Top Gas Recycling Blast Furnace Process

Top Gas Recycling Blast Furnace Process

In the area of production of hot metal (HM) by blast furnace (BF), the most promising technology to significantly reduce the CO2 (carbon di-oxide) emission is recycling of CO (carbon mono oxide) and H2 (hydrogen) from the gas leaving the BF top. CO and H2 content of the top BF gas has a potential to act as reducing gas elements, and hence their recirculation to the BF is considered as an effective alternative to improve the BF performance, enhance the utilization of C (carbon) and H2, and reduce the emission of CO2. This ‘top gas recycling’ (TGR) technology is mainly based on lowering the usage of fossil C (coke and coal) with the re-usage of the reducing agents (CO and H2), after the removal of the CO2 from the top BF gas. This leads to lower the energy requirements. Because of the advantages of high productivity, high PCI (pulverized coal injection) rate, low fuel rate, and low CO2 emission etc., the TGR-BF process is considered to be one of the promising ironmaking processes in future.

In TGR-BF, oxygen (O2) is blown into the BF instead of hot air to eliminate nitrogen (N2) in the top BF gas. Part of the top BF gas containing CO and H2 is utilized again as the reducing agent in the BF. CO2 from the BF top gas is captured and then stored. Several recycling processes have been suggested, evaluated or practically applied for different objectives. These processes are distinguished by (i) with or without CO2 removal, (ii) with or without preheating, and (iii) the position of injection.

The concept of the TGR-BF (Fig 1) involves many technologies which include (i) injection of reducing top BF gas components CO and H2 in the shaft and hearth tuyeres, (ii) lowering the consumption of fossil C input due to lower coke rate, (iii) usage of pure O2 gas instead of hot blast air at the hearth tuyere (removal of N2 from the process), and (iv) recovery of pure CO2 from the top BF gas for underground storage.  

Fig 1 Concept of TG- BF

The concept of top gas recycling has been experimentally tested at the LKAB’s experimental BF (EBF) in Lulea, Sweden. The EBF was modified and a gas separation plant based on VPSA (vacuum pressure swing adsorption) technology was constructed near the EBF.

History of BF top gas recycling

For lowering the rate of the reductant and for increasing the productivity of the BF, several new concepts based on the conventional BF process, have been introduced during the twentieth century.

Already in the 1920s, a concept was developed for the injection of the hot reducing gas into the BF. The lower gas volume from the tuyeres in the furnace was needed to be compensated by the injection of preheated reducing gas at 1000 deg C with 27 % CO, 33 % H2, and 26 % N2 into the lower shaft zone. As a result only 30 % of the coke (at that time 345 kg/tHM) was found to be necessary for the BF process. In the mid-1960s this idea was taken up again in Belgium and in the early 1970s the first trials were carried out at a 4.6 m hearth diameter BF in Cockerill-Seraing in Belgium. Specific amount of 400 N cum/tHM of reformed gas which was preheated upto 1000 deg C were injected into the lower shaft of the BF. A replacement ratio of 0.22 kg to 0.26 kg of coke per N cum of reducing gas was observed.

Further investigations were not carried out because of the economic reasons because of the high cost of natural gas. In the late 1970s, development work for a new process started in Germany in which injection of cold pure O2, fuel and recycled gas was carried out at two tuyere levels. Based on this idea a process was developed in Canada in1984 for a conventional BF without a second tuyere row. The main feature of this concept was injection of coal to substitute the coke. Both concepts were never realized and ended only as a study.

Nearly at the same time this idea was further used by NKK in Japan where a second row of tuyeres was installed in the middle of the shaft. Preheated reducing gas, consisting of recycled top gas without CO2 removal, was injected in these tuyeres. The gas was heated by partial combustion with oxygen. Cold O2, coal and cold recycled top gas were injected into the tuyeres of the hearth. NKK tested the process in an experimental BF with 3 tuyeres, an inner volume of 3.9 cum and a hearth diameter of 0.95 m. The coal injection rate could be increased to 320 kg/tHM, while the coke rate could be reduced to 350 kg/tHM. The shortage of fuel gas in the integrated steel plant because of the recycling the top BF gas stopped development of this process, since the price of electrical energy and natural gas was high in Japan.

The first commercial operation of a BF with top gas recycling was performed in the late 1980s in 12 campaigns by RPA Toulachermet in Russia at BF number 2 with a useful volume of 1033 cum. In this all-coke BF process concept, hot top gas and almost free of CO2, was blown into the hearth tuyeres, together with pure O2. The de-carbonated top gas was heated in hot stoves upto 1200 deg C. With this new process nearly 250,000 tons of hot metal (HM) was produced. The lowest coke rate achieved was 367 kg/tHM, compared to the reference of 606 kg/tHM which meant a reduction in coke rate of 239 kg (39 %). During these campaigns, serious tuyere burnouts were seen, leading to changes in the tuyere design. Difficulties with the CO2 cleaning system finally stopped the process. Taking these background investigations into consideration, the concept of ULCOS (ultra-low carbon di-oxide steelmaking) TGR-BF has been developed in 2004.

Development of ULCOS TGR-BF

The development work has been carried out in two phases. In the first development phase, which ran from 2004 to 2009, the process was named as ‘ULCOS new blast furnace process’. During this phase three new process concepts have been developed and tested. In the second phase which started in 2009 and which was named as ‘ULCOS top gas recycling blast furnace process’, two additional ULCOS TGR-BF campaigns were conducted.

