Neutrium 2: The Joule-Thomson Effect and Hydrate Formation

Hello and welcome to the second episode of the Nutrient Podcast. I'm Trevor Walker and I'm Matthew Campbell and today we're going to discuss the Joule-Thompson effect and its unwanted side effect, hydrate formation. Okay, so the Joule-Thompson effect, which is also known as the Joule-Kelvin effect or really any combination of the names Joule, Kelvin and Thompson, is the process by which a gas experiences temperature change as it is rapidly reduced in pressure or as we say in the beers let down. While the JT effect can result in either an increase or a decrease in the downstream grass temperature, it is typically the cooling effect we want to exploit. Yeah, there's easier ways than the Joule-Thompson effect to heat gases. You could use steam or direct fired or salt baths or pretty much anything. There's no practical use for Joule-Thompson as a heating mechanism. Just depends on what type of waste energy you've got lying around on site. So today we are going to discuss the JT effect, how it works, what applications it is used for and what the hazards are. We are also going to touch on hydrate formation, which is an unwanted side effect of JT cooling, and the cause of several industrial accidents such as the phase in explosion which we are going to discuss later in the episode. So Matt, why don't you tell us about your recent run in with the Joule-Thompson effect other than grabbing a cold one from the fridge? Yeah, so recently I was working on a gas plant where we use Joule-Thompson cooling to separate the heavy parts out of raw gas like propane and butane and we do that through a valve that they call the Joule-Thompson valve. But also we had a bit of an incident where we came in and the one of the shutdown valves was passing and you could see the ice all down the outside of the piping as it moved away from the valve and there was obviously a bit of Joule-Thompson cooling going on there too. Hopefully they were using low temperature carbon steel. Yeah. Okay so before we get into the implications of the JT effect let's Let's cover some of the basic principles. First off, when we say letting down the gas, what we mean is forcing it through a restriction such as a throttling valve with high pressure on the upstream side and low pressure on the downstream. Now this is going to get a little bit hairy but stick with me. Imagine a discrete packet of gas about to pass through a restriction. Now to actually get through the restriction, some work needs to be done by the upstream gas to push it through. And this work is equal to the volume of the packet times the upstream pressure. As our little packet makes it to the other side of the restriction, it needs to make way for itself by displacing some of the downstream gas, which involves doing work equal to the product of the downstream pressure and volume. For real gases, the amount of work done upstream will not equal the amount of work done downstream due to various intermolecular forces, but we are all beholden to the first law of thermodynamics, and therefore the difference in the upstream and downstream work must be accounted for to conserve energy. If depressuring is an adiabatic process, meaning the gas cannot exchange heat or work with its surroundings, the only thing that can satisfy the first law is a change in internal energy. Now, don't be put off by the mention of internal energy. It simply means the energy associated with the random motion of the gas molecules, which happens to be experienced by us as temperature. In a real gas, we have what is called Van der Waals forces, which are the intermolecular forces acting between the gas molecules. Just like people, gas molecules are subject to attractive and repulsive forces as they move past each other. As the pressure of the gas is lowered, the average distance between molecules increases and attractive forces become dominant over repulsive forces. This results in an increase in the potential energy of the molecules as they are further separated. It's a bit of the old, good from far but far from good. Now given that the potential energy of the molecules increase as the gas depressures, we need to make up the energy from somewhere else in the system to satisfy the first law. It just so happens that the best source of energy is the kinetic energy of the molecules randomly moving around, i.e. the internal energy. So as the gas depressures across the restriction, the internal energy decreases to compensate for the increase in potential energy, which means the molecules slow down and the temperature drops. It's important to note that most gases behave in this way at the temperatures and pressures at which we normally operate processes. However, not all do. At normal temperatures, say 20 degrees Celsius, hydrogen, helium and neon actually heat up when expanded. As you might have deduced, this is because the repulsive forces still dominate at low pressures and high temperatures. As these gases are expanded, the potential energy is reduced and the