REHAN AHMED, IFTIKHAR A. KIRMANI, M. SAQIB JAWED, Fauji Fertilizer Co. Ltd, Goth Macchi, Dist. Rahimyar Khan, Pakistan
This article was first published at Nytrogen and Syngas Conference, 21-24 February 2011, Düsseldorf, Germany
Fauji Fertilizer Company Limited (FFC), Pakistan, operates three ammonia-urea trains, all of Haldor Topsøe and Saipem (Snamprogetti) design. FFC Plant-I started commercial operation in June 1982. In the past 28 years, the Ammonia-I plant has exhibited excellent performance, both in productivity and efficiency. With more than 11 million tons of ammonia produced and 225,000 hours of operation, the plant has been the main contributor to the company’s growth since inception.
The plant has the unique feature of operating alloy IN-519 reformer tubes for more than 220,000 hours, at or around 135% of original name plate capacity. In September 2009, the plant underwent a mini-overhaul, which entailed changing the reformer tubes and the synthesis converter basket, revamping the HP ammonia separator, and replacing a few heat exchangers. This paper describes some of the key achievements of the plant, major changes that have been implemented during the course of time, the impact of the recent overhaul, and future prospects.
Fauji Fertilizer Company (FFC) is the largest urea manufacturer in Pakistan, operating three ammonia-urea plants; two at Goth Machhi and one at Mirpur Mathelo (Dist. Ghotki).
The first plant (Plant-I) was commissioned in 1982 at Goth Machhi with design capacities of 1,000 and 1,725t/d (metric tons per day) of ammonia and urea, respectively. The ammonia plant employed a conventional Haldor Topsøe design, while the urea plant was based on the Snamprogetti (now Saipem) ammonia stripping technology. The plant was successfully revamped to 122% of design capacity in 1992.
Plant-II was commissioned at Goth Machhi in March 1993, with design capacities of 1,100 and 1,925 t/d ammonia and urea, respectively. The ammonia plant was based on the Haldor Topsøe low-energy process.
Plant-III at Mirpur Mathelo was acquired in 2002 and was similar in design to the Plant-I. Design capacities were 1,000 and 1,740 t/d ammonia and urea, respectively. This plant was also successfully revamped to 125% of design capacity in 2008.
The ammonia plant being discussed in this paper is the Goth Machhi I unit (Plant-I).
Ammonia-I plant was a conventional Topsøe design of the late 1970s, featuring a high steam-to-carbon ratio(3.75), a hot potassium carbonate system for carbon dioxide removal, and an ammonia synthesis loop operating at high pressure.
It consisted of desulfurization, conventional reforming (large primary reformer and secondary reforming with air) and shift conversion (high and low temperature) sections at the front-end. The carbon dioxide removal section utilizes the Benfield technology from UOP and is followed by a methanation reactor.
The synthesis loop operates at a high pressure of 267 kg/cm2g. All the major compressors (process air, synthesis gas and ammonia refrigeration) are centrifugal machines driven by steam turbines.
A block flow diagram of the Ammonia-I plant is presented in Fig. 1.
The plant started production in 1982 and had the distinction of achieving the design capacity in the first year of its operation. The ammonia production from the plant was increased to 115% of the original design by 1990 with small modifications.
Modifications done in the past
Debottlenecking of the Ammonia-I plant was carried out in 1991-1992 and the plant was successfully revamped to 122 % of the original nameplate capacity. The following modifications were carried out:
• The original ammonia synthesis converter basket (Topsøe S-100 design) was replaced with a two-bed basket (Topsøe S-200 design). This was the major change to increase ammonia production.
• A steam superheater coil was added in the flue gas duct of the steam superheater furnace to attain the design temperature of high-pressure steam generated from the ammonia plant.
• A make-up gas chiller was added upstream of the synthesis gas compressor to decrease the temperature of make-up synthesis gas and thus to accommodate an increase in flow at the desired capacity.
The carbon dioxide removal section, based on conventional Benfield technology from UOP, was converted to the Benfield Lo-heat process by adding a flash vessel with a multi-stage ejector system (Fig. 2) in 2004 as an energy improvement project. This helped to decrease the steam-to-carbon ratio from 3.75 to 3.40.
