Technical Survey Of Steam Reformer To Control Tube Metal Temperature.

Tube metal temperatures (TMT) of many reformer tubes were higher than design limit, while methane slip and other process conditions of steam reformer were satisfactory in an ammonia plant. About 50 MTPD of ammonia production was lost, because of high TMT. An in-depth survey carried out in the steam reformer identified the root cause of the problem and explained how it contributed to high TMT. After implementing the recommendation of the survey, the plant is running normal. This paper includes the details of the problem experienced, the survey carried out and recommendations to overcome the problem of high TMT in the steam reformer tubes.

P. Ravichandran and Dr. Ijaz Ahmed Chaudhry Saudi Basic Industries Corporation (SABIC) catalyst of two different sizes are used in this reformer, with smaller rings at the top one third of the tube. Total volume of the catalyst charged in February 2002 was 16.1m3 (568cft). Natural gas is used for firing in the reformer in 162 arch burners arranged in nine rows. The reformer is equipped with eight of 92 mm (3.62 inches) inner diameter risers.

Introduction

ABIC has five ammonia-urea complexes functioning in the Eastern Province of Saudi Arabia. One of the ammonia plants with a name plate capacity of 1000 MTPD was commissioned in March 1983. It is based on conventional process and herein after called Ammonia Plant Jubail (APJ). Due to modifications carried out in stages over time, the ammonia production capacity has increased to 1300 MTPD. Natural gas supplied by Saudi Aramco is used as both feed and fuel stock. Steam Reformer in APJ is a typical top fired reformer with 416 reformer tubes arranged in eight rows. Reformer tubes, made of HK 40, with an internal diameter of 71.1 mm (2.8 inches) and thickness of 15.4mm (0.6 inches) were installed in 1997. Non alkalized nickel

High tube metal temperature

Tube Metal Temperatures (TMT), also known as tube skin temperature, of reformer tubes are measured on a regular basis in APJ. Measurement of TMTs is very useful not only in evaluating the process conditions of a reformer, but also in preventing premature failure of tubes. Reformer tubes are very costly and are generally designed for a life time of 100,000 hours.

2010

283

AMMONIA TECHNICAL MANUAL

The life of reformer tubes will reduce significantly, if the tubes are operated above design TMT level. In the third quarter of 2003, it was noticed that the TMT of reformer tubes started increasing. TMTs measured were in the range of 910oC to 960oC, versus the design value of 910oC. Hot bands were also noticed on many reformer tubes, not less than 5%, between two and five meters measured from the roof of the furnace. At the time, all process conditions at inlet and outlet were normal. Approach to equilibrium calculated for the catalyst of primary reformer was also normal. Production rate had to be decreased from 1300 MTPD by about 50 MTPD, because of high TMT. During the same period, the composition of the natural gas supplied to the plant was becoming leaner (less carbon). Typical composition of natural gas received compared with lean gas is given in table-1. Description Normal Gas Lean Gas Methane 77.30 89.61 Ethane 16.04 3.92 Propane 3.00 0.48 Butane 0.77 0.16 Pentane 0.15 0.06 Hexane 0.00 0.02 Nitrogen 2.23 5.45 CO2 0.51 0.30 Total 100.00 100.00 Mol Weight 19.95 17.56 C atom/ mol 1.227 1.0025 H2 mole/ mol 2.19 1.94 3 LHV, Kcal/nm 9974 8436 Table-1 Natural Gas Composition, mol % Change in the concentration of different components in natural gas and expected consequences are discussed in the following paragraphs:

1) Less Carbon Number of carbon atoms per mole decreased by 18%. Natural gas with such lower carbon content compared to normal level is known as lean gas and that with higher carbon content as rich gas. Natural gas supply became lean, because higher hydrocarbons (C2+) were recovered by the supplier. The lower content of higher hydrocarbons and the higher content of nitrogen reduce the quantity of hydrogen achievable from one nm3 of natural gas because less moles of hydrogen are produced per mole of feed natural gas. For same feed rate of carbon atoms, leaner natural gas requires more heat for reforming and higher reforming heat leads to higher tube skin temperature of reformer tubes. This is true with or without Pre reformer. 2) High Nitrogen Nitrogen content of the gas had increased by 144%. Nitrogen, one of the reactants for ammonia synthesis, is added in the form of air at Secondary Reformer. Thereby, oxygen required for the reforming of remaining methane in the Secondary Reformer is simultaneously supplied. The consequence of higher nitrogen content in natural gas is reduction in the air requirement, for the same methane slip. Reduced air input to Secondary Reformer shift the reforming load on Primary Reformer. By this way, capacity of the plant is reduced, if Primary Reformer is already running at its maximum capacity. 3) Reduced Heating Value (LHV) Since nitrogen is an inert as far as combustion process is concerned, increase in nitrogen content reduces the heating value of fuel natural gas. Heating value of natural gas is decreased by 15% which it contributes to; a) an increase in