internal energy is increased, and therefore the temperature increases as well. As hinted previously, the direction of the temperature change is dependent on the temperature and pressure the gas is at. Technically speaking, depressuring is an isenthalpic process, which means enthalpy stays constant. For any gas, the temperature could either increase or decrease depending on how the internal energy has to change to keep enthalpy constant. I know that's a lot to get your head around and might need a re-listen, but I promise that's the worst of it for this episode. In more practical terms, there's a handy rule of thumb for JT cooling for the conditions we normally handle natural gas at, you'll lose roughly half a degree Celsius for every one bar or 100 kPa. So if you were dropping 30 bar across a valve, you'll drop about 15 degrees Celsius. I guess in imperial units that would be something like 30 pounds per furlong? I'm not sure, I might have to go back and listen to the first episode again. But trolling the US listeners aside, you lose about 6 degrees Fahrenheit per 100 psi, or if you really want to mix your unit systems up, it's about 1 degree Fahrenheit per 100 kPa or 1 bar. To figure out whether a gas will cool or heat on expansion, we can check a parameter called the inversion temperature. The inversion temperature is the point at which decreasing the pressure causes no temperature change. At the end of the pressure drop step, it will be the same temperature it started at. So on one side of the inversion temperature, our gas will cool, and on the other side, our gas will heat as it goes through the pressure drop. Gases generally cool when they're expanded at lower pressures and at a middle band of temperatures. Outside of this, at high pressures and the extremes of temperature, the gas will heat when expanding. On a plot of temperature versus pressure, it forms a curve that kind of looks like a sideways parabola. You can check it out at the Nutrium.net article on Joule-Thomson cooling. It'll be much easier to understand when you see it on a graph. So looking at our gases that don't cool at 20 degrees Celsius, hydrogen, helium and neon, they are well above their upper inversion temperatures and well into the region where they heat up when the pressure is reduced. We would need to pre-cool these gases significantly to get them into a range where they'll experience JT cooling. It's no coincidence that these gases are very light and/or inert, qualities that give them low attractive forces. Yeah, you could actually cross the inversion temperature during a depressuring process, so you would have your gas heating for a while until it hit a low enough pressure after which it would start cooling. Yeah, but at the pressures and temperatures you see that at, it's more of a theoretical curiosity than a practical concern. It's worth noting here that there's another process that's also used in industry for cooling gases, the turbo expander. It's similar in principle to Joule-Thompson cooling, but the depressuring occurs through a turbine rather than a valve or orifice. When depressuring through a turbine, work is also extracted mechanically, which technically makes the process isentropic rather than isenthalpic like our pure JT system. So as our gas expands through the turbine, the temperature will decrease because of the increase in internal potential energy, just as we saw for JT cooling. But we're also taking out energy with our rotating turbine, and because this energy must come from somewhere, this will also decrease the temperature. With these two effects, it is possible to get even lower temperatures than pure JT cooling, while also using the energy extracted for some other purpose, typically running a compressor. So we know that we can use Joule-Thomson cooling and turbo expanders to make gas really cold, but what are we going to do with it? We humans are fairly crafty and we've come up with quite a few uses for these processes, mostly around gas separation and liquefaction. You can use JT cooling to liquefy air or its components. Carl Lind developed a process where air is pressurized, pre-cooled, and then expanded through a JT valve until it has reached a low enough temperature to liquefy. The process takes some of the liquefied air it has produced and boils it off to pre-cool the incoming pressurized air. This process was later enhanced by George's Claude, who added the turbo expander enhancement to the JT process to really get things cold in a hurry. With this process we can make liquefied oxygen and nitrogen. Argon can also be extracted from the air in this way. Argon is about 1% of the Earth's atmosphere and the boiling point is higher than oxygen or nitrogen, so while you are liquefying air, you might as well make some argon. Historically some other gases like chlorine and ammonia have been liquefied using Joule-Thompson cooling or turbo expanders, but these gases are pretty hazardous in liquid form and so it's preferable to keep them in their gaseous state. For example, I was once on a site that