New larger BFW pre-heater
The BFW pre-heater installed at the inlet to the low-temperature shift converter to cool the process gas was replaced in 1987 with a rod-baffle design (Phillips proprietary design) after leakage in the original exchanger due to high flow-induced vibrations. The exchanger installed in 1987 was replaced with a new rod-baffle design exchanger provided with vapour belts at the outlet nozzles in 2006 to accommodate increased plant loads.
The operating load of Ammonia-I plant was gradually increased with a few other modifications. It operated at 130-133% during 2006-2009.
RECENT MODIFICATIONS – 2009
The primary reformer (F-201) at Ammonia-I is a conventional Topsøe-design reformer with side wall mounted burners. The natural gas feedstock, after desulfurization, is mixed with process steam before entering the primary reformer. Natural gas fuel is fired in the burners.
The primary reformer consists of two chambers, each having 144 tubes – total 288 tubes. The 648 wallmounted burners are arranged in six rows and distributed uniformly across the two side-walls of each chamber. The original tubes were of alloy IN-519 (24Cr-24Ni-Nb) material.
FFC had operated the reformer in the optimum way to enhance the life of the reformer tubes and avoid failures of the tubes. The tube metal temperatures (TMTs) have always been kept well within the design limit, despite operation of the plant at much higher throughput.
The reformer tubes in service since commissioning in 1982 (IN-519; 24Cr-24Ni-Nb by Wisconsin-USA) provided a record service life of more than 220,000 hours, far surpassing the design life of 100,000 hours. During the service period, only six tube failure incidents occurred between 2003 and 2009; the first one after a service of more than 20 years. The leakage was mostly at the bottom (~10 inches from the floor) – i.e. the area of highest tube temperatures. All these tubes were nipped from both ends – top and bottom.
The higher life has been possible on account of excellent operational control with optimized fuel gas firing and vigilant monitoring to keep the TMTs below the design value of 925°C. Monitoring is carried out on a regular basis of values such as uniformity of firing and TMTs, differential temperature between chambers, corrective action for tubes with high TMT values, etc. The average TMTs since 1992 are presented in Fig. 3 and a typical TMT profile is presented in Fig. 4.
A plan for replacement of the tubes was prepared in 2004 to be ready for implementation without delay when the need arose – expected by 2009. Improvement in the tube design specification with respect to flexibility in operation, expected increase in plant load, and energy savings, was also contemplated while studying the replacement options considering the new metallurgical-development.
HP Microalloy (25Cr-35Ni-Nb-Ti) was selected for the new tubes, replacing the existing IN-519 (24Cr-24Ni-Nb) material. Its increased high-temperature strength allowed the tubes to be designed with thinner wall thicknesses, resulting in more cross-sectional flow area for a fixed outside diameter.
Considering the thinner wall thicknesses possible with Microalloy material, the option of increasing the inside diameter was chosen, allowing more catalyst to be loaded to accommodate higher process gas flow.
The new tubes were obtained from Doncasters Paralloy-UK. A comparison of the old and new reformer tubes is presented in Table 1.
The following advantages of using Microalloy tubes of larger ID (same OD) were foreseen:
• Low pressure drop and less fuel firing, resulting in energy saving;
• Low TMTs, giving enhanced tubes life;
• Less methane slippage, so enhanced urea capacity;
• Suitability for future capacity revamps.
The new tubes were replaced in November 2009 (Fig. 5) and loaded with Topsøe reformer catalyst (R-67-7H /R-67R-7H) using the Spiraload method.
Operational experience with the new tubes
The primary reformer operating performance with the new tubes and catalyst was as expected; with low TMTs, low methane slippage (improved heat gradient) and pressure drop, even though the plant was operating at a higher load.
A comparison of the operating parameters before and after the tubes replacement is presented in Table 2.
Comparisons of the average TMTs and TMT profile after tubes replacement are presented in Fig. 6 and a sample TMT profile is presented in Fig. 7.
The ammonia converter internals consisted of a Topsøe-designed S-200 converter basket, which was installed in 1991, replacing the original S-100 converter basket.