AMMONIA TECHNICAL MANUAL

284

2010

the amount of fuel required, b) an increase in the amount of flue gas flow and c) an increase in the temperature of flue gas. Also the increased fuel flow rate may limit the fuel header and burner system. Overall, the above changes contribute to the following, as natural gas is used as feed and fuel stock: Increased flows of feed and fuel natural gas. Increased heat load on primary reformer. Increased flow and temperature of flue gas.

an elevation corresponding to 3.05m (10feet) from the roof. For the bottom portion, TMTs were measured only on the west side (no peep holes on the east side) at an elevation corresponding to 8m (26.25feet) from the roof. A detailed Tube Wall Temperature survey was carried out using infrared optical pyrometer, Cyclops Land Model 52 and Gold Cup pyrometer. The measurements by Gold Cup equipment indicated TMTs higher than that indicated by optical pyrometer. Optical pyrometer readings need to be corrected to eliminate background radiation effects. So, temperatures of the background that includes furnace walls, roof and floor/flue gas extraction tunnels were also measured. For TMT correction purposes, a tube emissivity of 0.85 was used. Statistical analyses of the corrected temperatures were carried out. The set of TMTs indicated by optical pyrometer, with necessary corrections, had excellent match with the simulated value for the conditions prevailing at that time of operation and hence, only optical pyrometer readings were taken into consideration. Table-3 given below summarizes the temperature spreads observed at the top portion of reformer tubes, after correction for background radiation. Row # 1 2 3 4 5 6 TMT Deg C Avg. 884 875 881 862 848 862 Max. 908 926 920 883 895 903 Min. 860 822 850 822 821 828 Spread 49 104 70 61 74 76

Hot Bands
There were a few unscheduled shut downs in the plant. A remarkable observation made immediately after plant start up was that reformer tube hot bands disappeared. However, after about two months of operation, hot bands reappeared.

Reformer survey:
In order to determine the reason for high TMTs noticed in many reformer tubes and to sustain the production rate at 1300 MTPD level with lean natural gas, a detailed technical survey was conducted in the reformer on January 31 and February 01, 2005. Services of M/s. Johnson Mathey were hired for carrying out the survey. The reformer survey (Ref 1) included the following tasks: a) b) c) d) e) Measurement of reformer tube TMTs Process data collection and simulation Visual inspection Root cause determination Recommendations

Measurement of TMTs TMTs of the top portion of the reformer tubes were measured through peepholes available on both east and west sides of the furnace box, at

2010

285

AMMONIA TECHNICAL MANUAL

7 849 870 833 36 8 862 925 813 112 All 865.4 926 813 113 Table-3 Tube Metal Temperature - Top portion There were nine tubes with TMT above design level of 910oC, while seventeen tubes with TMTs 900oC and above. TMT spread (difference between maximum and minimum TMT) noticed was 113oC. It is clearly depicted in the plot for top level TMTs (Figure-1), which shows a cold zone at the north east corner of the furnace and a hot zone near rows 1 and 2 in the middle section of the furnace. TMT spread at the top elevation is quite high and it is due to the presence of hot bands.

3 820 832 4 5 817 831 6 808 814 7 820 833 8 828 855 All 816 855 Table-4 Tube Metal Temp. -

806 26 797 34 802 12 803 30 791 64 784 71 Bottom portion

Process data collection and simulation When the operating conditions were steady, a set of ammonia plant front end data was collected and consistency of the data checked. A computer simulation model for the radiant section of Primary Reformer PRIMARY (Ref 2) uses the corrected process data. Geometrical configurations of the radiant furnace, reformer tube layout, size and shape of catalyst, etc. were other inputs to the model. The PRIMARY program used in this survey incorporates reaction kinetics, correlations with experimental heat transfer data and advanced furnace calculation methods, that take into account many variables involved in the physical, chemical, geometrical and mechanical properties of the system. So, the whole steam reforming process, including TMTs along the length of the tube, can be simulated accurately. TMT value predicted by the simulation program at the top level was 858.6oC, which is very close to average value of 865.4oC, measured by optical pyrometer after necessary correction. Performance of convection section coils was checked using another simulation program and found to be satisfactory. Visual inspection

