paper title - Analytic Expert

Indal├ęcio Caldeira | Download | HTML Embed
  • Mar 4, 2011
  • Views: 25
  • Page(s): 5
  • Size: 611.16 kB
  • Report

Share

Transcript

1 The Impact of Hydrocarbon Dew Point on Sample Conditioning for BTU Analysis Shane Hale, Emerson Process Management Introduction No matter how good the analyzer is, the results can only be as good as the sample it analyzes. If the composition of the gas changes between the sample point and the analyzer, the analysis and the resulting calculations (such as energy flow) will not be representative of what is flowing through the line. The most common cause of the composition changing in the sample handling system is related to the hydrocarbon dew point (HCDP) of the sample. If at any stage the natural gas is allowed to go below the HCDP, the heavy components will drop out of the gas phase (and into the liquid phase) and the composition of the gas will change. Understanding the impact of HCDP and sample composition is critical for designing sample handling systems. Understanding the different methods for determining the HCDP and using it in practical situations can provide operational tools to improve measurement and protect against sample handling issues. Determining the HCDP The first step to designing a sample handling system is determining what the HCDP of the sample will be. Typically, an equation of state is used to determine what the theoretical HCDP of a composition is. When using an equation of state to determine the HCDP, the composition used is very important. Often the typical or nominal composition is used to design the sample system. However, the system should be designed for the worst case scenario. An analogy is the seat-belt. The seat-belt is designed to perform correctly when things are going wrong, not when everything is going well. The sample system should also be designed to perform when an upset or abnormal condition results in dirty or heavy sample and protect the analyzer while maintaining the composition so the abnormal mixture is correctly reported. The API 14.1 standard recommends that the natural gas sample should be maintained at least 30F above the expected HCDP (API 14.1, 2006, Section 6.6). Users should use the gas with the highest HCDP as a reference point and then design the system to maintain the sample 30F above that. Many people assume the HCDP tracks the BTU content. Figure 1 shows that for a large sample of more than 1000 natural gas analysis there is a loose relationship of higher cricondentherm with BTU; however, the relationship is not strong enough to use as a basis of design of the sample system. Figure 1 - BTU and Cricondentherm chart showing a loose relationship of BTU to the cricondentherm. Pipeline Quality Gas Transmission For pipeline quality gas transmission measurement, the large number of interconnects in the transmission system means the original source of the gas is often unknown and impossible to determine. A commonly used

2 tariff specification for HCDP in the United States is 15F (-9.4C), so one could assume this as the HCDP for the worst case scenario. However, this tariff is not used in all locations, is often not continuously monitored and thus not enforced, and producers are usually given relief to keep injecting off-spec gas while they bring their process under control. These factors result in the probability that gas with a HCDP greater than the tariff specification will be sampled at some stage. By definition, the HCDP of the gas in the pipeline must be at or below the flowing gas temperature and sample probes should be designed and located to only sample the gas phase. Using the flowing gas temperature is therefore a better basis to design the sample system to account for the car-crash occasions when the sample goes outside of the normally operating conditions. Most pipelines will have flowing temperatures lower than 90F (32C), so this can be used as the expected highest HCDP gas. Adding 30F to this temperature will result in an API 14.1 compliant sample system temperature of 120F (~50C). If pipeline temperatures greater than 90F (32C) are expected, the sample system design temperature should be altered to suit. Raw Production Gas The raw gas directly from the well can be significantly richer and contain more contaminates than pipeline quality gas. Very often, the stream is two-phase or very close to the HCDP. Special attention must be paid to removing the sample with as little liquid hydrocarbon as possible and removing the liquids at the same pressure and temperature as the flowing stream. If the temperature or pressure of the sample is changed prior to removing the liquids, then the composition of the gas phase will change and make any resultant analysis unrepresentative of the stream gas. The easiest way to achieve this is to reject the liquids at the inside the stream using sample probes with a liquid-membrane tip. Measured versus Theoretical HCDP The determination of the HCDP falls into two distinct methods. The measured HCDP method relies on cooling the gas sample until the formation of liquids can be detected in a chilled-mirror analyzer. The theoretical method uses the composition of the gas and an equation of state (EOS) to calculate the theoretical HCDP. With the chilled-mirror method, the formation of hydrocarbon liquids is detected on a mirror either by an operator (a manual HC dew-scope) or light-based detector technology. The accuracy of this method is determined by the rate of cooling of the mirror, and the repeatability of the hydrocarbon mist detection. The results of manual dew- scopes are heavily dependent on the experience (and patience) of the operator, but the latest generation of automated methods are highly repeatable. Figure 2 - Cricondentherm pressure for samples with cricondentherm between 0 F and 40 F using the PR EOS. The chilled-mirror is a physical measurement so will provide only the HCDP at the pressure at which the measurement is conducted. Typically the pressure on an automatic chilled-mirror system will be set at 400 PSI (2,757 kPa) that approximates the usual cricondentherm pressure (Figure 3) for pipeline quality natural gas so the results are close to the calculated cricondentherm. However, there is no method to extrapolate this result to other pressures (such as the pipeline pressure or delivery pressure). Additionally, a visible mist must be formed on the mirror for the detection circuit to register the dew point, which means that liquid hydrocarbons have already

