrid uh technology in it and i'll talk more about what that is and how it's done in the modern in the next slides so rather than um showing a lot of equations and boring people with all the details i'd like to just talk very uh generally about the approaches and formulations utilized in the sentiment model people that are interested in the details can go look at the technical reference document or if you have questions i can i can try to clarify them at the end of the presentation if there's time um just like in 1d the non-uniform sentiments are discretized into grain classes with specific properties such as size and density sediment transport is simulated with a 2d unsteady advection diffusion equation for total load concentrations per grain class there can be any number of grain classes a total load formulation is utilized instead of solving for separate bed and and suspended loads for efficiency because basically you have to solve half as many transport equations however it does make an estimate of the fraction of bet and suspended loads and it utilizes that in various calculations such as the diffusion coefficient for example um erosion and the position are computed for non-cohesives or computed with an adaptation approach the main advantage of this approach is that it allows us to reutilize the same transport potential formulas as 1d sediment transport and so it makes 2d7 transport consistent with 1d and there are other reasons why as well but that's the main one erosion of cohesives is treated with the same piecewise linear exoshear formulation as windy sediment the position of cohesives is slightly different in 2d sediment it utilizes a concentration-based flocculation formula formulation for silts and clays so particles are allowed to form flocks if the concentrations are high enough and when concentrations get too high then you start to have here settle line and it also does a temperature correction for the settling velocity of these flocks the user can specify bed layers this is a new feature in 6-0 for 2d sediment and then 2d sediment discretizes those vertical layers while preserving the boundaries that the user specified for those layers into computational bed layers and it uses a bed sorting and layering model to compute the the grain classes the fractions of each grain class and the layer thicknesses of every bed layer that's what the bed sorting layering model does it also uses an active layer approach in which this is basically that the top most layer is referred to as the active layer or the mixing layer and it's the layer that exchanges material with the transport 2d sediment can have any number of vertical layers but it's specified at the beginning of the simulation and held a constant as far as the numerical methods go the advection diffusion equation is solved with finite volume methods it uses an implicit journalized euler scheme in time and it has several addiction schemes from uh simple linear different schemes to less diffusive non-linear flux limiter schemes um anisotropic diffusion is simulated with a linear two-point flux approximation and it has because then it's implicit in solved matrix there's several options for matrix solvers in it um flow and sediment are semi-coupled meaning that they're they're coupled at the time step level but flow is computed first and then sediment and then it does it assaults continuity again and corrects the continuity equation and flow for the bed change basically um the bed sorting and layering equations are solved with finite differences um and there are many uh methods for the subgrid uh methods and algorithms for the subgrade calculations uh really too many to describe here but um you could people can look into the technical reference document for those in order to explain how the subgrid calculations work for 2d segment i'd like to briefly review how subgrid works in the flow model so on the left you have a computational style that is partially wet and dry and the terrain is bend into discrete elevations which can be used to create a curve of horizontal area as a functional elevation so basically you have terraces of constant elevation each bend represents the area corresponding to specific elevation and can be used to compute the wetted area for a given stage right this area elevation curve can then be integrated to compute a piecewise linear volume elevation curve and these are the hydraulic curves that the flow model uses during the simulation for a cell to go back and forth between stage wetted area and volume spaces are treated similarly and the model computes other variables at faces like hydraulic radius conveyance vertical area and other variables okay 2d uses 2d sediment transport uses the same computational mesh as the flow model for the transport equation so basically solving for sediment concentrations but it computes bed change sorting layering and some hydrodynamics at a subgrid level this plot the the red curve shows the area elevation curve that's used for hydraulics and those curves can be really high resolution there can be more than 100 area bins so those curves are really too high resolution to be used for sediment transport so the approach that is adopted in 2d sediment is that those bins are grouped so that you have a coarser set of curves for sediment and then sediment the bed processes are computed on those coarse curves and bed elevations are interpolated or updated [Music] from sediment to the flow curves so now i'd like to show some um verification validation data sets um these two simple analytical data sets are of uniform flow with different upstream boundary conditions uh stepwise and boxcar functions um i'm not going to go into the details of the datasets but it's just to show that we've done verification datasets to analyze the performance of the different advection schemes and look at grid convergence and things like that a more interesting data set to look at is the next is on the next slide so these uh this data set is a a base and that has circular flow the water is just going around and around and then a concentration field is initialized at a point at an around the circle the constant concentration of one and because the flow is going around it should go back to its original position and if if if there's no numerical errors it should preserve its shape and stay monotonic and so by using different advection schemes we can test whether how numerically numerically diffusive they