Figure-1 Corrected TMTs at the top portion. TMTs observed at the bottom portion are given in Table-4, after necessary correction. Average temperature at the bottom level was 816oC with a TMT spread of 71oC observed at the bottom level. In Row 4, the last tube was located in front of the peephole door, blocking the view for the other tubes in that row. Row # 1 2 TMT Deg C Avg. 815 800 Max. 827 811 Min. 795 784 Spread 33 28

Visual inspection was made through peep holes, available at the top and bottom. The following observations were made in the radiant section:

AMMONIA TECHNICAL MANUAL

286

2010

At the very top section of reformer tubes (up to 2m or 6.6feet from the roof), all the tubes appeared normal and there were no hot bands. All those patches noticed in the reformer tubes were seen between 2m and 5 m (6.6 and 16.6 feet) from the roof (except the tube #15 in row no. 4 which had a long hot band of ~3 m running up to 50% of the tube ht. from top). Flames and burners were generally good. Though there were a few burners in the fourth and fifth rows having slightly longer flames than others, there was no observed impingement on the tubes. Most of these tubes hot bands do not appear to have been caused by flame impingement. No unusual movement of the flue gas.

shut-down, but reappeared after a few months of operation. This means during normal continued operation of the plant, carbon formation occurs, whereas during shutdown/startup, carbon is removed due to steaming step. Endothermic reaction for the formation of carbon from methane is given below: CH4 C + 2H2 In the above equilibrium reaction, forward reaction forming carbon is favored at high temperature and low partial pressure of hydrogen. Considering various zones of reformer tubes:A) In the top 25 % of the tubes of most reformers, process gas is heated up and temperature is not high enough to have a net carbon forming region. B) In the bottom half of the tube in most reformers, good amount of reforming reaction is already done and as a result hydrogen partial pressure is too high to have a net carbon forming region. C) But, in the 25-50% region, it is very likely to have both conditions favoring net carbon formation namely, high temperatures and low hydrogen partial pressure. This is due to high heat load on the furnace and/or insufficient activity of the catalyst and hence not yet sufficient hydrogen present or produced. In the case of APJ, the location of the hot bands fell in the last category zone discussed above. Methane slip and calculated approach to equilibrium of the reformer catalyst were within the acceptable limits. This implies the overall activity of the catalyst, especially the catalyst at the bottom half portion of reformer tube is active. However, the activity in the top 50% might have deteriorated to such an extent that

Convection section could not be inspected, as the plant was running during the reformer survey period. However, during turnaround in July 2005, visual inspection through inspection doors was done and it revealed that there was no abnormality in the convection section. Slight deposition of burnt salts was found on the external surface of mixed gas feed preheating coil. Root cause determination Utilizing the TMT measurements and computer simulations, detailed analysis of the problem was done and the following paragraphs explain the theories involved and the root cause of the problem: Carbon formation It was observed that the hot bands noticed on the reformer tubes had disappeared after every

2010

287

AMMONIA TECHNICAL MANUAL

the generation of the reforming products is slightly less than expected. This in turn leads to the hydrocarbon-bearing reactants moving into the zone where the temperature of the process gas is taken to high enough levels that carbon forming reaction predominantly takes place and this apparently is seen by way of hot bands.

increased considerably. Conditions mentioned in the above table for APJ are for 1270MTPD. For a production rate of 1300TPD, catalyst loading will be much higher than 7.5. On the other hand, for the original name plate capacity of 1000MTPD, the catalyst loading was 6.2 only. As Primary Reformer feed rate or catalyst loading is increased,

Catalyst loading Further, the survey identified that the root cause of the problem is increase in Catalyst Loading, which is defined as the ratio of total wet gas flow (kg/hr) to total volume of catalyst (litre) charged in Primary Reformer. A table comparing the catalyst loading in different ammonia plants around the world with steam reformers is given in Table-5. Plant C with Pre-ref a. b. c. sensible heat load increases reaction heat load increases residence time reduces

Quantity of heat transfer is a product of heat transfer coefficient, area and temperature difference. For the conditions of reformer tube under discussion, as the heat load increases, the temperature difference or the driving force increases. Hence, for maintaining certain process gas temperature in reformer, TMT increases. As discussed earlier, the higher the tube wall temperatures in the 25-50% region, the more the carbon forming reaction is favored. As residence time reduces, partial pressure of hydrogen at any given point will reduce, which also favors carbon formation. So, in any reformer, if the catalyst loading is increased beyond certain level, hot bands are expected to occur.