The catalyst was also installed in 1991 along with new S-200 basket required replacement owing to reduced performance with respect to lower conversion efficiency and higher approach to equilibrium after rendering 18 years of satisfactory performance. In order to gain maximum benefit of plant outage owing to catalyst changeout, replacement of S-200 with S-300 was also synchronized.
Study for three-bed basket
The three-bed concept with cooling between the catalyst beds gives a high conversion for each converter pass, since for each bed the achievable conversion is limited by the equilibrium of the ammonia synthesis reaction. The pressure drop is slightly higher but is outweighed by the advantage of higher conversion and lower loop pressure. A further advantage is that the synthesis gas chilling duty moves from the ammonia refrigeration circuit to the synthesis loop water cooler because of the higher ammonia concentration at the converter outlet. Operation of the synthesis loop under milder conditions leaves room for a possible future capacity increase. Thus the ammonia converter with a three-bed converter basket was indeed found feasible.
Topsøe’s S-300 radial flow converter basket was selected for installation in the original ammonia synthesis converter pressure shell, being a highly efficient converter based on three adiabatic converter beds with interbed cooling. All of the inlet gas passes through all three catalyst beds, resulting in a higher conversion. The mechanical design was based on the well-proven S-200 converter basket. Therefore, the same reliable operation was expected with the S-300 converter as with the S-200 design. No modifications / replacement of the shell, piping or other loop equipment were required.
The S-300 converter has also a lower heat exchanger. A sketch of the S-300 converter basket is presented in Fig. 8 (overleaf).
The following advantages were foreseen:
• Higher conversion per pass, leading to increased ammonia production
• Reduced synthesis loop pressure, resulting in compression-energy saving
• Reduced inert level, therefore lower purge gas rate
• Suitability for future capacity revamps.
The new S-300 ammonia converter basket was installed inside the original pressure shell in November 2009 (Fig. 9) and loaded with Topsøe ammonia synthesis catalyst (KM1R) using Showerhead method.
Two revamp conditions were forecast at two inert levels (16.5% and 22.5%), obtaining maximum converter outlet temperatures of 380°C and 360°C, respectively (360°C being the original design temperature).
The operating performance of the ammonia synthesis converter with the new S-300 basket exceeded the predictions regarding ammonia production rate. Moreover, other operational parameters also remained as expected or better despite plant operation at higher load.
This was possible after extensive optimization carried out by FFC team after completion of the catalyst reduction activities. A comparison of the operating parameters before and after the basket replacement is presented in Table 3.
The ammonia separator (V-501) is installed in the high-pressure ammonia synthesis loop to remove ammonia product from the converter effluent gas. The original ammonia separator was a conventional knock-out vessel with mesh-type mist eliminator at the top of the vessel and is presented in Fig. 10.
Entrainment of liquid ammonia in the recycle synthesis gas from V-501 was suspected owing to a high concentration of ammonia at the inlet of the reactor. Carry-over of liquid ammonia from the separator at high plant load was found to be the most plausible cause; this was resulting in a direct loss of ammonia production.
The original separator (designed for 1,000 t/d ammonia production) was found to be inadequate for operation at high loads on account of high velocities.
There was a potential of 10 to 15 t/d additional ammonia production with efficient operation of the separator.
Replacement with a new larger separator
The possibility of replacing the existing vessel with a new, larger separator was reviewed which could easily cater for future revamps as well. However, on account of the long delivery period quoted by the vessel fabricators for this special type of equipment, this option could not be implemented in the 2009 turnaround-and replacement had to be deferred to the next turnaround in 2012.
Retrofitting with a vane unit
An alternative option of retrofitting the existing vessel with vane unit internals was evaluated in consultation with Peerless- Singapore, as that option could be synchronized with the 2009 turnaround.
As an interim, technically feasible way of improving the separation efficiency at current operating loads, the existing V-501 was retrofitted with a vane unit (Peerless-USA). A sketch of the retrofitting with vane unit is presented in Fig. 11.
