Analysis of the reasons for low argon production and countermeasures

Thread Content

Analysis of the Reasons for Low Argon Production and Corresponding Countermeasures Sun Quanhai (Oxygen Production Unit, Olefin Plant, Yangzi Petrochemical, Dachang Campus, Nanjing, 210048) Abstract: In light of the fact that the argon production volume from Yangzi’s 20,000 m3/h oxygen production equipment was low under the original MPC control system, this paper uses the concepts of vapor-liquid equilibrium and relative volatility to analyze and identify the main reason for this low production level – namely, the low argon content in the argon fraction. It is proposed that the argon content in this fraction should be increased to above 10×10-2. By adjusting the control strategies and data of the MPC, the argon extraction rate was brought up to the design target of 90×10-2, resulting in annual economic benefits of over 5 million yuan. Fig 1, Table 2, Ref 2. Keywords: air separation unit, argon production, analysis, MPC. Introduction: The Yangtze 20,000 m3/h air separation unit was introduced by our company from Praxair in the United States. It adopts an internal compression process with expanded air fed into the lower column; regular packed columns are used in the upper column and the argon column, and full distillation is employed to extract argon from the air. Generally speaking, air separation units that employ the aforementioned advanced technologies should be able to achieve a high argon recovery rate. However, in reality, the argon extraction rate of this air separation unit under the original MPC control was less than 60×10-2, which is far below the design target of 90×10-2. Although argon constitutes only a small amount in air (0.932×10-2), and the various chemical manufacturing facilities in the Yangtze River region do not require it, it is a valuable product that remains in high demand in today’s gas market. Therefore, if it is possible to increase the argon extraction rate to levels close to or reaching the design targets, thereby increasing the production of this by-product, the economic benefits would be significant. Through calculation and analysis, this paper concludes that it is entirely possible to achieve an argon extraction rate of around 90×10-2 by making certain adjustments to the original operating conditions. 2 Analysis of the original operating conditions 2.1 Introduction to the argon system The schematic diagram of the argon column in the Yangtze 20,000 m3/h air separation unit is shown in Figure 1. An oxygen-aragon mixture (argon fraction) containing almost no nitrogen is drawn from the lower part of the column in the air separation unit and introduced into the bottom of the argon column; as it rises within this column, almost all of the oxygen is removed through distillation by the liquid flowing back in the column. By the time the rising gas reaches the top of the argon column, its oxygen content is only around 1×10-6. The gas at the top of the argon column, after being condensed into a liquid in the argon condenser, is sent entirely back to the column as reflux fluid for the distillation process. The liquid argon product is obtained from a level several stages below the top of the column. Let the flow rate of the argon fraction be V, and let the argon content in the argon fraction be yA ； The flow rate of the reflux liquid at the bottom of the tower is L, with an argon content of xA ； The flow rate of liquid argon in the product is D, and the argon content is xD. It is clear then that there are the following two material balance equations: …………………………………………………………(1) ……………………………………………(2) 2.2 Comparison between the original operating conditions and the design conditions By comparing the original operating data with those under design conditions, it is possible to identify the reasons for the low argon production in the original operating conditions; therefore, several parameters closely related to liquid argon production from both the original operating data and the design conditions are listed in Table 1. In Table 1, “original operating condition” refers to the average value over the 24 consecutive hours under MPC control from 0:00 to 24:00 on May 16, 2000. For ease of comparison, the air flow rate into the cold box has been increased to the design value, while the flow rates of the argon fraction, contaminated nitrogen gas, and oxygen-argon product have been adjusted accordingly based on this increase in air flow. The purity of argon products is indicated by its oxygen content, expressed in units of 10-6 ； Flow rate unit: m3/h (converted to gaseous form, the same applies below). Table 1 Comparison between the original operating conditions and the designed conditions: Argon fraction, Liquid argon product, Oxygen product, Waste nitrogen gas – Flow rate, Concentration, Flow rate, Purity, Flow rate, Purity, Flow rate, Oxygen content. Designed values: 22107, 0.156213; 800, 1; 20000, 0.998000; 13143, 0.000047. Original operating conditions: 23756, 0.065000; 503, 0.90; 19957, 0.999000; 15852, 0.001500. As can be seen from Table 1, the concentration of the argon fraction under the original operating conditions is much lower than the designed value, while the purity of the oxygen product is higher than the designed value. The oxygen content in the waste nitrogen gas is also significantly higher than the designed value. As will be explained in the following sections of this paper, it is precisely the low concentration of the argon fraction and the high oxygen content in the contaminated nitrogen gas that limit the increase in argon production. 