Detail

Plant A

Plant B

APJ with Pre-ref

Catalyst volume, 21.6 22.7 20.6 16.08 1000 lt. S/C 3.9 3.4 3.5 3.3 Ratio Catalyst Loading, 5.76 6.12 5.7 7.5 Kg/hr/lt. Table-5 Ref catalyst loading in different plants. As per the above table, the catalyst loading on the reformer is the highest in APJ for the reformers of comparable size, with natural gas as raw material. Catalyst loading is an important factor that is causing the carbon formation in the catalyst, leading to hot patches. As the leaner gas contains lower carbon and hydrogen contents, natural gas feed rate and hence total wet process gas entering Primary Reformer

Recommendations
Recommendations of the survey are given below: 1) Balancing fuel firing in primary reformer Reformer tube TMT spread is an indication of how well balanced the reformer firing is. In general, TMT spreads can be classified into the following categories:

AMMONIA TECHNICAL MANUAL

288

2010

a) Less than 50C - A well run furnace with little room for improvement. b) Between 50 and 100C - An average furnace with some room for improvement. c) Greater than 100C - Needs improvement. In order to reduce the high TMT spread observed at the top peep-hole level, it is necessary to balance fuel firing in the reformer. Balancing in firing could be achieved by gagging the isolation valves of burners close to hot tubes and un-gagging the valves of burners close to Cold tubes. 2) Large size reformer tubes In order to solve the root cause of high catalyst loading, the volume of primary reformer catalyst charge needs to be increased. This can be achieved by using thinner wall reformer tubes made of HP micro alloy, which has improved creep rupture strength and it can withstand high tensile stress. Since the temperature difference between the inside and outside surfaces is smaller in a thin walled tube than it is in a thicker one, the thinner tube can run with a greater margin between the normal operating TMT and the design limit. 3) Tubes inspection It is also recommended that a detailed diagnostic tool, such as LOTIS technique, be used to inspect integrity of the current tubes, while planning catalyst replacement.

Availing the shutdown opportunity in July 2005, catalyst replacement was done in 126 reformer tubes (total 416). Eddy current measurements were done and defects in a few reformer tubes identified. Totally eight reformer tubes were replaced and loaded with fresh catalyst.

After the start up, reformer tube TMTs were observed to be less than 900oC, as against the design value of 910oC for the same throughput at 1270 MTPD. In parallel, necessary efforts were taken to install new reformer tubes. During the turnaround in July 2008, larger diameter tubes made of HP micro alloy were installed with a fresh charge of non alkalized reformer catalyst of volume 21.68m3. Since then the reformer has been running normal with TMTs well below design limit. A comparison of catalyst loading is given in table-6. Feed gas composition is not same for the conditions considered. Production Catalyst MTPD Loading Original nameplate 1000 6.17 After debottleneck 1300 7.0 With lean gas 1270 7.5 Lean gas &new 1270 5.56 tubes Table-6 Improvement in catalyst loading

Safety and reliability

Management of Change (MOC) is one of the elements of Safety, Health and Environment Management Systems (SHEMS). MOC is intended to manage/control the changes in the quality of raw material also. Though, natural gas continues to be used as the raw material, change in its composition significantly affects the reformer, as discussed in the previous

Implementation
As the reformer tube replacement requires considerable time to implement, the following short term remedial actions were taken: Balancing the fuel firing to bring down the TMT spread, as low as possible.

2010

289

AMMONIA TECHNICAL MANUAL

paragraphs. Therefore, MOC procedure is recommended to be applied to such changes. By means of a prior and proper study, high TMT and load reduction can be avoided. Also, an important lesson to reformer owners is not to stretch the catalyst loading level beyond certain limit, as high catalyst loading can lead to tube failures and hence, can hamper the safe operation of the plant.

Conclusion
Many reformer tubes in APJ had TMTs higher than design level and ammonia production rate was curtailed, subsequent to change in the quality of natural gas. A detailed technical survey carried out in the reformer identified that high TMT problem had occurred due to carbon formation which in turn due to high catalyst loading. Main solution for this problem, as recommended by the survey, was to replace all HK-40 reformer tubes with larger diameter HP micro alloy tubes. Accordingly, all the reformer tubes were replaced. Now, the reformer is running normal and expected ammonia production rate is achieved.

References: 1. JM Reformer Survey (31st Jan & 1st Feb 2005) Report for SABIC Ammonia Plant S Balasubramanian JM Ref. A5130 2. Advanced modeling tools and techniques for primary reformers M Davies, B Fisher & Bill Cotton Johnson Matthey Catalysts AIChE Ammonia Seminar 2004

AMMONIA TECHNICAL MANUAL

290

2010