Main features of the proposal were:
• A new vane unit (64″ x 48″) installation; replacing the existing demister pad:
• No welding on V-501 shell (heat treated)
• New vane unit in pieces and match-marked to pass thru existing man-way for reassembly inside the vessel
• New vane unit supported on existing demister sealing ring
• Installation of a new horizontal quieting plate
• Changes in level instruments positions / settings – presented in Fig. 12(opposite).
The normal operating level was also lowered by the identical margin as the high-level switch (from 1531 mm to 1152 mm above BTL). Normal level indication on the LIC remained as original (30%), as the location of the transmitter was lowered.
Response times for level variation changed; however, the critical timings remained as such. The response timings are presented in Table 4.
The following advantages were foreseen:
• Decreased entrainment of liquid ammonia in the recycle synthesis gas
• Better ammonia conversion in the reactor
• Increased ammonia production.
The operating performance of the retrofit of the ammonia separator with the vane unit fully met the expectations and the ammonia in the recycle gas at the outlet dropped to the equilibrium ammonia concentration at the operating conditions; as represented in Fig. 13.
Performance of all heat exchangers has remained satisfactory since plant commissioning. However, a few exchangers have been replaced due to frequent leaking problems.
Synthesis gas cooler
The synthesis gas cooler (E-503) is amongst a series of exchangers installed to cool down the ammonia converter outlet gas for liquefaction of produced ammonia, and is installed upstream of the first ammonia refrigeration chiller. The original cooler with CS-bundle (U-tubes) was replaced twice in 1986 and 1999 owing to leakage problems.
Severe corrosion and pitting was observed in 2006; accordingly the decision was made to replace the tube bundle with SS tubes in view of the better performance of SS tubes with respect to fouling, reliability, etc.
Consideration was also given to reducing the tube thickness with SS material tubes but the idea was dropped as the benefit to the thermal performance would have been insignificant. Moreover, reducing the thickness could have resulted in less damping potential for vibrations, which was not desirable.
The new cooler (tube bundle completely fabricated from and old bundle at FFC workshops in Goth Machhi) was also installed in 2009.
Methanator feed / effluent exchanger
A tube leak occurred in the methanator feed-effluent exchanger (E-311) for the first time in May 2007. The main cause of leakage was attributed to high gas velocities at the inlet / outlet areas, in particular.
A new exchanger with a changed configuration – that is, with the shell and tube side fluids swapped round (hot gas on the shell side instead of originally on the tube side and vice versa), was decided upon. The tube pitch was decreased from 36 mm to 25.4 mm to mitigate the risk of damage from flow-induced vibrations.
The new exchanger was also installed in 2009. As a hallmark of FFC in-house capabilities, the equipment was designed and fabricated completely at FFC’s workshops in Goth Machhi.
Plant control system
The control and emergency shut-down systems at the Ammonia-I plant were also successfully shifted to new DCS- and PLC-based systems, respectively.
Steam generation coil
The steam generation coil in the flue gas duct of the primary reformer furnace had bare tubes in view of the likelihood of fins burning at the high flue gas temperature. New coils with finned tubes were installed in November 2009, taking advantage of advances in metallurgy, to achieve better heat transfer with more steam generation and to overcome induced draft fan limitation at high loads.
The overall effect of the above modifications and the expected increase in plant loads on the remaining equipment and piping with respect to line velocities, conversion parameters, ρv2 values at exchangers and vessels, etc., were carefully evaluated and simulations were done.
The modifications carried out in the maintenance turnaround of 2009 at the Ammonia-I plant proved very successful, not only in overcoming the problems and limitations that were being faced in the above areas, but also because the overall effect of these modifications also resulted in boosting ammonia production rates to new heights. The capacity factor of Ammonia-I over its history is presented in Fig. 14 (opposite).
After the successful implementation and commissioning of these modifications ammonia production increased by more than 60 t/d (6%); along with an energy index improvement by 5% (8.40 Gcal/t ammonia achieved compared with 8.85 Gcal/t ammonia before the modifications of 2009).
The total investment was US $ 0.139 million per t/d ammonia and the return comes out to be US $ 0.077million per t/d ammonia, indicating a payback of less than two years.