2.3 Relationship between the concentration of the argon fraction and argon production: For an argon-rich gas with argon content of yA, its concentration becomes yA1 after passing through the first theoretical tray; since argon is a more volatile component than oxygen within the argon column, it is clear that yA1 ≥ yA ； According to the concept of theoretical trays, the vapor and liquid phases leaving a given theoretical tray should be in phase equilibrium; therefore, the oxygen-argon mixture with an argon content of yA1 should be in phase equilibrium with the reflux liquid at the bottom of the crude argon column (with an argon content of xA). If the relative volatility of argon to oxygen at the bottom of the crude argon column is αAO, then according to the definition of relative volatility, we have: ………………………………………………………（3） From (3), it follows that: ……………………………………………（4） Since yA ≤ yA1, replacing yA1 with yA in the fraction on the right-hand side of equation (4) should result in a decrease in the value of that fraction (the numerator decreases while the denominator increases, so the fraction must decrease) or no change. Thus, the following inequality is obtained: ………………………………………(5) Since the purity of the argon product, xD, is approximately 1, it is easy to determine, using equations (1), (2), and (5), that there exists the following relationship between the argon production volume D, the relative volatility αAO, the flow rate of the argon fraction gas V, and the concentration of argon in that fraction yA: ……………………………………………(6) The relative volatility of argon to oxygen at the bottom of the crude argon column is about 1.5 (in most cases it is actually slightly less than 1.5). By substituting αAO = 1.5 into equation (6), we obtain: ……………………………………………………(7) Equation (7) shows that, under normal circumstances, the argon production volume cannot exceed one-third of the total amount of argon present in the argon fraction. 2.4 Several main factors limiting argon production As can be seen from Equation (6), even without considering the factors related to the argon column itself, the argon production rate D is influenced by three factors: the relative volatility of argon to oxygen at the bottom of the crude argon column, αAO; the flow rate of the argon fraction gas, V; and the argon content in this fraction, yA. Each of these three influencing factors will now be discussed separately. Increasing the relative volatility of argon to oxygen, αAO, can improve the yield of argon, but generally, this relative volatility can only be increased by reducing the operating pressure. In reality, since the regeneration of the molecular sieve purifier requires a certain pressure of contaminated nitrogen, nitrogen compressors, water evaporation cooling towers, and similar equipment all need a certain nitrogen pressure at the exit of the cooling box. Moreover, the resistance in the nitrogen and contaminated nitrogen pipelines, as well as the resistance in the towers, is practically impossible to change during actual operation; as a result, it is almost impossible to reduce the pressure of the argon fraction. In other words, in practical operations, it is hardly possible to increase the argon yield by raising the relative volatility of argon compared to oxygen. Increasing the volume of the argon fraction gas V can also boost the argon yield, but this volume is limited by the flow capacity of the argon column; an excessive volume will degrade the performance of the column and, in severe cases, may cause flooding. On the other hand, the more argon fraction gas is drawn out from the upper column, the less upward gas flow there is above the extraction point of this argon fraction. This can lead to an increase in the nitrogen content within the argon fraction, thereby allowing nitrogen to enter the argon column. As is well known, nitrogen has a lower boiling point than argon; therefore, the nitrogen that enters the argon column cannot be removed by distillation in the same way as oxygen. As a result, the nitrogen content in the liquid argon product exceeds the allowable limits, and in severe cases this can prevent the argon condenser from functioning properly. Of course, in practice it is still possible to make full use of the design margins of the argon column and the upper column by appropriately increasing the volume of the argon fraction. This not only benefits the yield and purity of argon but also has a positive effect on the purity of oxygen (as it allows the same amount of argon to be produced at a lower concentration of the argon fraction). The main approach to increasing argon production should be to