produced a lot of chlorine based products, although they had stopped producing liquid chlorine by the time I was there. Even then it was still common to hear about personics who was incapacitated by liquid chlorine rapidly vaporising into their face as they serviced a valve. Similar to Lind and Claude's process for liquefying air, a combination of JT valves and a turbo expander is often used to separate the heavy components from raw natural gas. You can use a process like this to simply remove heavy components so that the raw gas is brought to pipeline quality or is suitable for use as a fuel gas. The most basic example would drop the pressure of the gas through a JT valve to get it cold enough to knock out most of the components heavier than propane as liquids. Then you have decent quality natural gas to run fired equipment like heaters and generators. The same cooling processes are also used in an LPG plant in conjunction with three distillation columns, deethanizer, depropanizer and debutanizer, to make pipeline quality natural gas, liquid propane, liquid butane and condensate. Liquid propane and butane are used for gas stoves, lighters, barbecues and all mixed together to make automotive LPG. Condensate is everything left over and is the lighter parts of gasoline. We can also use JT cooling for refrigeration, although it's uncommon. Most refrigeration uses a compression and expansion cycle that compresses a gas until it's a liquid and then reduces the pressure to take advantage of the boiling of that liquid for greater heat transfer and more stable temperatures. There are some applications however which use a straight JT compression cycle without any phase change for the refrigeration. Air conditioning is basically the same as refrigeration, and all the common air conditioners you'll find in your house or office will use a phase change process. But in aircraft, they use our JT and turbo expander technologies. This is because they already have a great source of high pressure air - the engines. The first stages of a jet engine compress the incoming air for the combustion chamber. A side stream of this high pressure air is taken and used to pressurize and air condition the cabin. When the compressed air is too warm, it's put through an expansion to cool it to a comfortable temperature. Yeah, that's probably why it's so terrible to be stuck at the gate in an aircraft, because they haven't got the engines running and they're just blowing air at you with a fan. Yeah, they might have a small conventional air conditioner on the plane somewhere, but it doesn't really make sense for them to carry a full-size conventional air conditioner through the sky all that time when they get all the cooling for free. Yeah, so we've told you how JT works and what it can be used for, but what are the process safety impacts of all this cooling? Let's talk about depressuring systems in metallurgy for a second. When depressuring a gas there is a risk that cooling will be significant enough to drop the temperature below the minimum design metal temperature or MDMT of the piping or equipment. The MDMT of a vessel or piping is often an arbitrary value selected during design but roughly speaking it is the lowest temperature the integrity of the metal is guaranteed at a specific pressure. The MDMT varies with the materials of construction and the metal thickness. So when your process includes a gas depressuring stage such as a chiroplan or gas distribution letdown station, you need to carefully consider whether your JT effect is significant enough to push the temperature below the MDMT and adjust your metallurgy appropriately. Otherwise you can get low temperature embrittlement which can lead to loss of containment. And these considerations don't just need to be made for steady state processes, you also need to consider infrequent operations such as vessel blowdown and relief valve operation. aren't always aware that during vessel blowdown the temperature in the vessel will also drop. Luckily on the vessel side you have a bit more headroom. Although the falling temperatures can lower the steel yield strength, the forces exerted on the vessel decline as it depresses. So as the vessel weakens, the forces it has to bear reduce. Often dynamic modelling and astute understanding of standards such as ASMI B31.3 can help you optimise the mechanical design. On the piping side, if you're expecting temperatures lower than about -20C, you need to go to low temperature carbon steel, which can be safe down to temperatures in the range of -40C to -100C depending on the grade. Much beyond that, you will have to consider stainless steel which lets you get down to about -195C or even more exotic materials. But low temperature embrittlement can be a complicated topic and it's best to talk to a competent mechanical engineer. Moving on to liquid dropout, Depressuring a gas can cause the system to fall below the dew point temperature and cause gases to liquefy, which is bad when you aren't planning for it and can lead to problems such as liquid damage