are and whether the the non-linear schemes are are are monotonic and are we conserving also the con you know mass are we conserving concentrations um and the the upper animation shows a very diffusive upwind scheme by the time the concentration goes around it's been diffused it's barely recognizable and the one in the bottom shows a high resolution scheme in this case it's a super b uh scheme and that one is very good at preserving the shape of the concentration field as it goes around all right next slide this is a classic uh laboratory data set of channel and filling and migration um done in in the 80s it used fine sand the flow was going basically from left to right the black line represents the initial bed elevations and then the the dots the circles and diamonds are the measured elevations at two different times and you can see that the computed results follow very closely the measurements this next slide shows an animation of that same data set if you go ahead and turn it on the vertical layers are the bed layers and the colors represent the median grain size within each uh layer i believe there were three to five green classes in the simulation but you can see as that as the channel infills and migrates you it deposits uh you have finer sediments within the channel and the places that are eroding tend to armor and tend to have higher medium grain sizes so right there there's a snapshot of the measuring computed bed profiles it's going to end at the final time step and show the final bed elevations in just a second so this gives you an idea of how the bed layering algorithm works how it splits layers and how it merges layers as you have erosion into position okay this is a very interesting data set of a settling column done by tormund and berlin they um mixed a slurry of um with a it was three percent by uh by volume it had 3.5 percent sand they mix it in the settling column two meters high and then let it settle and consolidate and the plots on the bottom show the evolution of the dry bulk density over time and over the first day it settled rather quickly and consolidated rather quickly but then it took uh you know almost two weeks to really uh uh consolidate much uh it consolidated much more slowly afterwards um the kind of the tail at the bottom of those dry bulk density profiles is due to the sand that settled very quickly at the beginning of the experiment and that's why you have that little tail the way uh consolidation is is specified in uh in the interface is through a consolidation curve which is what you see in the upper right hand corner of the slide the user specifies dry bulk densities as a function of time and this is the curve that the model uses to determine the consolidation rates during the simulation this is a lab experiment that is useful for looking at model performance over non-erotical surfaces and also avalanching so it had it was a rectangular phloem with a very strong clear water jet coming in from the left hand side which produced this really beautiful recirculation pattern within the rectangular flume and it scoured all the way to the concrete bottom and produced the sound wave that started to migrate forward the sides of that scour hole were very steep and it and it was uh sliding it was avalanching inside the uh the scour hole um and that plot to the lower left you see a comparison of the computed the solid lines uh solid and dashed compared to the measurements the the diamonds and squares you can see that the model performance was was really good compared to the measurements this is a field data set and the purpose of this data set is to demonstrate uh really the benefit and power of the subgrid approach within the 2d settlement model this is uh the west fargo data set which is a diversion channel 6.8 miles of a diversion channel there was an existing 1d sediment transport model that was set up for this this project and we wanted to see how the 2d sediment model would perform for this data set if we used just one cell wide for the whole channel you know we wanted to see if we could capture the varying bet change along the cross sections or for the different terrain elevations and how that compared to the measurements this uh the sediments consisted of fine silts and clays and the the figures that you see to the lower lower left is the computational mesh it's basically a a 1d model really because it only has one cell across and the plot to the right is a typical cross section of the channel which has a low flow channel um yeah all right next slide please all right uh you can start the animation so there's this animation has several plots that are moving dynamically the one to the upper left shows kind of like the area elevation curve that i talked about for cells but this is for faces so it's length elevation the horizontal blue line represents the the stage the water level at that face and you can see how it goes up and down and uh on the lower right you see the current velocity at average current velocity for that face and how it's changing over time um and then the uh the lower bottom two plots are the bed change for that uh length length elevation curve and also for the cross section the cross section is not actually stored in the 2d sediment model it only has this length elevation curve but it can back out where the bed change occurs for the cross section because um it knows um though that's a post-processing thing how it does that it just knows that for certain elevations have certain amount of bad change and then if you if you have the original cross section you can you can back out where the bed change occurred along that cross section so that's what this animation is showing and what's important here is that the field measurements showed exactly this behavior that there was erosion at the lower elevations at the low flow channel lower part of the channel and the position in the higher parts of the channel when the water was high and we were able to reproduce this behavior with just one cell across the the channel this is a very neat um example all right can you go yeah so this is a little different uh example this is an experimental watershed and in arizona it's very small it's one acre it's it's an experimental watershed um and here i'm showing the uh the flow and the measured sediment loads at the outlet which is located at the lower right portion of that watershed and in the animation well it was a very strong event very short