find ways to raise the argon content yA in the argon fraction. As can be clearly seen from Table (1), the volume of the argon fraction under the current operating conditions is already 7.5×10-2 higher than the designed value; it is therefore difficult to increase this value further (the design margin is generally around 10×10-2). However, the argon content in the argon fraction under the original operating conditions was only 41.6×10-2 of the designed value; clearly, it was the low argon content in this fraction that led to the low argon production. This can also be illustrated by substituting specific values into the inequality. If the relative volatility αAO is 1.5, the gas flow rate of the argon fraction V is 23756 m3/h, and the concentration of argon in this fraction yA is 0.065 (Table 1). Substituting these values into equation (6) yields… In reality, however, the argon production rate is 503 m3/h, which is already quite close to the maximum production rate that could be achieved theoretically (with no reflux and an infinite number of theoretical trays required for the argon column, assuming no tray resistance). This indicates that when the argon content in the argon fraction is low, a phase equilibrium state is nearly reached between the gas from the argon fraction and the liquid flowing back from the bottom of the crude argon column to the upper column. If the concentration of the argon fraction is increased to 10×10-2, and the volume flow rate of this argon fraction rises to 24,000 m3/h (which is roughly equivalent to the maximum flow rate allowed by the argon column), then the theoretical maximum value of the liquid argon production, without taking into account the limitations of the argon column, is as follows: Since the theoretical number of theoretical plates in the argon column is actually limited, and the relative volatility of argon at the bottom of the crude argon column with respect to oxygen is actually slightly less than 1.5, it is generally not possible to achieve a liquid argon product flow rate of 800 m3/h under a flow rate of 24,000 m3/h for the argon fraction and a concentration of 10×10-2. In other words, to achieve an argon extraction rate of around 90×10-2 (which corresponds to a production volume of about 800 m3/h at full capacity), it is necessary to increase the argon content in the argon fraction to above 10×10-2. 3 Adjustment of operating conditions 3.1 Adjustment of the upper column operating conditions As indicated in the above analysis, increasing the yield of argon is possible only by raising the argon content in the argon fraction; since this argon fraction comes from the upper column of the air separation unit, it is necessary to adjust the operating conditions of that upper column first. When the amount of air entering the cryogenic tank remains constant, increasing the amount of liquid oxygen taken from the upper column and appropriately reducing the purity of the oxygen product causes the argon-rich zone within the upper column to move downward. As a result, the argon content in the argon fraction increases, while the oxygen and argon contents in the contaminated nitrogen gas decrease. However, if the amount of oxygen product removed is increased too much, the oxygen purity will fail to meet the requirements ； For the argon fraction, if the sum of the oxygen yield and the argon yield is too high, nitrogen components are likely to end up in this fraction, which will result in an excessive nitrogen content in the argon product. In severe cases, this can cause a nitrogen blockage in the argon condenser, disrupting the operation of both the argon column and the main column. Therefore, during actual adjustments, operations must be carried out carefully after a thorough understanding of the air separation process, and a certain margin should be reserved to prevent the oxygen purity from exceeding acceptable levels or the operation of the argon column from being disrupted due to fluctuations in operating conditions caused by factors such as purifier switching. Regarding the factors affecting the argon content in the argon fraction, we established a mathematical model for the lower part of the tower (i.e., the section below the argon fraction sampling point). The model indicates that the greater the number of theoretical plates in the lower part of the upper column, the lower the pressure (i.e., the greater the relative volatility); the smaller the liquid-to-vapor ratio in this section, the lower the oxygen purity, and consequently the higher the argon content in the argon fraction. Regarding the factors affecting the nitrogen content in the argon fraction, the mathematical model we developed indicates that this nitrogen content is primarily determined by the operating conditions in the section of the column above the argon fraction sampling point and below the inlet to the evaporation air in the argon condenser. The more theoretical plates there are in this section, the lower the pressure (the greater the relative volatility), and the smaller the liquid-to-vapor ratio. As a result, the nitrogen content in the vapor-liquid phase