to turbines if they're not designed to handle it, or slugging liquid flow in pipelines. Slugging may or may not be a problem depending on the design of the process, but it's generally desirable to avoid it due to the instability it can cause. Whether Joule-Thomson cooling is caused deliberately or accidentally, you need to think about the properties of the gas being cooled. If the gas contains water, it might drop out as a liquid and then freeze. If the gas contains water and natural gas, it may form hydrates. Both of these can block up piping and equipment, which can have all kinds of knock-on safety effects. Hydrates are a fairly complicated animal and a big hazard in the natural gas industry, so we're going to spend some time talking about what they are and what you can do about them now. So what are hydrates? Hydrates are a solid crystal structure made up of water and light hydrocarbon molecules. You can think of it as a hydrocarbon molecule, like methane, trapped in a cage of water molecules. For a typical methane hydrate, you'll find the ratio of methane to water molecules is about 1 to 6, depending on the occupancy rate of the cages. There are three structures of hydrates, Structure 1, Structure 2 and Structure H. These correspond to the molecules that have been trapped in the cages, and differences in the cages that result. Structure 1 forms when smaller molecules are trapped, such as methane, ethane, carbon dioxide and hydrogen sulfide. There are two structures of the cage, the 12-sided dodecahedron and the 14-sided tetrakaidecahedron. Structure 2 forms when larger hydrocarbons are trapped, such as propane and butane, and sometimes even nitrogen. Type 2 hydrates are typically formed from the smaller 12-sided dodecahedrons and the larger 16-sided hexakai decahedron. Structure H forms when certain isoparaffins and cycloalkanes are trapped. The structures are much larger and require rarer conditions to form. Smaller molecules such as methane form smaller hydrate structures like we had for structure I, which can then be built up into larger cages, which trap the larger molecules. The cages are made up of the 12-sided dodecahedron, irregular dodecahedrons, and a 20-sided irregular icosahedron. You can think of these cage structures as soccer balls of varying sizes, with a guest molecule trapped inside. When you can see hydrates, they sort of look like normal ice, although the appearance of ice can vary depending how it's formed, but is like a packed snow kind of appearance. In practice, it is somewhat difficult to see hydrates, as they're only stable at low temperatures or high pressures. Sometimes you can see them fleetingly as you blow down a line. In these cases, where they're more of a suspension than a solid, they can appear to have the consistency of snot. There's actually vast amounts of methane hydrates in the cold parts of the earth and at the bottom of the ocean, places where it's very cold or has a high pressure. It is speculated that the stored energy in methane hydrates, technically methane calthrate, is about 2 to 10 times the energy available in all the world's natural gas reserves. There are projects underway to attempt to commercially exploit these reserves, but I'm not aware of any operations in production. So what are the conditions under which hydrates form? It really boils down to the availability of two key components, water and hydrocarbons. The hydrocarbons as we mentioned earlier can be light alkanes and alkenes, like your methane, ethane, propane and butanes, or some isopropanes and cycloalkanes, everything commonly found in the natural gas stream. Now the water will already be present in the gas stream as a side effect of the extraction process, but the presence of water alone isn't enough for the hydrates to form. We need the stream to be at a temperature below its dew point, bearing in mind that the dew point is a function of pressure. As you might remember from earlier in the episode, the JT effect typically results in a reduction in temperature. So you can see why the JT effect and hydrate formation go hand in hand in the gas processing industry. In a typical natural gas processing or distribution system, we see pressures somewhere between 0 and 10,000 kPa. For this pressure range we expect to see hydrate formation at temperatures between about -80C for low pressure and 20C for high pressures. We see hydrate formation when the system temperature falls below the hydrate formation temperature. When the temperature increases back above the hydrate formation temperature, hydrates decompose rather than melt, with both the water and the hydrocarbon subliming back into the gas. There are several factors governing hydrate formation. Now make sure to pay attention as these are going to be important when we look at how to prevent hydrates forming. The primary factors are the dew point, low temperature, high pressure and the gas composition. Factors with a more minor effect on hydrate formation are mixing, nucleation sites, kinetics and salinity. So we now know how hydrates