just a few minutes but it had intensities of like six uh inches per hour um and as the storm went by you see the high concentrations initially and as the water recedes the concentrations go down it kind of looks like a heart beating and so this data set includes sheet and splash erosion obviously because it uh there's precipitation that's that's a very important contribution to the erosion all right so uh can you move to the next slide okay now going to the other uh end of the of this scale for for elaborate for uh validation data sets this is the upper mississippi river we modeled 18 miles of the upper mississippi river and model two years of bed change for this data set it's a portion of the river that is has a lot of training structures there's chevrons and spur dikes and many other structures so it's quite challenging and and there's a lot of different uh bed roughnesses going on it's a large mesh but um we did pretty good and the the calibration i don't i have the flow calibrations here but i'll show you the bed bed change results on the next slide please so on the left those are the initial bed elevations and it's really cool to see the the bed forms and the multi-beam survey they're beautiful and then when you compare the measured and computed bed change uh obviously the the measured ones show that the bed forms in them and we cannot reproduce those because of the scale that they're they're at compared to the computational mesh um but we generally do a good job of reproducing the areas of erosion and the areas of deposition and the magnitudes are pretty good as well this is a two-year simulation um yeah so if there's still time i'd like to talk a little bit about the future work so obviously we still have work to do in terms of verification and validation that's uh report is in progress uh we'd like to be able to couple 2d settlement with 1d that's something we can't do yet um we like to do more visualization and erasmapper we can't visualize any subgrid uh results you know it would be nice to see how the the hydraulic curves are changing over time um and all the sediment curves as well oops um it would also be very useful to be able to update the terrain based on the computed uh uh bed change you can't do that right now 1d sentiment has a really cool hot start feature we don't have that yet for 2d that's something that we would like to add for the next release we would like to be able to spatially interpolate initial bed properties that we don't have that right now subgrid methods and algorithms the way it works right now right now is that the user has to choose which methods are going to be used and they're used everywhere all the time um we'd like to for it to be smarter and be able to adapt depending on the flow conditions and resolution and the sediment properties and decide for itself what methods are most appropriate um and there's there's a lot of work to do on that um it doesn't have secondary or helical flow effects we like to add dredging tools we'd like to transfer some of these 2d 2d sediment algorithms and features to 1d settlement so that they can be more consistent with each other and there's a lot of work to do with speed improvements right now it's parallelized with openmp but i would say just just barely there's a lot of work to to do that right now um as like a rule of thumb uh uh running flow with sediment will take approximately twice as long as only doing flow although that that varies a lot because the sediment computations are approximately they probably scale approximately linear with the number of grain classes so if you have three green classes or 12 it makes a big difference in the computational times okay so to give you a quick taste on what the interface looks looks like and how things work the way and um if you're familiar with wendy's sediment the bed gradations are specified at cross sections a bit for 2d we need them at faces and cells so the way this is done is with a classification layer in brazmapper and that can be created with a raster and then having override regions like when you do like the way you do setting sentiment mendings and layers or you can just specify polygons like i've done here where i've created a channel region and island and banks and once you have those sediment bed layers you can go to the sediment data editor and there's a new tab called 2d sediment gradations and there you can associate the classifications with actual bed gradations or bed layer groups so that that's how the initial degradations and bed layers are specified in the interface as i said previously a lot of the interfaces shared between 1d and 2d sediment you know the way you specified the bed gradations is the same the user defined green classes all that's the same but there is some things that are different for example there's this menu called 2d options which contains all the options that are only applicable to 2d like sheet and splash erosion the morphologic acceleration factor a base beds look coefficient hindered settling and that's where those things are specified as far as the computational options just like for hydraulics those are under the unsteady flow analysis editor you can select the menu sediment computation options and tolerances it takes you to this editor that everybody's already familiar with that has one done 1d but now there's this new tab called 2d computation options and it has all the options for transport like affection scheme iterative solver implicit waiting factor the options for the outer loop convergence parameters computational sediment layer parameters so this is like the resolution for the vertical layers number of vertical layers computational bed layers and as i mentioned before there's different algorithms or or methods for the subgrid erosion and deposition that's and this is those are specified here and here you can also control the spatial resolution of the subgrid calculations within sediment by specifying a maximum number of subgrid regions and also a length scale so that in areas where your mesh is very high resolution you have less subgrid resolution and in areas where your cells are very large you have more sub subgrid resolution that's how that is controlled those are optional parameters and to give you a taste of the kind of output you can get from the sediment model on the left side there's a field computed concentration field on the right is a bed change and then you can also do time series plots from