WEBVTT 00:00:00.240 --> 00:00:05.200 All pipes carrying fluids experience losses  of pressure caused by friction and turbulence 00:00:05.200 --> 00:00:09.120 of the flow. It affects seemingly simple  things like the plumbing in your house 00:00:09.120 --> 00:00:14.080 all the way up to the design of massive, way  more complex, long-distance pipelines. I’ve 00:00:14.080 --> 00:00:18.720 talked about many of the challenges engineers  face in designing piped systems, including 00:00:18.720 --> 00:00:23.760 water hammer, air entrainment, and thrust forces.  But, I’ve never talked about the factors affecting 00:00:23.760 --> 00:00:28.160 how much fluid actually flows through a pipe  and the pressures at which that occurs. So, 00:00:28.160 --> 00:00:32.640 today we’re going to have a little fun, test  out some different configurations of piping, 00:00:32.640 --> 00:00:36.880 and see how well the engineering equations  can predict the pressure and flow. 00:00:36.880 --> 00:00:41.200 Even if you’re not going to use the equations,  hopefully, you’ll gain some intuition from seeing 00:00:41.200 --> 00:00:47.040 how they work in a real situation. I’m Grady and  this is Practical Engineering. In today’s episode, 00:00:47.040 --> 00:00:57.840 we’re talking about closed conduit  hydraulics and pressure drop in pipes. 00:00:59.120 --> 00:01:03.680 This video is sponsored by HelloFresh,  America’s number 1 meal kit. More on that later. 00:01:08.800 --> 00:01:14.320 I love engineering analogies, and in this case,  there are a lot of similarities between electrical 00:01:14.320 --> 00:01:19.440 circuits and fluids in pipes. Just like all  conventional conductors have some resistance 00:01:19.440 --> 00:01:24.560 to the flow of current, all pipes impart some  resistance to the flow of the fluid inside, 00:01:24.560 --> 00:01:29.920 usually in the form of friction and turbulence.  In fact, this is a lovely analogy because 00:01:29.920 --> 00:01:34.960 the resistance of a conductor is both a function  of the cross-sectional area and length of the 00:01:34.960 --> 00:01:40.880 conductor—the bigger and shorter the wire, the  lower the resistance. The same is true for pipes, 00:01:40.880 --> 00:01:46.400 but the reasons are a little different. The fluid  velocity in a pipe is a function of the flow rate 00:01:46.400 --> 00:01:51.200 and the pipe’s area. Given a flowrate, a  larger pipe will have a lower velocity, 00:01:51.200 --> 00:01:55.920 and a small pipe will have a higher velocity.  This concept is critical to understanding the 00:01:55.920 --> 00:02:01.840 hydraulics of pipeline design because friction and  turbulence are mostly a result of flow velocity. 00:02:01.840 --> 00:02:06.800 I built this demonstration that should help  us see this in practice. This is a manifold 00:02:06.800 --> 00:02:11.680 to test out different configurations of pipes and  see their effect on the flow and pressure of the 00:02:11.680 --> 00:02:16.080 fluid inside. It’s connected to my regular  tap on the left. The water passes through 00:02:16.080 --> 00:02:20.880 a flow meter and valve, past some pressure  gauges, through the sample pipe in question, 00:02:20.880 --> 00:02:24.560 and finally through a showerhead. I  picked a showerhead since, for many of us, 00:02:24.560 --> 00:02:29.200 it’s the most tangible and immediate connection  we have to pressure problems in plumbing. It’s 00:02:29.200 --> 00:02:33.280 probably one of the most important factors in the  difference between a good shower, and a bad one. 00:02:33.280 --> 00:02:37.280 Don’t worry, all this water will be given  to my plants which need it right now anyway. 00:02:37.280 --> 00:02:40.960 I used these clear pipes because they  look cool, but there won’t be much to see 00:02:40.960 --> 00:02:46.080 inside. All the information we need will show  up on the gauges (as long as I bleed all the 00:02:46.080 --> 00:02:50.960 air from the lines each time). The first one  measures the flow rate in gallons per minute, 00:02:50.960 --> 00:02:54.720 the second one measures the pressure  in the pipe in pounds per square inch, 00:02:54.720 --> 00:02:59.120 and the third gauge measures the difference  in pressure before and after the sample 00:02:59.120 --> 00:03:04.480 (also called the head loss) in inches of water. In  other words, this gauge measures how much pressure 00:03:04.480 --> 00:03:09.600 is lost through friction and turbulence in the  sample - this is the one to keep your eye on. In 00:03:09.600 --> 00:03:15.040 simple terms, it’s saying how far do you have to  open the valve to achieve a certain rate of flow. 00:03:15.680 --> 00:03:19.760 I know the metric folks are giggling at these  units. For this video, I’m going to break my 00:03:19.760 --> 00:03:24.640 rule about providing both systems of measurement  because these values are just examples anyway. 