/ News & Press / Video / Dynamic Simulation of Mining, Mineral Processing and Extractive Metallurgical Plants
Dynamic Simulation of Mining, Mineral Processing and Extractive Metallurgical Plants
WEBVTT Kind: captions Language: en
00:00:01.879 are you an engineer a scientists or a 00:00:05.87000:00:05.880 manager finding it hard to make 00:00:08.78000:00:08.790 decisions because your process flow 00:00:11.18000:00:11.190 sheet may be technically quite complex 00:00:14.23900:00:14.249 and the financial impact of your 00:00:17.12000:00:17.130 decisions may be significant now that's 00:00:20.75000:00:20.760 true 00:00:21.41000:00:21.420 you are not unique what it actually 00:00:23.51000:00:23.520 means is that you are one of those 00:00:26.33000:00:26.340 people working out in the process 00:00:28.25000:00:28.260 engineering field that's relied upon to 00:00:31.25000:00:31.260 make important decisions because of the 00:00:33.82900:00:33.839 skills you have so high they my name is 00:00:37.57900:00:37.589 johan having myself worked in the mining 00:00:40.97000:00:40.980 industry for over 20 years i know that 00:00:44.47900:00:44.489 these technical challenges are 00:00:47.02000:00:47.030 appreciated across all the different 00:00:49.13000:00:49.140 sectors obviously none more so than the 00:00:53.20900:00:53.219 operations themselves but then also the 00:00:56.27000:00:56.280 process engineering companies the R&D 00:00:59.05900:00:59.069 companies and the equipment suppliers 00:01:01.04000:01:01.050 and manufacturers now all of us in order 00:01:06.83000:01:06.840 to survive we have to innovate but 00:01:10.01000:01:10.020 innovation is hard and quite risky so i 00:01:13.99900:01:14.009 want to tell you more about how we use 00:01:16.94000:01:16.950 computer modelling and simulation to 00:01:19.85000:01:19.860 innovate so for the rest of this video 00:01:23.66000:01:23.670 let's work from the screen so why is it 00:01:29.14900:01:29.159 so difficult to connect the dots in 00:01:31.39900:01:31.409 industrial processes now that's because 00:01:33.74000:01:33.750 the physics and the chemistry are 00:01:35.33000:01:35.340 complex even for processes that may be 00:01:38.14900:01:38.159 perceived as being simple we as process 00:01:41.87000:01:41.880 engineers impose restrictions on mass 00:01:44.84000:01:44.850 and energy flow through pipes valves 00:01:48.55000:01:48.560 reactors and so forth because we 00:01:51.52900:01:51.539 ultimately want to maximize yield and 00:01:53.77900:01:53.789 minimize costs before the physical world 00:01:58.34000:01:58.350 of friction although we cannot exist 00:02:00.52900:02:00.539 without it is also our Achilles heel in 00:02:03.88900:02:03.899 order to solve problems and innovate we 00:02:07.21900:02:07.229 first have to mathematically acknowledge 00:02:10.21900:02:10.229 that our processes and equipment operate 00:02:13.33900:02:13.349 in a physical domain here's an arbitrary 00:02:18.97900:02:18.989 and quite simple comminution flow sheet 00:02:21.38000:02:21.390 model labeled for the purposes of this 00:02:23.50900:02:23.519 video so let's press the play button you 00:02:29.63000:02:29.640 can see the time starting to tick at the 00:02:31.43000:02:31.440 bottom here that's in second everything 00:02:33.71000:02:33.720 is our units we could look at the slurry 00:02:37.52000:02:37.530 flow rate delivered by the pump and we 00:02:39.61900:02:39.629 could for example look at the pressure 00:02:42.16000:02:42.170 created at the inlet of the hydrocyclone 00:02:44.93000:02:44.940 so as the controller tells the pump to 00:02:47.11900:02:47.129 switch on and other words slurry starts 