will form, but what does this mean for our process? Some of the consequences of hydrate formation include reduction of flow in pipelines, blocking of pipelines, the fouling of equipment, the blocking of instruments, or trapping pockets of fluid or pressure. These consequences range in severity from nuisance efficiency losses, such as a restriction in a flow line or the fouling of a heat exchanger, all the way to critical safety hazards, such as the blocking of safety critical instrumentations or valves. So you've gone and formed hydrates. What do you do now? The simplest fix when you have time and accessibility on your hands is to remove the conditions that allow hydrates to be stable, that is reduce the pressure or increase the temperature. This has the effect of melting or subliming the hydrates away and is a guaranteed fix. There are practical problems with this approach though. Firstly we'll look at increasing the temperature. To increase the temperature of your hydrates you need to wait for the ambient temperature to heat it up, or you need to apply heat yourself. Applying heat directly to the hydrate is unlikely to be a practical solution. If your hydrate is in a gas plant, which is cold, it's probably insulated, so you can't get to it. If your hydrate is in a pipeline, it is probably many kilometres long, so how are you going to find it? And if you're doing undersea gas production, and you make a hydrate at the wellhead, well, you're in a world of hurt getting to that. Heating a hydrate can also result in a rapid increase in pressure and gas volume, which is something that must be kept in mind to prevent a dangerous situation from occurring. We can also try to decrease the pressure to remove our hydrate. As the pressure decreases, the dew point temperature increases. Because of this, the hydrate will begin to melt or supplime and slowly disappear. There are a few practical problems with this approach too. Firstly for natural gas, your product is both valuable and environmentally harmful, so reducing the pressure would usually mean sending the gas to atmosphere or flare, which you want to avoid if you can. This issue is compounded when you are talking about long pipelines or undersea productions where there can be many kilometers of piping that must be blown down, which is costly and time consuming. You may also get to a situation where there is pressure trapped by the hydrate, which makes it impossible to reduce pressure on both sides of the blockage. If you only blow down the pressure on one side of the hydrate, you need to be careful, because you have made an ice cannon, and when the hydrate starts to loosen, a sudden rush of gas can propel the hydrate forward. A methanol or glycol injection is another potential fix. This is usually only possible because someone has seen the problem coming and provided you with a suitable injection point, but the application of methanol or glycol directly to the hydrate will accelerate decomposition. This fix is also governed by pipe layout and positioning, since it only works if you can bring the chemical into direct contact with the hydrate. If you have a hydrate stuck up in a high point, it's going to be very difficult to get the liquid methanol or glycol to contact it without completely liquid filling your system. At this point you might be thinking, so hydrates can block some pipes, what's the big deal? Well let me tell you a little story about an oil refinery near the town of Faisanne, France. On the 4th of January 1966 a large explosion occurred at the Elf owned and operated refinery which was in part caused by hydrates. An operator was draining water from a propane vessel. The drain line on this vessel featured two valves in series. Correct procedure would see the valve closer to the tank remain open with the second valve being used to control the flow rate. This means that if the lower valve freezes, the warmer upstream valve could be used to isolate flow. When draining was nearly complete, the operator closed the upstream, rather than downstream valve and then slightly cracked it back open to complete draining. As he cracked the valve open again, there was no flow so he went on to fully open the valve. The blockage in the upper valve, which has since been assumed to be a hydrate plug, suddenly cleared and led to rapid release of propane. This splashed the operator and gave him frost burn slowing his response. When the operator tried to close the upstream valve the handle came off and could not be put back on and by the time he attempted to close the downstream valve it had frozen open. A propane vapor cloud which is heavier than air began to form and spread along the ground. The alarm was raised and traffic on a nearby motorway was stopped but despite this the plume is thought to have been ignited by a car 160 meters away from the storage vessel. After ignition the flame propagated back to the vessel which eventually blevied. Ultimately the explosion killed 18 people and injured 81. Given this disastrous event industry is now acutely aware of the