00:03:24.640 --> 00:03:29.120 They are just nice round numbers that are easy  to compare with no real application outside the 00:03:29.120 --> 00:03:34.080 demo. Substitute your own preferred units if you  want, because it won’t affect the conclusions. 00:03:34.080 --> 00:03:39.280 There are a few methods engineers use to estimate  the energy losses in pipes carrying water, 00:03:39.280 --> 00:03:44.080 but one of the simplest is the Hazen-Williams  equation. It can be rearranged in a few ways, 00:03:44.080 --> 00:03:49.120 but this way is nice because it has the variables  we can measure. It says that the head loss (in 00:03:49.120 --> 00:03:54.400 other words the drop in pressure from one end of a  pipe to the other) is a function of the flow rate, 00:03:54.400 --> 00:03:58.640 and the diameter, length, and roughness of  the pipe. Now - that’s a lot of variables, 00:03:58.640 --> 00:04:03.280 so let’s try an example to show how this works.  First, we’ll investigate the effect the length 00:04:03.280 --> 00:04:07.760 of the pipe has on head loss. I’m starting  with a short piece of pipe in the manifold, 00:04:07.760 --> 00:04:15.920 and I’m testing everything at three flow rates:  0.3, 0.6, and 0.9 gallons per minute (or gpm). 00:04:15.920 --> 00:04:22.000 At 0.3 gpm, we see pressure drop across the pipe  is practically negligible, just under half an 00:04:22.000 --> 00:04:32.240 inch. At 0.6 gpm, the head loss is about an inch.  And, at 0.9 gpm, the head loss is just over 3 00:04:32.240 --> 00:04:38.640 inches. Now I’m changing out the sample for a much  longer pipe of the same diameter. In this case, 00:04:38.640 --> 00:04:44.320 it’s 20 times longer than the previous example.  Length has an exponent of 1 in the Hazen-Williams 00:04:44.320 --> 00:04:49.120 equation, so we know if we double the length,  we should get double the head loss. And if we 00:04:49.120 --> 00:04:54.400 multiply the length times 20, we should see the  pressure drop increase by a factor of 20 as well. 00:04:54.400 --> 00:05:00.880 And sure enough, at a flow rate of 0.3 gpm, we  see a pressure drop across the pipe of 7.5 inches, 00:05:00.880 --> 00:05:06.000 just about 20 times what it was with the short  pipe. That’s the max we can do here - opening 00:05:06.000 --> 00:05:10.560 the valve any further just overwhelms the  differential pressure gauge. There is so much 00:05:10.560 --> 00:05:14.880 friction and turbulence in this long pipe that I  would need a different gauge just to measure it. 00:05:15.680 --> 00:05:20.480 Length is just one factor that influences the  hydraulics of a pipe. This demo can also show 00:05:20.480 --> 00:05:25.200 how the pipe diameter affects the pressure  loss. If I switch in this pipe with the same 00:05:25.200 --> 00:05:29.920 length as the original sample but which has  a smaller diameter, we can see the additional 00:05:29.920 --> 00:05:34.720 pressure drop that occurs. The smaller pipe  has ⅔ the diameter of the original sample, 00:05:34.720 --> 00:05:40.720 and diameter has an exponent of 4.9 in our  equation. That’s because, as I mentioned before, 00:05:40.720 --> 00:05:46.480 changing the diameter changes the fluid velocity,  and friction is all about velocity. We expect the 00:05:46.480 --> 00:05:54.880 pressure drop to be 1 over (⅔)^4.9 or about 7  times higher than the original pipe. At 0.3 gpm, 00:05:54.880 --> 00:06:00.480 the pressure drop is 3 inches. That’s  about 6 times the original. At 0.6 gpm, 00:06:00.480 --> 00:06:06.720 the pressure drop is 7.5 inches, about  7 times the original. And at 0.9 gpm, 00:06:06.720 --> 00:06:12.320 we’re off the scale. All of that is to say, we’re  getting close to the correct answers, but there’s 00:06:12.320 --> 00:06:17.200 something else going on here. To explore this  even further, let’s take it to the extreme. 