3 dropped out before the detection circuit is tripped. For this reason, the results of the physical measurement of HCDP will be lower than those calculated from the theoretical models. Another practical factor for consideration is that a chilled-mirror device is a dedicated analyzer for a single measurement and the user must determine if the additional cost of the analyzer and associated sample handling system can be justified. An alternative is to use the theoretical EOS method to calculate the HCDP using the composition of the gas. The composition is almost always reported using a gas chromatograph (online or offline) for the determination of gas quality, energy, compressibility, and other physical properties, so extending the use of the composition for the determination of HCDP can be an attractive option. Several equations of state (EOS) can be used for calculating the theoretical HCDP for natural gas mixtures with the most common being the Peng-Robinson (PR) and the Soave-Redlich-Kwong (SRK). The SRK equation will tend to calculate values about 5 F (3 C) higher than the PR equation (refer Figure 3), but the PR is generally accepted to calculate values more representative of typical natural gas mixtures. Figure 3 - Difference between the cricondentherm from the SRK and PR Equations of state for samples between 0 F and 40 F. The advantage of using the theoretical method is it allows the calculation of the cricondentherm temperature and pressure as well as the HCDP at any pressure. Gas chromatographs are available that will calculate the HCDP at four user set pressures and the cricondentherm temperature and pressure. However, the level of analysis is very important to determining the HCDP with reasonable accuracy. The typical gas chromatograph used in natural gas custody transfer applications performs as C6+ in which the individual hydrocarbon species up to normal-pentane, nitrogen, and carbon dioxide is measured, and any hydrocarbon larger than n-pentane is lumped together as C6+. For the determination of energy content and other physical properties, the C6+ value is then mathematically split into fractions of hexanes, heptanes, and octanes. This is acceptable for energy content, relative density, compressibility, and the other typical calculations as the C6+ fraction is very small (less than 0.1%) and using fixed assumptions for these components results in very small errors. However, the concentration of the heavier hydrocarbons dramatically affects the hydrocarbon dew point value as it is the heavy components that will drop out into the liquid phase. Using fixed ratios for the determination of the C6, C7, and C8 concentrations will result in large errors the HCDP calculation, well beyond practical use (Error! Reference source not found.). Extended analysis gas chromatographs capable of measuring the heavier hydrocarbon species permit a much more accurate calculation of the HCDP. In many laboratories, measurement of the hydrocarbons up to C12 or even C14 is conducted using a Thermal Conductivity Detector (TCD) for the C1 to C5 measurement, and a Flame Ionization Detector (FID) for the heavier hydrocarbons. The analysis cycle time can be from 10 minutes to over 40 minutes that is usually not practical for online use in custody transfer where the typical C6+ analysis is between three and five minutes.

4 Figure 4 - The HCDP calculation errors from using a C6+ fixed ratio split to determine the C6, C7 and C8 values. Online gas chromatographs that use parallel analysis paths, dual TCDs, and analysis times of five minutes to measure up to C9+ have been available for more than ten years. This level of extended analysis provides an alternative to the poor HCDP determination using the C6+ analysis, and the excessively long analysis times for C12 or C14 analysis found in the laboratory. Using the C9+ analysis to determine the HCDP results in a more accurate and consistent values that can be used practically in the field. Further accuracy can be achieved using fixed fractions of C9, C10, and C11 for the C9+ component that can be based on experience or routine extended laboratory analysis. Using HCDP in Operations Whereas the main purpose for HCDP determination is for monitoring the gas for tariff specifications, the online calculation of the HCDP can also provide useful diagnostics for measurement operations. Flow meters do not measure the flow accurately if there is two-phase flow. The drop-out of heavy hydrocarbons causes loss-and- unaccounted-for errors across the pipeline as the energy value of the gas exiting the pipeline is less than the energy value of the gas entering the network. Liquid hydrocarbons can result in hydrate formation and other operational issues. Determining the HCDP online provides an effective tool to mitigate these issues. Using the stream pressure as one of the pressures for the calculation of the HCDP (through an analog input or Modbus serial link) provides a useful indicator of potential problems. For example, if the calculated HCDP is the equal to or higher than the pipeline temperature, it is a sign that there are liquid hydrocarbons in the pipeline that should be addressed by operations. Figure 5 shows a real-world example of where the stream was two-phase when the sample was extracted. In this situation, the sample probe did not exclude liquids, and the sample was heated prior to the pressure reduction. The reported composition will be incorrect as some liquids have been

5 sampled and then vaporized by the sample system, resulting in a sample that is analyzed by the gas chromatograph richer than the gas phase in the pipeline resulting in an over-reporting of energy content. Figure 5 - Example of a sample taken when the stream was two phase. The incorrect composition will result in errors in the energy and other physical properties calculations leading to incorrect fiscal reporting. Additionally, the flow-meter may also read incorrectly because of the interference of the liquid hydrocarbons in the measurement (depending on the technology used). Modern ultrasonic meters can determine when there is two-phase flow from the flow profile and other diagnostic indicators and provide an alarm. However, trending the HCDP from an online GC allows operations to determine when the HCDP is being approached and take measures to avoid the issue escalating to a two-phase flow condition (such as shutting in a rich gas supply). Conclusion When designing a sample system, knowing what the HCDP of the sample can be expected to be will help determine what components are required to ensure a good sample reaches the analyzers. The online analysis with an online HCDP analyzer provides an accurate and repeatable determination for HCDP for use in reporting gas quality tariff values. However, the determination of HCDP at the pipeline pressure and the cricondentherm can only be done using the theoretical EOS models. The typical C6+ compositional analysis used in natural gas custody transfer is not sufficient to calculate the HCDP with enough accuracy, as the heavy components higher than normal-pentane affect the HCDP dramatically and assuming them from a fixed ratio of the C6+ value introduces excessive errors. The extended C12 or C14 analysis available in the lab provide accurate results, but the extended times are impractical for use in online custody transfer. Using an online C9+ analysis provides a compromise that permits an accurate determination of the cricondentherm and the HCDP at flowing conditions. Using the HCDP value at flowing conditions and comparing to the stream temperature can determine if the sample handling system is changing the composition of the gas, can be used to avoid two-phase flow, and thus avoid measurement and operational issues.

Load More