00:02:49.49000:02:49.500 to flow you can see how the dynamic 00:02:52.19000:02:52.200 pressure increases and at the heart of 00:02:55.13000:02:55.140 shotgun and that's obviously because of 00:02:56.99000:02:57.000 friction because of the internal 00:02:58.61000:02:58.620 workings of the hydro cyclin so this bad 00:03:02.27000:03:02.280 pressure created by the hydro circlin in 00:03:04.52000:03:04.530 turn influences the efficiency of the 00:03:07.16000:03:07.170 pump so this is just a very simple 00:03:09.92000:03:09.930 example to show our physical modeling 00:03:14.17000:03:14.180 allows us the capture they're highly 00:03:17.03000:03:17.040 integrated behavioral characteristics of 00:03:19.94000:03:19.950 the process simultaneously as you would 00:03:24.11000:03:24.120 have noticed from the simulation I just 00:03:26.36000:03:26.370 showed you and as anyone being involved 00:03:28.97000:03:28.980 with plant operations with no 00:03:30.55900:03:30.569 steady-state operation is really a 00:03:33.05000:03:33.060 theoretical concept in a real world 00:03:35.96000:03:35.970 everything is dynamic and this becomes 00:03:38.87000:03:38.880 more true as we zoom in from a 00:03:41.24000:03:41.250 macroscopic to a microscopic scale and 00:03:44.15000:03:44.160 as we shorten our sampling timeframes so 00:03:47.87000:03:47.880 there's a better way to design and 00:03:49.58000:03:49.590 economically optimize our processes and 00:03:52.09900:03:52.109 equipment then treating material flow 00:03:55.00900:03:55.019 physical and chemical changes as steady 00:03:57.77000:03:57.780 state phenomena let's rerun the 00:04:01.90900:04:01.919 simulation now focusing on the small 00:04:04.56900:04:04.579 sample area here 00:04:06.81900:04:06.829 I've also increased the simulation time 00:04:10.25000:04:10.260 to 43,200 seconds which is 12 hours now 00:04:16.03900:04:16.049 firstly one cannot simulate flashy 00:04:19.37000:04:19.380 dynamics if you haven't also 00:04:21.72000:04:21.730 included the dynamics of the various 00:04:23.76000:04:23.770 controllers so in this very simple 00:04:27.42000:04:27.430 example we've got two controllers this 00:04:30.06000:04:30.070 controller here tries to control the 00:04:33.15000:04:33.160 label in a sample by changing the pump 00:04:36.51000:04:36.520 speed and this controller here tries to 00:04:39.51000:04:39.520 control the popping city in the feed to 00:04:42.81000:04:42.820 the cyclin by changing the feed make up 00:04:47.64000:04:47.650 water valve position let's press play 00:04:52.62000:04:52.630 again opening the scope here which 00:04:57.99000:04:58.000 indicates the level in the sump in 00:05:00.12000:05:00.130 meters over time you'll notice that the 00:05:04.38000:05:04.390 level increases rapidly during this 00:05:07.47000:05:07.480 initial simulation period so the 00:05:10.95000:05:10.960 conditions are highly transient as one 00:05:13.62000:05:13.630 would expect during the startup now you 00:05:19.20000:05:19.210 will also notice that control is not 00:05:22.17000:05:22.180 particularly good in this example for 00:05:24.57000:05:24.580 two reasons firstly our sump is grossly 00:05:29.04000:05:29.050 oversized compared to the rest of the 00:05:31.11000:05:31.120 circuit and secondly our controller 00:05:33.87000:05:33.880 parameters are and optimized but let's 00:05:36.95000:05:36.960 leave the simulation to run up and 00:05:41.79000:05:41.800 because the point I want to make here is 00:05:43.62000:05:43.630 that we approach steady-state operation 00:05:47.63000:05:47.640 over done so you can see that 00:05:51.62000:05:51.630 steady-state circuit behavior is really 00:05:56.01000:05:56.020 a