00:06:17.200 --> 00:06:22.400 We’ll swap out a pipe with a diameter 5 times  larger than the original sample. In this case, 00:06:22.400 --> 00:06:29.040 we’d expect the head loss to be 1 over 5^4.3,  basically a tiny fraction of that measured with 00:06:29.040 --> 00:06:34.480 the original sample. Let’s see if this is the  case. At 0.3 gpm, the pressure drop is basically 00:06:34.480 --> 00:06:41.520 negligible just like last time. At 0.6 and 0.9  gpm, the pressure drop is essentially the same as 00:06:41.520 --> 00:06:46.320 the original. Obviously, there’s more to the head  loss than just the properties of the pipe itself, 00:06:46.320 --> 00:06:50.400 and maybe you caught this already. There is  something conspicuous about the Hazen-Williams 00:06:50.400 --> 00:06:55.280 equation. It estimates the friction in a pipe,  but it doesn’t include the friction and turbulence 00:06:55.280 --> 00:07:00.720 that occurs at sudden changes in direction or  expansion and contraction of the flow. These 00:07:00.720 --> 00:07:05.520 are called minor losses, because for long pipes  they usually are minor. But in some situations 00:07:05.520 --> 00:07:10.000 like the plumbing in buildings or my little  demonstration here, they can add up quickly. 00:07:10.000 --> 00:07:15.120 Every time a fluid makes a sudden turn (like  around an elbow) or expands or contracts (like 00:07:15.120 --> 00:07:20.080 through these quick-release fittings), it  experiences extra turbulence, which creates 00:07:20.080 --> 00:07:24.960 an additional loss of pressure. Think of it like  you are walking through a hallway with a turn. You 00:07:24.960 --> 00:07:30.160 anticipate the turn, so you adjust your path  accordingly. Water doesn’t, so it has to crash 00:07:30.160 --> 00:07:34.720 into the side - and then change directions.  And, there is actually a formula for these minor 00:07:34.720 --> 00:07:39.760 losses. It says that they are a function of the  fluid’s velocity squared and this k factor that 00:07:39.760 --> 00:07:45.040 has been measured in laboratory testing for any  number of bends, expansions, and contractions. 00:07:45.040 --> 00:07:50.400 As just another example of this, here’s a sample  pipe with four 90-degree bends. If you were just 00:07:50.400 --> 00:07:56.000 calculating pressure loss from pipe flow, you  would expect it to be insignificant. Short, 00:07:56.000 --> 00:08:01.200 smooth pipe of an appropriate diameter. The  reality is that, at each of the flow rates tested 00:08:01.200 --> 00:08:06.480 in the original straight pipe sample, this one has  about double the head loss, maxing out at nearly 00:08:06.480 --> 00:08:13.920 6 inches of pressure drop at 0.9 gpm. Engineers  have to include “minor” losses to the calculated 00:08:13.920 --> 00:08:19.680 frictional losses within the pipe to estimate the  total head loss. In my demo here, except for the 00:08:19.680 --> 00:08:25.280 case of the 20’ pipe, most of the pressure drop  between the two measurement points is caused by 00:08:25.280 --> 00:08:30.000 minor losses through the different fittings in the  manifold. It’s why, in this example, the pressure 00:08:30.000 --> 00:08:34.960 drop is essentially the same as the original. Even  though the pipe is much larger in diameter, the 00:08:34.960 --> 00:08:39.840 expansion and contraction required to transition  to this large pipe make up for the difference. 00:08:40.400 --> 00:08:45.440 One clarification to this demo I want to make:  I’ve been adjusting this valve each time to keep 00:08:45.440 --> 00:08:50.720 the flow rate consistent between each example  so that we make fair comparisons. But that’s not 00:08:50.720 --> 00:08:55.360 how we take showers or use our taps. Maybe  you do it differently, but I just turn the 00:08:55.360 --> 00:09:00.240 valve as far as it will go. The resulting flow  rate is a function of the pressure in the tap 00:09:00.800 --> 00:09:06.400 and the configuration of piping along  the way. More pressure or less friction 00:09:06.400 --> 00:09:10.800 and turbulence in the pipes and fittings  will give you more flow (and vice versa). 00:09:11.440 --> 00:09:15.040 So let’s tie all this new knowledge  together with an example pipeline. 