anywhere on your mesh just like you do for hydraulics you can also create transits and extract profiles along those transects that's the bottom right plot on the slide so where do you get it you can get it from our website if you go to the downloads section hcc raz is free the beta is out there for everyone to download and test it really benefits us for us when people download the beta versions and test them so please do that please give us feedback if you have any questions or or things you'd like to talk to stan and i please email us the next slide has um our email addresses so that's out i'll i'll take any questions unless we want to go to the next uh portion of the webinar um yeah why don't we just hit a couple of things real quick because some uh some people may need to leave i watch the numbers kind of drop off right on the hour so there are going to be some who need to leave and come back to the recording later on so welcome back if you have come back to youtube um so stanford is going to get into a couple more things on the slides but i do want to highlight if you want to do this yourself um we do have an on-demand course available which marty teaches on the 1d components we will be gradually getting this into a 2d course as well but if you want to get into the basics of it you can take that 1d course online if you want just the regular course in what uh in the basics for hecarim's 1d or 2d for fluid flow for water flow we have those on demand as well and we've got a special session coming up tomorrow for those who have taken the 1d and 2d courses or if you would like to just dive into it from scratch as well um a live session tomorrow three hour session that covers what's new in heck grass 6-0 and we will take the demo models that we build in our on-demand courses and highlight uh some of the new features uh the really cool new things so if you get a chance to sign up for that um whether or not you've taken the courses before um that'll be happening tomorrow so we invite you to do that um so before we dive into the next piece of uh stanford's presentation then um i just wanted to highlight one maybe one more question that's next up voted and see if either stanford or alex can answer that and then we'll bring all the other panelists on after to highlight some of the questions that they have answered everybody in the background has been very diligently answering questions on the text and we want to give you a chance to highlight those uh live as well but the one i wanted to ask and put out there deals with lateral uh erosion and i know we've you know if you go back and watch part 100 our 100th webinar we talk about how uh sediment particles move in relation to water and those sediment particles being big rip-rap like you see in the video behind me how does water move riprap and what would you have to do around your bridge pier or your bank erosion you know your bank protection bank stabilization uh to prevent that from moving you don't want that transported um so stanford and alex if you just wanted to comment briefly on you know when would you use a you know a mobile bed model like like you've been showing and when would you need to use something different to look at like localized uh bridge pier abutment sky you know abutment scour or pier scour um or sizing the riprap and maybe you could highlight even that riprap sizing tool that might be coming you know what are the limitations of this and when would you move into just like an equation based an empirical kind of uh equation based uh approach the issue is dimensionality any sort of bridge scour analysis is a three-dimensional and turbulent phenomena and so a one-dimensional or a two-dimensional model can't handle that and so um it doesn't matter if you're using mobile bed um if you're doing a localized bridge scour analysis you want to be using either like a complicated cfd model or some sort of empirical equation that accounts for the dimensionality in the turbulence now as far as riprap sizing you kind of you threw me a softball there um yeah we are actually adding a rip-rap sizing um tool into raz it won't be in 6-0 but it's actually um we have a working prototype it will be in six one um and so if you're interested in using kind of the cores methods to size rip-rap on bends or in straightaways that's going to be built that's going to be baked right into hc ras in 6.1 awesome awesome okay and so yeah definitely wanted to highlight that and if you do want to look at bridge scour and things like that locally we do have on-demand courses on that as well so i'm going to take literally three minutes before we go to q a all right so i just want to make sure we give marty and gary and chris and others a chance to highlight the questions that they've done so yeah let's go back to you stanford let's go to your slides and then we'll bring everybody on thanks okay just um you know alex showed you very briefly the uh the interface and how to set up a 2d model we actually have a i put together a very short video demo video of setting up a 2d model which i'm putting in the chat right now um so if you'd like to actually like just watch someone set one of these things up um that that that's probably the best way to get to that now i was going to talk about wendy's segment but i kind of knew we wouldn't have a lot of time um because the other stuff is just the new stuff that people want to hear about but there are like 35 new 1d sediment features including mainly you can view we have a new 1d sediment viewer yes it's a new new sediment viewer for those that that's an inside joke for the razz users for the red sediment users but you can do things like multiple plans and multiple variables multiple accesses it's just a much more robust viewer every update we're actually making this better um and so uh for those of you who have been um maybe underwhelmed with the experience of viewing wendy's segment we've uh we've done a lot of we're doing a lot better now the other thing is that we can do quasi and study um we can you can define quasi and steady dss data instead of just putting in putting this data in and that might seem like a minor feature except what that does is that allows you to use 1d sediment transport in the watershed analysis tool which is our stochastic engine at hcc which means that now you can do um sediment transport stochastically and so that's that's what i'm showing here um and this is from a paper that i'm showing