If you are unaware of the Joule–Thomson effect, but have some technical knowledge, you might conclude that the temperature is unchanged by the throttling process. There appears to be nothing happening to change the temperature of the fluid. This view is incorrect, except in a few relatively rare cases.

On the other hand, it should be clear to all that no work is done as the fluid crosses the valve. The potential work available from going from high pressure to low pressure is simply lost.

Those with some technical training might say that the fluid cools upon such an expansion and that the temperature leaving the valve is lower than that entering.

Mathematically this is given by:

When this quantity is positive, the fluid cools upon expansion, and when it is negative, the fluid warms.

From classical thermodynamics, we can derive the following expression for this quantity:

The derivation of this expression is given in most textbooks on classical thermodynamics and will not be repeated here.

We have reduced the expression to one that contains only the pressure, temperature, molar volume (and derivatives of these quantities), and the heat capacity. An equation of state of the form P = f(v,T) can be used to evaluate the numerator.

The detailed thermodynamics of the Joule–Thomson expansion are well understood and will not be repeated here. Those interested can check almost any book on classical thermodynamics or the paper by Carroll (1999).

Substituting this expression into Eqn (8.2) reveals that the Joule–Thomson coefficient is zero. Thus, for an ideal gas, the isenthalpic expansion does not affect the temperature. However, for real gases and for liquids, that is not the case.

There are three cases where negative Joule–Thomson values may be encountered in engineering practice. These are:

In horizontal slurry flow, there are four flow regimes: (1) homogeneous flow, (2) heterogeneous flow, (3) saltation flow, and (4) flow with a stationary bed (Turian and Yuan, 1977).

In homogeneous flow, the slurry flows almost as a single phase with no separation into either phases or layers. This occurs when the settling velocity of the particles is small in comparison with the velocity of the fluid.

If the particles are somewhat larger or if the settling velocity of the particles is larger, then vertical gradients in the particle concentration can occur. This is the heterogeneous flow regime.

If the settling velocities become too large, then the particles with start to settle. The initial settling is the saltation regime. This regime is intermittent, with leaping movement of particles. Typically, the solids are transported by two mechanisms: bulk movement of the particle bed and movement with the fluid above the bed. To visualize this, imagine wind blowing sand off the top of a sand dune.

In the stationary bed regime, the particles have settled to the bottom of the pipeline. This is similar to the bedding of hydrates described below. In this case, the particles will only flow if they are swept along with the fluid flow.

The criteria for determining the flow regime is a function of the drag coefficient, the solids concentration, the velocity, the pipe diameter, the particle size, and the ratio of the densities. Different pressure drop correlations were developed for each flow regime.

In the slurries studied by Turian and Yuan (1977), there was no tendency for the particles to adhere. For example, in a coal-water or a coal-oil slurry, the coal will not agglomerate into a large mass. However, this is indeed the case with hydrates, thus, for hydrate slurries.

In addition, in typical slurries, the solid phase is denser than the liquid. For natural gas hydrates, if the liquid phase is water, then the hydrates are most likely less denser than the water, and will have some buoyancy and will not likely settle to the bottom of the pipe. If the fluid phase is a condensate or an oil, typical density less than 800 kg/m³, then the hydrate will be of greater density, more like the traditional slurry.

The flow of the gas through the reservoir is governed by Darcy's law (or nearly so). For the radial flow of a gas, Darcy's law is (Smith, 1990):

Furthermore, when this equation was integrated, it was assumed that the properties of the fluid do not change significantly with the changes in the pressure and temperature of the fluid.

In addition, moving from a depth to the surface, the surrounding temperature decreases the so-called geothermal gradient. Typically, the geothermal gradient is approximately 25 °C/km (1.5 °F/1000 ft).

Fig. 11.3 shows the pressure and temperature as a function of the depth for a hypothetical gas well. The well has a depth of 3000 m (9842 ft) and the pressure at the bottom of the well is 15 MPa (2176 psia), which includes some pressure drop because of the flow through the reservoir. The temperature at 3000 m is 75 °C (167 °F) and the geothermal gradient is 25 °C/km.