00:09:15.040 --> 00:09:18.320 Rather than just knowing the total  pressure drop from one end to another, 00:09:18.880 --> 00:09:23.680 engineers like to draw the pressure continuously  along a pipe. This is called the hydraulic grade 00:09:23.680 --> 00:09:28.800 line, and, conveniently, it represents the  height the water would reach if you were to tap 00:09:28.800 --> 00:09:33.840 a vertical tube into the main pipe. With a  hydraulic grade line, it’s really easy to see 00:09:33.840 --> 00:09:38.960 how pressure is lost through pipe friction.  Changing the flow rate or diameter of the pipe 00:09:38.960 --> 00:09:44.240 changes the slope of the hydraulic grade line.  It’s also easy to see how fittings create minor 00:09:44.240 --> 00:09:49.440 losses in the pipe. This type of diagram  is advantageous in many ways. For example, 00:09:49.440 --> 00:09:53.920 you can overlay the pressure rating of the pipe  and see if you’re going above it. You can also 00:09:53.920 --> 00:09:59.200 see where you might need booster pump stations  on long pipelines. Finally, you can visualize how 00:09:59.200 --> 00:10:04.880 changes to a design like pipe size, flow rate,  or length affect the hydraulics along the way. 00:10:11.520 --> 00:10:17.040 Friction in pipes? Not necessarily the  most fascinating hydraulic phenomenon. But, 00:10:17.040 --> 00:10:22.560 most of engineering is making compromises, usually  between cost and performance. That’s why it’s so 00:10:22.560 --> 00:10:28.800 useful to understand how changing a design can  tip the scales. Formulas like the Hazen-Williams 00:10:28.800 --> 00:10:33.840 and the minor loss equations are just as useful  to engineers designing pipelines that carry 00:10:33.840 --> 00:10:38.560 huge volumes of fluid all the way down to  homeowners fixing the plumbing in their houses. 00:10:38.560 --> 00:10:44.560 It’s intuitive that reducing the length of a pipe  or increasing its diameter or reducing the number 00:10:44.560 --> 00:10:49.360 of bends and fittings ensures that more of the  fluid’s pressure makes it to the end of the line. 00:10:49.360 --> 00:10:54.960 But engineers can’t rely just on intuition.  These equations help us understand how much 00:10:54.960 --> 00:11:00.000 of an improvement can be expected without having  to go out to the garage and test it out like I 00:11:00.000 --> 00:11:04.320 did. Pipe systems are important to us,  so it’s critical that we can design them 00:11:04.320 --> 00:11:08.960 to carry the right amount of flow without too  much drop in pressure from one end to the other. 00:11:10.800 --> 00:11:12.800 It’s time for everyone’s favorite segment of 00:11:12.800 --> 00:11:16.000 me trying to cook while my wife  tries to capture that on video. 00:11:16.000 --> 00:11:17.040 “And… Action!” 00:11:20.320 --> 00:11:22.400 “Who cut this tiny hole in the cheese?” 00:11:24.160 --> 00:11:27.280 Goofing around in the kitchen is one  of our favorite things to do together. 00:11:27.280 --> 00:11:30.480 That’s why we’re thankful for  HelloFresh, the sponsor of this video, 00:11:30.480 --> 00:11:34.320 for converting cooking from a chore into  our favorite thing to do on date night. 00:11:34.320 --> 00:11:35.440 “So delizioso!” 00:11:36.960 --> 00:11:39.920 Sometimes, the hardest part about  dinner is just deciding what to have, 00:11:39.920 --> 00:11:45.120 so it’s nice to have HelloFresh curating  delicious and healthy recipes so we don’t have to. 00:11:45.120 --> 00:11:46.000 “How’s it feel?” 00:11:49.120 --> 00:11:53.520 The pre-portioned ingredients mean there’s less  prep and less food waste, and the packaging is 00:11:53.520 --> 00:11:59.280 mostly recyclable or already recycled content.  HelloFresh also helps us get dinner ready quickly 00:11:59.280 --> 00:12:04.000 on the days we don’t feel like planning, prep, and  shopping. We get to skip straight to the fun part. 00:12:04.000 --> 00:12:04.160 “Ewww!” 00:12:10.320 --> 00:12:12.960 Go try it yourself at HelloFresh.com and use 00:12:12.960 --> 00:12:17.200 code PRACTICAL12 to get 12 free  meals, including free shipping. 00:12:17.200 --> 00:12:21.680 Supporting our sponsors helps support this  channel. That’s HelloFresh.com and use code 00:12:21.680 --> 00:12:37.840 PRACTICAL12. Thanks, HelloFresh, and thank  YOU for watching. Let me know what you think.