at the bottom that i wrote with will layman and uh and dan perdal and the uh what we've done here is we've just run a sediment transport model with bootstrap hydrology 300 times and then developed um statistics for the bed change um from this sort of stochastic approach now um the last thing that i want to point out is that we have gone to online documentation which is wonderful because you know we're developing a lot of videos there's a lot of videos on our website um and uh we're just embedding them right in the documentation we're updating the documentation as we update the model um it's excellent um and so i'm just gonna put the uh the hec ras documentation link in the chat and um you know the the 1d sediment transport is already up it's complete 2d7 transport we're working on right now that'll be up pretty shortly but then the other thing that i want to that i want to show you is that i have added a page that's just new 1d segment features in htc ras and this is the other 33 slides that i didn't show you and so i'm going to just put that up in the chat as well and if you just want to kind of work through that powerpoint and see all the other cool 1d functionality that's in there um i that i just think that would be awesome um so uh craig with that why don't i wrap up and turn it over to our panel all right sounds good um as everybody's uh coming on chris if you want to turn your video gary come back on marty uh as well uh be thinking about the questions that you've answered that we would want to highlight uh for the youtube recording my my number one question um maybe maybe it goes back to the government of the u.s why are we measuring sediment grain size in millimeters and why do you and you're putting depths in feet um no that's uh that's more of a universal question no that's actually it's actually a really good question because also we've run into that with the non-newtonian parameters we are not going to make you put non-newtonian parameters in in dines per whatever um and so the non-newtonian parameters are all an si2 basically what it comes down to is if you if you are unfortunate enough to have to model sediment and u.s customary units i am sorry i feel your pain um we are not going to make you translate grain sizes into feet that's not something we're going to do if you're if you're a u.s customary imperial you know it's purest you can just swallow that no that's that sounds good um the one thing i just wanted to highlight um we'll get into the questions is that some of the figures you showed with the increasing availability of lidar data and just you know the ability to just go out there and pretty much pick up every little cobble now uh you know as long as you've got a dry system afterwards and see what happened when you talk about what we're doing in the future calibrating to those you know some of these uh before and after events um lidar data sets is that something that you see um you know being the wave of a wave of the future in sediment transport calibration alex yeah for sure yeah yeah for sure and it brings up an important point that uh when we're comparing models such as such as ras to those measurements we're not going to pick up a lot of those bet forms right and so they they should be removed from the the initial bed elevations to have more fair comparisons of morphology change cool all right so with that um i'll stop my video there in the background um so it stops moving around and let's just introduce everybody else who's on the call here uh chris uh the author of the uh rad solution gary if you haven't watched gary's presentation and cam ackerman's presentation from a couple of weeks back it is number 102 in our video library so uh definitely check that out um let's start let's see we'll go from marty then gary and then chris that's how it shows up on my screen if you can uh over to you first marty the questions that you've answered um highlight a few of those and um and and let people know uh what what the answers are and where they might turn for more resources sure yeah there are a couple really good ones a couple of them had to deal with river meandering we've talked about and can we model that and i think it kind of goes back to what stanford was saying also you know it's it's a three-dimensional phenomena and trying to reproduce that in 1d or even 2d if you're doing you know depth average assumptions you're not going to get an accurate result so um you know i would there's other programs i mentioned in my response that you know purport to do that but some of them are are not 3d but there's a lot of assumptions that are kind of built into that so you have to be a little bit careful i think in how you use those um one one question i do want to point out that was asked by mike from new zealand was you know uh predicting bank erosion you know without riprap there but if you had uh uh banks or even parts of the upper bank that are dry but maybe uh the side of your channel might be eroded you know can can you use uh rouse for that and uh i might put that for stanford in terms of 2d but i would my response would be no i would probably look at using something like the b-stem capabilities which people may or may not be familiar with it's intended to work with 1d but you can have different layers of different materials in the bank and it uses computed shear stresses and some other equations to figure out if the banks eroding and then that material goes into the channel and it's available to be transported using the uh equation so that's a nice feature and uh stanford you think thank me later for the plug because we were talking you know it's a lot of work went into that uh putting that into raz and it's maybe not being you know used as much as it should be so i give a plug to that also i think i'll i'll pass there on to other folks yeah let's go on to gary team leader for the raz uh the razz team anything that you've highlighted um on on the uh the questions that have come in that you did you wanted to mention sure there there were several things that well there were quite a few questions so that's great first of all um there was one question about can hc ras do both bed load and suspended load and the answer is yes but you have to pick an appropriate sediment transport equation that is a total load equation okay um so that was one there was a question about bulking and concentration that i answered but i wasn't 100 sure my answer so i'm going to