within the column at the inlet to the evaporation air in the argon condenser is lower, and consequently, the nitrogen content in the argon fraction is also lower. Increasing the flow rates of oxygen and argon products, as well as increasing the volume of the argon fraction gas, will all result in an increase in the liquid-to-vapor ratio in this section. The air evaporated in the argon condenser, as well as the unevaporated liquid air, contain a high level of nitrogen and a low level of argon. When more of this evaporated air and unevaporated liquid air enters the upper column, it increases the nitrogen content in the vapor-liquid phase within the upper column at the inlet to the argon condenser, which hinders the accumulation of argon in the middle and lower sections of the upper column. Generally, in a compression process that uses structured packing in the upper tower and expanded air in the lower tower, the argon content in the argon fraction is higher than that in conventional air separation units where expanded air is used in the upper tower, while the nitrogen content can be lower than in the latter. We believe the main reasons for this are that the theoretical number of theoretical plates in a column using regular packing is usually high; furthermore, the liquid-to-vapor ratio in the section below the liquid-air feed inlet in the process where expanded air enters the column from below is lower, and the pressure at the lower part of the column is also lower (meaning that the relative volatility between the various components is higher). 3.2 Adjustment of the argon column operating conditions The control of the argon column’s operating conditions mainly involves regulating the amount of argon vapor produced and the quantity of liquid argon product obtained. Generally, when the concentration of the argon fraction and the amount of liquid argon product remain constant, an increase in the volume of argon gas within a certain range leads to an increase in the reflux ratio inside the argon column, resulting in improved argon purity (i.e., a decrease in the oxygen content in the liquid argon product) ； When the volume of the argon fraction gas and the argon content in that fraction remain constant, increasing the amount of liquid argon product extracted reduces the argon reflux ratio, which in turn increases the oxygen content in the liquid argon product (resulting in a lower purity of argon). On the other hand, since the argon fraction gas comes from the upper column, and the reflux liquid at the bottom of the crude argon column returns to the upper column, any adjustment to the operating conditions of the argon column will inevitably affect the upper column. Specifically, when the volume of the argon fraction gas is increased, the upward flow rate in the section above the argon fraction extraction point in the upper column decreases; as a result, nitrogen, which is the component with a lower boiling point compared to oxygen and argon, moves downward, and this may lead to an increase in the nitrogen content in the argon fraction ； When the amount of liquid argon product taken out is increased, it is easy to see from material balance that the argon content in the reflux liquid returning from the bottom of the crude argon column to the upper column decreases, and as a result, the argon content in the argon fraction also decreases. Since, with a constant liquid-vapor ratio in the lower part of the upper tower, the lower the argon content in the argon fraction, the higher the purity of the oxygen product, increasing the amount of argon product extracted will raise the oxygen purity. Therefore, as long as it is ensured that nitrogen does not enter the argon fraction, the volume flow rate of the argon fraction should be increased as much as possible within the limits permitted by the flow capacity of the argon column ； As long as the purity of the argon product remains within acceptable limits, efforts should be made to increase the amount of liquid argon obtained; this not only boosts the production volume of liquid argon but also is highly beneficial for the purity of the oxygen product. 3.3 Analysis of the data for the two test conditions Before changing the MPC control strategy and related parameters, we had already obtained several test conditions by using manual control on the DCS. Table 2 shows two of the more typical operating conditions; “Condition 1” is an operating condition designed to bring the system as close as possible to the design conditions, while “Condition 2” is one aimed at ensuring the purity of the oxygen product and maximizing the yield of liquid argon. For ease of comparison, the flow rate data in Table 2 have also been appropriately scaled based on the air flow rate, with the same units as those in Table 1. Table 2 Two typical operating conditions: Argon fraction, liquid argon product, oxygen product, and nitrogen waste gas – Flow rate, concentration, flow rate, purity, flow rate, purity, flow rate, oxygen content. Condition 1: 22100, 0.150000; 792, 1; 20010, 0.997500; 13150, 0.000045. Condition 2: 23960, 0.110000; 806, 1.20; 19996, 0.998200; 