consequences of hydrate formation. So given the nature of hydrates prevention is better than the cure but curing hydrates can be hard work. The best prevention is to address the factors promoting hydrate formation which recapping quickly are dew point, temperature, pressure, gas composition, mixing, nucleation sites, kinetics and salinity. For processes where you get a shot at controlling the composition of your gas stream, you have the opportunity to reduce the water to prevent hydrate formation. The lower the water content in the gas stream, the lower the dew point, so the colder you can let it get before you start forming hydrates. There's also a more practical effect, with less water available in the stream, so hydrates must form more slowly, because there's just less water to to make up those crystals. Some common technologies for removing water include TEG systems and molecular sieves. TEG systems, where TEG stands for triethylene glycol, rely on the natural affinity of TEG for water and run liquid TEG against the gas stream in a counter-current contactor. The TEG is then taken off to a reboiler where it's heated and the water is driven off. The TEG is cooled and sent back to the contactor to absorb more water. Molecular sieves are solid structures with holes that are large enough for water to fit in but too small for the methane molecules. This material catches water in the pores and is usually regenerated by passing a hot gas over it periodically. The hot gas drives the water back into the vapour and thus readies the pores to catch more water. Another way to prevent hydrates is to increase the temperature. Hydrates cannot form unless the temperature of the gas falls below the dew point of water. By keeping the temperature high it is possible to prevent this dew point from being reached. When you need to drop pressure by a large amount, like you might through a pressure control valve or a choke valve on a well, the trick is to heat the gas first. Typically a water bath is used to heat the gas up. After first heating the gas, the pressure can be dropped so the temperature falls but if the system is set up right it will stay above the dew point and hydrates will not form. Where a system such as an LPG plant is deliberately reducing pressure and temperature, there is some scope to control the system to limit hydrate formation. By By using a moisture analyzer, the dew point of the gas can be detected and the process can be controlled to avoid reducing temperature to the point which hydrates will start to form. Moving on to pressure, there are two ways that pressure reduction can help you prevent hydrate formation. Firstly, if you reduce the pressure at a location where the hydrates are likely to form, the dew point will be lower, and so you may not get the hydrates forming at all. Secondly, you can decrease the pressure upstream of your Joule-Thomson cooling. If you start at a lower pressure, then drop to your desired pressure, you will have less of a temperature drop and so you might stay out of the hydrate formation zone. So these prevention methods all address the bulk physical properties and state of the gas and are rather straightforward to implement, but for certain conditions and processes, they may not work. If this is the case, you need to move on to something fancy such as thermodynamic inhibitors, kinetic rate inhibitors and anti-agglomerates. Kinetic inhibitors can chemically depress the formation temperature of hydrates. And for this we use our old friends methanol and ethylene glycol. These are the same chemicals we mentioned before that can be used to directly break down hydrates. When using them in a preventative way, you don't need bulk quantities but a significant amount is still needed, typically 40-60% by weight of the water in the gas stream. Kinetic rate inhibitors are surface active compounds that greatly decrease the rate of hydrate formation. Using this kind of technique, you might find hydrate formation reduced to a point where you can tolerate the rate of build up until such time as they can be removed, such as during a plant shutdown. Interglomerates prevent the hydrates from combining together and attaching to fixed surfaces, allowing them to remain transportable through the pipeline. If you can keep hydrates moving, then you can take them to a location is convenient to deal with them on your own terms, such as somewhere where you can pass the stream through a water bath heater. So that brings today's show to a close. Joule-Thomson cooling can be very useful and is an important part of producing liquefied gases but it can also create hazards. One of the most common hazards associated with JT cooling is hydrate formation. Hopefully today we've shown you how you can use JT cooling to your advantage while avoiding some of the risks. As always you can check out the show notes at engineered.network where you can also find other great engineering shows such as Causality which this week discusses the BP Texas City refinery disaster. (upbeat music) (upbeat music) [BLANK_AUDIO]