make sure it was right but we'll get stamford to come in but the person asked if you um add in bulking of the flow from mud and debris flow does that interact with the concentration and i said no it changes the volume but the concentration has to stay currently in 6-0 a single value which will be changing that in the future though so i'm hoping that was the right answer is that the right answer yeah that's right and this this can be pretty confusing alex alex actually pushed me a couple weeks ago to change this interface because it can be confusing you have two options in raz uh in in non-newtonian you could say hey i know what the water flow is and then i'm going to give it a concentration and then i want you to bulk the volume with the solids or you can say hey i know what the complete um fluid mixture is so don't bulk it but i'll still give you the concentration to compute the other non-newtonian parameters so if you specify the complete fluid volume then you still com you still specify the concentration but it's just used elsewhere if you only specify the water flow then you turn bulking on give it the concentration and we'll increase the volume all right um any others you want to highlight gary before we turn over yeah there was one last question that didn't get answered yet so since it was just a recent question i thought i would save it for either alex or stanford but to let them answer but um the question was does the 2d sediment transport allow you to model river delta growth yes i would think that that's within the capabilities of the model certainly so okay well that's uh that's something that you know and and as we watch these things you know you get these satellite imagery you know going back you can again see how this happens and you can determine you know are we doing a good job and you can probably calibrate that as well um chris as we turn over to you and then we'll just open it up for general discussion um maybe if you could highlight uh chris the resources available through the raz solution in particular what i think has been kind of under utilized lately is the forum um because it's kind of got a different format and i see a lot of questions coming up on the facebook pages and the linkedin pages that probably could go into the forum as well and get a more detailed uh response there so maybe start with that um as far as turning people over to some of the resources that are available and then dive into some of the questions that you might have answered yeah well thanks for that plug craig um so for those of you who don't know the rad solution is a uh online help site for heck razz uh there are a lot of articles up there there is a forum as craig mentioned and i'll apologize for those of you who are on it a lot about the spam i'm constantly battling to get off of there it's almost a daily battle to keep the spam off of there so sorry about that we're actually working on some improvements to hopefully make that happen a lot less going forward but yeah if you want to post questions there there are lots of readers um i think even some of the hec-ras team get on there occasionally and my suggestion would be when you post a question keep it brief and state your your point keep it brief write it well so it's easily understandable if you don't people are likely not to answer it so um try to keep it brief and easy to understand um so yeah thanks for that cray and um i've got a question though too um marty took mine the lateral um that was the big one i had on my mind so thanks for for um asking that marty but uh another question i had was um for output you said that you can't yet see changes in terrain your terrain in raz mapper but that's coming how would you suggest somebody who's getting into it right now doing some 2d sediment transport model what's the best way to look at your output to view it to understand what's happening in your simulation you can visualize right now uh cell average bed elevations and bed change but that you can't uh update the terrain based on the subgrid bed change um i i think what's there now is is sufficient for lots of applications if people are interested and and more of the more detail and want to look at the subgrid output they can mine the hdfi file for that output and the user manual describes where that data is and how to access that they can do that with matlab or python or some other scripting language at the moment but if you're interested in any sediment output on the cell scale you can just see it you can just view it in ras mapper and you can plot it as a transect it's only if you care about like the sub grid scale and what's going on there which i think most users won't most users want to like see the map at the grid scale and and that's enough um one things we will say is that you know there is so much output in sediment transport that um a lot of people do like to write code against the hdf5 files of course you know um you know chris has made a cottage industry of this um but the uh but there like there's a lot of folks that use either python or r or matlab to um essentially a lot you know a lot of the visualizations you saw alex put together were put together with matlab the ones that i put together were put together with r and you know if you go to the sediment user's manual we have some sample code in there we're also collecting sample code from other folks so if you write some sample code like i'll tag it onto the raz users manual we want to provide that resource so that anything that razz can't do yet you can do until we can do it thanks dan that's good that's great um a couple other resources i wanted to point out here um i'll just maybe share my screen real quick so you can see this um again on the australian water school webinars you'll see uh them numbered uh and you know up through i think we're at 103 or 104 now we did about a year ago um do a sediment transport modeling uh webinar in preparation for our 1d course that marty teaches um and so stanford and marty were both part of that one that we talked more about the basic principles and this was before the 2d was available even in the beta version but it is worth going back i think in watching that one to get some of the principles also um bmt and the wbm guys at bmt and 2flow have put together a webinar that they just had about a month ago which talks about some of the 2d and 3d uh morphological modeling and it's it's a bit different from what we're talking