12950, 0.000050. A high argon extraction rate can be achieved in both test conditions; however, the oxygen purity in “Condition 1” is slightly lower (the design value being 0.9980). This may be due to an insufficient number of theoretical plates in the packing section below the argon fraction extraction point in the upper column, resulting in a decrease in the oxygen product purity below the design value when the argon content in the argon fraction approaches the design target. In \"Condition 2,\" by appropriately increasing the amount of argon fraction gas and reducing the purity of the liquid argon product, not only is a higher argon extraction rate achieved, but the purity of the oxygen product is also maintained. Therefore, we consider \"Condition 2\" to be an ideal operating condition, and it is also the operating condition controlled by the current MPC system. After the argon extraction rate increased, the oxygen content in the contaminated nitrogen gas decreased significantly, which implies that the oxygen extraction rate will also increase. This is because, as the amount of liquid argon taken out increases, the argon content in the contaminated nitrogen gas leaving the tower decreases; since the boiling points of oxygen and argon are relatively close to each other, the oxygen content in that contaminated nitrogen gas also decreases. In fact, a higher argon extraction rate and a lower oxygen content in the contaminated nitrogen gas are complementary to each other; the latter is both an inevitable result of the former and a necessary condition for it. 3.4 Changes to MPC data: Both of the above test conditions were achieved with the MPC (Model Predictive Control System) not in use, and adjusting to such conditions requires certain technical skills. At the same time, since the oxygen content in the contaminated nitrogen gas is one of the factors that contribute to the stability of the air separation distillation tower, a decrease in the oxygen level in this gas can lead to reduced stability in the distillation tower system. This requires the system to be able to respond promptly to various disturbances; otherwise, it may result in an oxygen purity level that exceeds the specified limits. The test results only demonstrated that the argon extraction rate of this air separation unit can reach around 90×10-2, which is the specified design value. However, to maintain such a high argon extraction rate and to make the above two operating conditions feasible in practice, reliable control methods are necessary as a guarantee. The MPC (Model Predictive Control System) used in the Yangzi 20,000 m3/h air separation unit has fairly reliable stability; however, it originally had two significant drawbacks: first, a low output volume of liquid argon could lead to a decrease in oxygen purity, and second, the argon recovery rate was low. We believe that the issue of poor oxygen purity stability is related to the original MPC control strategy. The MPC developed in September 1999 by Praxair experts primarily uses the method of controlling the middle section of the tower to regulate oxygen purity. However, when the amount of liquid argon extracted is low, argon accumulates in the lower part of the upper column; at this time, the oxygen content in the middle section of the upper column usually remains unchanged or changes only slightly. As a result, the MPC cannot receive timely signals indicating that the oxygen purity is about to decrease, and no corresponding adjustments are made, which often leads to a sudden and significant drop in oxygen purity. During the ten-month period from early October 1999, when MPC was officially put into use, to early August 2000, before we modified the MPC data, there were several instances in which the oxygen purity exceeded acceptable levels for this reason; each time such an excess occurred, 50 to 100 tons of liquid oxygen were wasted. At the beginning of August 2000, we changed the MPC to control oxygen purity by regulating the oxygen content in the argon fraction, which effectively solved the problem of poor stability in oxygen purity. We believe that the level of oxygen content in the argon fraction is more closely related to the purity of the oxygen product at the bottom of the column; before any factor causes a decrease in oxygen purity, the concentration of the argon fraction changes first. This allows the MPC to receive a signal in a timely manner, enabling it to take appropriate corrective actions to prevent a drop in oxygen purity. After the purity of the oxygen product was reliably ensured, we modified some of the control parameters within the MPC to reduce the oxygen and argon content in the contaminated nitrogen gas, resulting in an argon extraction rate that increased by approximately 60×10-2 compared to before. The operational results over the past six months and more have proven that our changes to the MPC data were successful: not only was the argon extraction rate significantly improved, but the stability of the MPC also became even better than before, and since then the oxygen purity has never exceeded the specified limits again.