simplified case of the dynamic system 00:05:59.46000:05:59.470 so if steady-state behavior is required 00:06:02.34000:06:02.350 the feed flow rates are kept constant 00:06:05.04000:06:05.050 and the simulation is run until the 00:06:07.41000:06:07.420 outputs stabilize as we've just 00:06:11.01000:06:11.020 Illustrated so if bats or semi-batch 00:06:14.34000:06:14.350 behavior is required the relevant feed 00:06:17.16000:06:17.170 streams are closed and the unit block 00:06:19.80000:06:19.810 batch responses are recorded so you can 00:06:25.65000:06:25.660 see that by doing dynamic modeling and 00:06:28.17000:06:28.180 simulation we don't sacrifice any 00:06:31.62000:06:31.630 information we only gain information as 00:06:34.71000:06:34.720 compared to steady-state and modeling 00:06:37.86000:06:37.870 and simulation now thirdly most of us 00:06:42.90000:06:42.910 have built some sort of model to help us 00:06:45.66000:06:45.670 solve our problems 00:06:47.18000:06:47.190 now irrespective of the package you 00:06:49.95000:06:49.960 prefer to use the following holds true 00:06:52.41000:06:52.420 the more detail you add to your model 00:06:55.23000:06:55.240 the more data you require to back up 00:06:57.93000:06:57.940 your assumptions and calibrate your 00:06:59.43000:06:59.440 model the reality though is that 00:07:02.01000:07:02.020 experimental data is often quite 00:07:04.98000:07:04.990 inaccurate and careful control of our 00:07:07.92000:07:07.930 processes may not be possible so what we 00:07:11.31000:07:11.320 do to mitigate our risks we conduct more 00:07:14.07000:07:14.080 testing and embark on more expensive 00:07:16.38000:07:16.390 pilot plant campaigns the problem is 00:07:19.98000:07:19.990 that without the appropriate physical 00:07:21.90000:07:21.910 dynamic tools we cannot really 00:07:25.08000:07:25.090 capitalize on our expensive data now for 00:07:30.24000:07:30.250 various reasons we use the mathworks 00:07:32.52000:07:32.530 suite of products to build our models 00:07:36.06000:07:36.070 the specifics of which I shall elaborate 00:07:38.70000:07:38.710 on in upcoming videos an important 00:07:42.18000:07:42.190 feature of this platform is its powerful 00:07:45.27000:07:45.280 ability to regress our models to the 00:07:48.45000:07:48.460 experimental and real plant data now 00:07:52.11000:07:52.120 it's not necessary to rerun our 00:07:53.91000:07:53.920 simulation we could simply open the 00:07:56.73000:07:56.740 results Explorer and all the data 00:07:59.79000:07:59.800 generated by the simulation is available 00:08:02.79000:08:02.800 at the click of a button so by the way 00:08:05.10000:08:05.110 we could also easily export all of this 00:08:08.52000:08:08.530 data to for example Microsoft Excel if 00:08:12.90000:08:12.910 you prefer to work in that environment 00:08:15.18000:08:15.190 to generate your reports so as an 00:08:19.20000:08:19.210 example we could select the ball mall 00:08:21.75000:08:21.760 block and we can scroll down to the 00:08:24.96000:08:24.970 variable you're interested in in this 00:08:28.29000:08:28.300 case we're looking at the mass 00:08:31.80000:08:31.810 distribution of each size class in a 00:08:35.37000:08:35.380 mall over time in this case we're 00:08:38.45900:08:38.469 looking at the mineral pirate another 00:08:42.81000:08:42.820 example say we 00:08:44.87000:08:44.880 are interested in what the pressure and 00:08:47.45000:08:47.460 the temperature did at night a which is 00:08:50.15000:08:50.160 the inflow to the hottest I plan we 00:08:53.18000:08:53.190 could simply find the cyclone block and 00:08:56.86000:08:56.870 then click on node a and there you can 