about uh here today but certainly an application like around ports and harbors and uh and and coastal uh coastal currents and things like that and and how it flows i absolutely recommend that one too um and and their whole series any any of the series that uh that that bill and chris and others have done uh with with uh through tooth flow um awesome awesome content there so i'll stop sharing that i just wanted to point out a few more resources that you might want to go back and watch and certainly join these courses if you want to do it yourself because watching it is one thing but getting hands-on with your actual model and clicking on the things instead of just seeing screenshots you know that's that's where you're going to learn it uh and and but your head up against the uh some of the the errors the instabilities um it doesn't always run smoothly i did see a comment there and marty you may want to uh respond on that one just in light of some of the things on the courses about instabilities and whether that's common to get instabilities in some of the uh sediment transport modeling and i i almost had to to laugh at that one being like oh man uh how many times if i wanted to throw my computer out the window go ahead many many fewer instances of stable models than unstable models right i think we've all uh anyone who's been modeling has run up especially you know unsteady flow modeling uh most of the time it's unstable right you got to work at it to get a nice model and get it stable uh so there's there's plenty of examples all over the place of unstable models it's uh it's the uh examples of the stable models that are hard to come by so we always like to look for those uh but as long as i have the the foreground i was gonna ask these guys um you know there's a lot of example data sets that that come with whereas to try out different things uh how about 2d settlement yes there is one example data set that has shipped in this uh beta release we're we're going to expand that and add more data sets to that yeah so be sure when you download the software um that you install the sample beta data sets um that come with it and if you've done it before you might want to do it again because there's new some new content that's that's come in there uh back to what you were talking about marty with the instabilities though i wanted to just uh throw the question out there with uh the the new features in in 6-0 that for example allow a uh the water to shut off and go dry which helps us out here in western australia quite a bit when i try to do long-term series i've had clients ask for a 10 000 year series of what happens at mineclosure and boy we i think a couple years ago i got in touch with all you guys trying to figure out how to do that um and you know we got it done but we had to break into many many pieces when you're doing that you can just sort out and take out all the dry periods um but you know in the past you weren't able to get your uh model to even go dry uh at all and so you had to put in some additional features to let that happen and so maybe i'll go to chris or gary on this one you know the new features that allow you to um you know the the new computational methods that allow you to go to zero flows um will that help sediment transport model stability as well um i'll give my uh two cents opinion on that but and then i'll let the sediment guys weigh in but um i've i've said this to stanford since you started working there you shouldn't even start modeling sediment transport until you have a good hydraulics model that's very stable otherwise you're just asking for trouble right because the sediment transport and various parameters are going to potentially lead to more instability so if you have a semi-stable hydraulics model that might put it over the edge okay so that's the first thing to work on the hydraulics model first make it a good stable model we have added a new 1d finite volume capability which is much more stable than our traditional 1d finite difference and so yes it does allow you to to go down to low flow or dry start drying and go to dry and it handles low flow way better and steep systems way better so that's going to be quite a quite a good feature for sediment transport also and maybe from there i'll let stanford weigh in a little bit on that because he has a lot of experience with this also yeah i mean i if you're if if you're just moving the water and your borderline stable what's going to happen when you move the bed like i'm not even embarrassed about that that's just a reality of physics right um and so you need to you need to make sure that your hydraulic model is stable um and calibrated right you need a because one of the things that happens with with sediment transport is things that are just kind of um that the hydraulic model can be very forgiving the hydraulic equations they're self-correcting they can be very forgiving you can have bad infected floor areas and bad um bad end values places you know ugly cross sections and uh brass can just blow through it and give you a pretty good result sometimes um but then when you add sediment it it uncovers a whole range of errors um because sediments can be much more sensitive to some of those things and so you just need to make sure you have a high quality sediment transport model or a high quality hydraulic model that's calibrated before you add sediment um it it's you know because we have it's kind of easy to add sediment um a lot of people will just add the water then add the sediment and run it all together and that's definitely not the way you should approach this yeah i'll take that one step further that um you also need to you need a you need to have also a very good quality terrain often with hydraulic models that um you can get you can get by with a poor or like pretty bad terrain model but once you're doing sediment and and and bed change you need a really good um terrain model otherwise the results are going to be terrible that's yeah you can't stress that enough and when you go back to some of these original methods and you know when you had a few cross-sections you know hundreds of meters apart and and that's all you had uh to define everything that you're working with as a baseline you know you can you can kind of see where the accuracy is heading and just make sure that we're keeping up with that if you go back into that um you'll you'll see on that sediment transport modeling webinar that we