Fig. 11.4 shows the same data but on a pressure-temperature plot. Also shown on this graph is the hydrate curve for this gas mixture. It can be seen that the pressure-temperature curve intersects the hydrate curve, which means a hydrate can form in the wellbore. The intersection of the two curves is at approximately 19 °C (66 °F), which is equivalent to a depth of about 750 m (2460 ft). From the surface to about 750 m, hydrate formation is possible in this well. For depths greater than 750 m, the pressure-temperature of the gas is outside the hydrate region and freezing will not occur at greater depths.

In this situation, the injection of an inhibitor chemical is required. Some wells are completed with an injection string that allows for the injection of chemicals at the bottom of the well. This injection is typically continuously preventing production interruption because of plugging of the wellbore. Other wells have a pump at the surface for the injection of chemicals. Unlike pipelines, which are more or less horizontal, the injected chemical can flow down the well string, even if the flow is blocked, because of gravity.

A number of depleted gas reservoirs located in northern Alberta, Canada, and elsewhere, have pressure and temperature conditions that are identified as potential sites for CO₂ storage in gas hydrate form. These reservoirs are in the range of pressure from 2 to 5 MPa (290–725 psia) and temperatures from 1° to 10 °C (34°–50 °F). The CO₂ must flow into the reservoir for a sufficient distance and then solidify into a hydrate. If the gas solidifies in the wellbore region, the flow would be impeded.

This technology is very new and has only been studied in the laboratory and in simulation (Linga et al., 2009; Zatsepina and Pooladi-Darvish, 2012).

For long distance transmission lines, pressures are typically 1000 psia or about 7 MPa. In addition to the pipeline, this involves compression of the gas and, depending upon the distance of the line, may require intermediate recompression.

Transportation of liquid natural gas (LNG) is done at very low temperature but at near atmospheric pressures. It is expensive to liquefy natural gas and requires a vaporization facility at the delivery point.

The industrialization of the transportation of natural gas in the form of hydrates is in its infancy, with several problems that still must be worked out. However, this is not a new idea. Amongst the first to suggest this technology were Cahn et al. (1970).

One of the problems with this method is to produce a hydrate with minimal entrained water, ice (which contains no gas), and a high hydrate concentration. Gudmundsson (1996) describes one possible method for making hydrates on a large scale.

Table 11.1 gives a comparison of the amount of methane contained in 1 m³ (35.3 ft³) of hydrate, 1 m³ of pipeline gas, and 1 m³ (35.3 ft³) of LNG. All of the values are for pure methane and the densities of methane are taken from Wagner and de Reuck (1996). One cubic meter of hydrate contains more than twice as much methane as the pipeline gas, but only about one-fourth of the methane in LNG.

These hydrates are found throughout the world and not only in colder locations. For example, hydrates have been found in the Gulf of Mexico, the Caribbean Sea, off the cost of South America, India, and many other locations.

Estimates of how much hydrocarbon is locked in these resources are astronomical. For example, Dickens et al. (1997) estimated that there is about 15 GT7 equivalent of methane in the seabed tied up as natural gas hydrates. This is approximately equivalent to 3 × 10¹³ m³[std] or 1 × 10⁹ MMSCF of methane. However, such estimates range widely depending upon who is doing the estimate and what is their basis. For example, Kvenvolden (1999) estimates this amount of gas in seabed hydrates as between 1 × 10¹⁵ and 50 × 10¹⁵ m³[std]. Lerche (2001) provides an interesting analysis of the estimates of the hydrate resource.

Others, such as Laherrere (1999), argue that the hydrates are of poor quality and may not be of commercial quality.

The production of the gas locked in the hydrates of such a reservoir is an interesting problem. Perhaps a thermal recover method, such as those used for heavy oil, could be used to melt the hydrate. Alternatively, perhaps a miscible flood of an inhibitor chemical could unlock the gas. There have been several papers discussing the production of natural gas from hydrate reservoirs by depressurization. One such paper is by Jia et al. (2001).