00:09:01.34000:09:01.350 see the pressure and the temperature 00:09:03.43000:09:03.440 variation at the inflow obviously round 00:09:08.15000:09:08.160 about 2400 seconds into the simulation 00:09:12.22000:09:12.230 the pump switched on and succulent 00:09:17.30000:09:17.310 started to operate how it was supposed 00:09:21.14000:09:21.150 to operate so there are also very nice 00:09:25.87000:09:25.880 visualization tools available year I've 00:09:29.87000:09:29.880 created the shortcut key that you 00:09:34.10000:09:34.110 standard MATLAB plot functions to be the 00:09:39.65000:09:39.660 data so in the left hand side we got the 00:09:44.09000:09:44.100 mass distribution over time and particle 00:09:47.78000:09:47.790 size when a long thin scale and we can 00:09:51.56000:09:51.570 compare what that distribution would 00:09:54.32000:09:54.330 lock to the overflow from the cyclone 00:09:58.21000:09:58.220 and you can see at 2400 seconds we that 00:10:03.88000:10:03.890 sucker and feed pump switched on we 00:10:06.95000:10:06.960 started to get rid of all the pore sizes 00:10:10.01000:10:10.020 in a cyclone overflow another example 00:10:14.72000:10:14.730 here's a figure of the classification 00:10:19.40000:10:19.410 function of the cyclone again it's after 00:10:22.99000:10:23.000 2400 seconds the cyclone started to 00:10:26.42000:10:26.430 operate as it was designed to do you can 00:10:29.48000:10:29.490 see the the characteristic a shape of 00:10:31.97000:10:31.980 the classification function as reference 00:10:35.81000:10:35.820 cases you may want to have a look at 00:10:38.15000:10:38.160 what the aerospace and especially the 00:10:40.55000:10:40.560 automotive industries have done you 00:10:43.55000:10:43.560 would be astonished to see the 00:10:45.35000:10:45.360 difference that's physical dynamic 00:10:47.54000:10:47.550 modeling of individual components and in 00:10:50.60000:10:50.610 building simulations of integrated 00:10:52.88000:10:52.890 clutch gear brake systems etc have made 00:10:56.96000:10:56.970 today 00:10:57.87900:10:57.889 industry think about all the improved 00:11:00.42900:11:00.439 safety aspects reduced emissions and 00:11:03.69900:11:03.709 reduce development tons of new models so 00:11:08.10900:11:08.119 we as process engineers in the mining 00:11:10.23900:11:10.249 industry require visionary mindsets to 00:11:13.11900:11:13.129 understand our building physical dynamic 00:11:15.93900:11:15.949 models of processes will provide us with 00:11:18.78900:11:18.799 that competitive edge test data and 00:11:22.05900:11:22.069 experience are and always will be 00:11:24.06900:11:24.079 essential aspects however these new 00:11:27.40000:11:27.410 generation models provide the only real 00:11:30.40000:11:30.410 methodology to capture information 00:11:33.47900:11:33.489 because by including the physics and 00:11:35.91900:11:35.929 dynamics of material and energy flow the 00:11:39.12900:11:39.139 degrees of freedom of the system is 00:11:40.96000:11:40.970 reduced especially after the model has 00:11:44.13900:11:44.149 been calibrated to the real plant 00:11:46.38900:11:46.399 behavior we believe in building bespoke 00:11:49.98900:11:49.999 models our experience is that it is the 00:11:53.37900:11:53.389 only way that we can effectively deal 00:11:55.92900:11:55.939 with the unique complexities that each 00:11:59.13900:11:59.149 case presents each model becomes 00:12:02.22900:12:02.239 essentially a living report 00:12:04.15000:12:04.160 that requires no in our expertise to use 00:12:07.38900:12:07.399 effectively to conduct what-if studies 00:12:09.75900:12:09.769 etc now collaboration is a key aspect 