did number 71 we start out with the potentially appropriate story of albert einstein whose son helped found the department have went to school at berkeley um and he was a sediment sedimentation engineer and the story goes that uh einstein was help helping him guide his helping his son uh and guide him toward what field to go into and he said no no don't go into sediment engineering it's it's it's too much too difficult you know this from the guy who brought us uh you know relativity um but just because of those feedback loops and the things that stanford was talking about i mean if you're already looking at instabilities um you know then you tend to start fishtailing um with some of these uh once you introduce sediment because of those feedback loops that come into it um so yeah we'll just open it back up again we've got about you know five minutes to go here 30 questions that have been answered at least or more um so we're not gonna obviously have time to hit them all but maybe just go back to like one one each uh to each of you and then we'll do some closing comments so uh maybe over to you chris um grab it grab a question that you might have answered um or have for the uh panelists and then we'll we'll pass it along i've got one for you stanford where is that usgs demonstration flume and how can i get there to go watch some of that that's awesome yeah so we visited a little while ago so it's in it's in northern washington um and it's uh you know they the uh professor iverson just retired so i'm not sure it's gonna be active um i've never actually seen it in in action i just climbed it um just to kind of be there like i don't know like like a pilgrimage or something but uh it was um it's a it's a it is that's cool work with the work they did is has it it's the best it's the best work out there that's awesome uh and by the way i didn't mention that chris you've told me before that you don't think consider yourself a sediment guy but i've seen your name in the manual and it has to do with the sediment functions right oh boy that was a long time ago and i just barely scratched the surface um yeah so stanford and alex have got that area figured out all right sounds good um marty uh any questions that uh any la you know maybe pick one of the questions and uh yeah yeah there was one there was a question about using raz for coastal sediment modeling and uh gary and i both both responded to that i mean raz is not intended as a coastal model uh my response was well you could try but you you know all the transport equations were all developed from riverine conditions not not coastal and you'd have to have the appropriate boundary conditions uh wherever you're going to model which would be complicated um and you know you'd have a lot of things that really brows was not intended for you know not not that we've never used the model for things for which it was never intended uh right chris lots of packs yes but you know you're extrapolating so that that's always you know dangerous territory so check out twoflow's webinar there where they get into the 3d bits of it as well because um that's an awesome presentation uh gary any uh any last question that you wanted to highlight that's been answered good question for mud and debris flow on the question was have you performed 2d sediment mudflow models with bridge structures or a culvert conveyance any validation models of papers and the answer is yes the santa barbara models had lots of culvert crossings that got mostly blocked and had to do some unique solutions to solve that with with putting in blockages for the culverts so that's an area though that we need further work and research and development in though to do a better job and maybe i'll let stanford comment a little bit on that yeah and let's let's just do it i think we're just about out of time so you'll do we'll do closing comments stanford just uh you know maybe answer that question uh give your closing comments um alex we'll have you do uh the the final send off and and then we'll we'll stop the recording here so stanford over to you yeah culverts are culverts and crossings are the biggest complexity and the biggest uncertainty you know we made a blanket assumption that the culverts were all 66 blocked or something like that and um you know it was this wasn't something we used to calibrate it was an a priority assumption but uh it's that it's that's complex and that's something we're act that the team is actively working on cool well i just want to say thank you for giving us the opportunity to talk and i appreciate it i i enjoyed it and that's it excellent do take the polls so everybody all you uh for all the attendees out there and those watching this on youtube take the poll let us know what you want to hear about in the future we'll cater that content around you um and around your needs sign up for these courses um in this online world now you know where we don't really go to conference as much anymore probably this year take advantage of the resources that are out there and take advantage of the hands-on things not just these presentations but let's dive into the models and so we're excited to offer you some of this content that you'll see on your screen take the polls take these uh you know send us your feedback we look forward to interacting with you again on future webinars thanks everybody for attending uh we'll sign off now thank you bye-bye thanks for watching subscribe by clicking the link below and click on the notification bell to stay up to date with new releases for the latest in significant innovative and critical advances in water science technology and management subscribe now to build your skills enhance your technical knowledge and learn from leading experts in water visit the australianwaterschool.com.au and discover our online training courses both live and on demand you
Info
Channel: Australian Water School
Views: 10,850
Rating: undefined out of 5
Keywords: USACE, Hydrologic Engineering Center, HEC-RAS, Krey Price, Stanford Gibson, HEC-RAS 6.0, Alex Sánchez, sediment transport modelling, 2D sediment, mud and debris flow, newtonian, non-newtonian, geologic flows, flood modelling, USGS, debris flume, advection, scour, morphological
Id: k08c9bgKT-4
Channel Id: undefined
Length: 85min 26sec (5126 seconds)
Published: Wed Feb 10 2021
Related Videos
Note
Please note that this website is currently a work in progress! Lots of interesting data and statistics to come.