00:12:14.53000:12:14.540 when creating project specific models of 00:12:17.79900:12:17.809 existing new or conceptual plants our 00:12:21.34000:12:21.350 role is to provide focus support to 00:12:24.10000:12:24.110 companies that want to innovate to help 00:12:27.36900:12:27.379 their smart young engineers and 00:12:29.57900:12:29.589 experienced individuals in progressive 00:12:32.25900:12:32.269 companies understand existing or new 00:12:35.15900:12:35.169 process problems to innovate and reduce 00:12:38.61900:12:38.629 the technical and financial risks you 00:12:42.46000:12:42.470 know you would have noticed from our 00:12:43.84000:12:43.850 comminution fly sheet demo example that 00:12:47.04900:12:47.059 each unit block represents a piece of 00:12:49.86900:12:49.879 equipment found on a specific 00:12:51.92900:12:51.939 metallurgical plant a custom bolt 00:12:55.50900:12:55.519 interface allows the input parameters to 00:12:58.62900:12:58.639 be adjusted by the inducer so this is 00:13:01.59900:13:01.609 what our example library looks like 00:13:04.65900:13:04.669 let's double click on a ball mill unit 00:13:07.90000:13:07.910 block at this step the inducer with 00:13:11.40000:13:11.410 the initial conditions any reaction rate 00:13:15.93000:13:15.940 constants that may be of interest 00:13:17.90000:13:17.910 equipment properties and volumes species 00:13:21.66000:13:21.670 parameters like densities and material 00:13:24.68900:13:24.699 parameters like for example the 00:13:28.29000:13:28.300 characteristic brackets matrix of the or 00:13:31.61900:13:31.629 in this specific model now user-friendly 00:13:37.13900:13:37.149 drag-and-drop flow connectors allow easy 00:13:40.76900:13:40.779 flow sheet construction and I shall post 00:13:43.92000:13:43.930 upcoming videos to illustrate just how 00:13:47.16000:13:47.170 simple it is to build these flow sheets 00:13:50.55000:13:50.560 once the custom library has been 00:13:53.97000:13:53.980 developed now an attractive feature of 00:13:57.66000:13:57.670 this approach is that it gives us 00:14:00.17900:14:00.189 flexibility to match each end users 00:14:03.92000:14:03.930 situation or experience so by the way 00:14:08.34000:14:08.350 these are all arbitrary images used to 00:14:12.56900:14:12.579 mask our unit blocks for illustrative 00:14:15.60000:14:15.610 purposes I obtained these images from 00:14:19.94900:14:19.959 the grab cat open source library as 00:14:23.40000:14:23.410 acknowledged here at the top finally 00:14:27.90000:14:27.910 this diagram explains the typical 00:14:31.43900:14:31.449 project workflow the point here is that 00:14:34.80000:14:34.810 value can be created at the initial 00:14:38.24900:14:38.259 first level with minimal available data 00:14:41.72000:14:41.730 there's always some initial data 00:14:44.22000:14:44.230 available importantly this first level 00:14:48.24000:14:48.250 model can also be used to direct the 00:14:51.92900:14:51.939 test program and extract maximum value 00:14:55.23000:14:55.240 from expensive R&D the detail or second 00:15:00.62900:15:00.639 level is typically used to add more 00:15:03.72000:15:03.730 complexity where required and to 00:15:05.81900:15:05.829 calibrate the model this level 00:15:08.42900:15:08.439 represents the living report 00:15:10.47000:15:10.480 of the project and is used to optimize 00:15:13.63900:15:13.649 design and operating strategies for the 00:15:16.92000:15:16.930 process thank you for watching this 00:15:20.28000:15:20.290 video please visit our website 00:15:23.31900:15:23.329 you require more information
Office location
Engineering company LOTUS®
Russia, Ekaterinburg, Lunacharskogo street, 240/12