During the development heat and mass balance models and a 3-D axi-symmetrical model of the BF were used for the calculation of the main data and the inner state of the process for the selection of the best operating parameters. Four alternatives were defined and examined for the possible reachable C saving and the feasibility of running the BF under these new concepts. The conclusion was that the alternatives 1, 3 and 4 should be able to achieve a fossil C saving of 21 % or higher with a high pulverized coal injection level. Alternative 2 was rejected because of the low expected C saving and the necessity of the challenging technology to heat the recycle gas in two steps first in a recuperator and then further heating by partial oxidation. All the alternatives included CO2 removal and the injection of CO rich product gas into the hearth tuyeres, the usage of pure O2 and the injection of coal together with the reducing gas. In all the alternatives, use of bio mass, partly reduced ore and hydrogen rich gas had been considered as a possibility in future.

In the alternative 1 (Fig 2), the de-carbonated product gas is injected cold with pure O2 and coal at the hearth tuyeres and hot at the shaft tuyeres. One critical point in this alternative was the small cold gas flow rate at hearth tuyere level leading to smaller raceway sizes and higher flame temperatures compared to the normal BF process. Also, a new tuyere design was necessary because of the small gas flow rates.

Fig 2 ULCOS top gas recycling blast furnace – Alternative 1

In alternative 3 (Fig 3), the de-carbonated product gas was injected hot at the normal hearth tuyeres together with O2 and coal. To reach high carbon saving it was necessary to operate with low RAFT (raceway adiabatic flame temperature) and at the same time with high coal injection rate.

Fig 3 ULCOS top gas recycling blast furnace – Alternative 3

In the alternative 4 (Fig 4), the de-carbonated product gas was injected hot at the hearth tuyeres and hot at the lower shaft. The temperature of the recycled gas varied from room temperature to 1250 deg C.

Fig 4 ULCOS top gas recycling blast furnace – Alternative 4

In alternatives 1 and 4, product gas is also injected through shaft tuyeres. The differences are the gas injection temperature, and the position of the injection points. In all the cases, least part of the gas was heated in a regenerative system. The expected fossil C savings for alternative 1 was 21 % at a coal rate of 170 kg/tHM, alternative 3 was 24 % at a coal rate of 180 kg/tHM and alternative 4 was 25 % at a coal rate of 150 kg/tHM.

Mathematical modelling of the raceway conditions and gasification tests were then carried out and both laboratory and pilot scale investigations were done for the design and engineering of the tuyeres under the constraints of simultaneous injection of recycled gas, pure O2 and pulverized coal. The geometry of the tuyere has been improved based on the results of the calculations to avoid hot spots and failure during operation and to keep a sufficient impulse of the gas stream to form a raceway with a sufficient depth.

Campaigns of TGR-BF process in the experimental BF

The objective of the trials was to demonstrate the operation of the EBF in a complete TGR mode with pure O2 and PCI at the hearth tuyeres. This was carried out under the three defined alternatives. The ferrous burden consisted of 30 % of pellets and 70 % of sinter. The production rate of HM was kept at a constant level of 1.5 tons per hour and the PCI rate was varied between 130 kg/tHM to 170 kg/tHM. During the different trial periods the volume of the recycled top gas was maximized in order to obtain maximum fossil C saving. The results obtained during the trials in EBF were then compared with model calculations.

Alternative 3 and alternative 4 were tested in 2007, during the first campaign after a conventional start up. Alternative 3 was tested for optimization during the autumn of 2009 in the second campaign which was followed by a test of alternative 1. In the third and last campaign during 2010 the focus was on alternative 4 at 900 deg C. Alternative 4 was considered to be a preferred alternative for the follow-up ULCOS BF demonstration project on industrial scale. The 900 deg C limit for the temperature of the re-injected gas was set to avoid problems of silica reduction by H2 of the refractory materials.

During each campaign in-situ measurements of temperature and gas composition as well as samples from the burden material were taken from the EBF with the two in-burden probes. For investigating the burden material behaviour under the new operating conditions, baskets with different sinter and pellet materials were charged into the EBF just before the stoppage of the campaign. These baskets were recovered from the EBF during the dissection of the furnace after the quench with nitrogen. This had been done during the campaigns in 2007 and 2010. In the second campaign in 2009 only the quench could be performed due to an unprepared end of the campaign because of problem in the EBF charging system and hence no baskets were charged.

 During the campaigns, samples from the cohesive zone were recovered and size and shape of the raceway measured for further investigation. The campaigns were started with one week of conventional BF normal operation (hot air blast) with sinter, first for heating up and thermal stabilisation and then for establishing a reference for conventional BF operation. After the reference, a stop was made for disconnecting the blower and connecting the product gas to the pebble heaters (regenerators). The start up in the TGR-BF mode was made by artificial blast consisting of cold O2 from the lances and hot N2 from the pebble heaters. In progressive steps the N2 from the pebble heaters was replaced by de-carbonated top gas (product gas).

Results of the campaigns of ULCOS TGR-BF

The first conclusion which emerged from the campaigns is that it is possible to operate the ULCOS TGR-BF process. No safety related issue had occurred during the campaigns with the new process. The operation of the VPSA unit, which is the second important facility was smooth and without any major failures. The EBF coupled with the VPSA unit worked very well during the campaigns.

However, it has to be noted that the operation of the VPSA unit was influenced by the changes in top gas composition and the volume of the gas from the EBF. Hence, both the units were to be operated in a very close relationship. The maximum ratio of recycled top gas which was achieved during the campaigns was around 90 %.

During the campaigns of the TGR-BF process, stable operation of the BF was experienced with a smooth descent of the burden and it was easy to maintain the thermal stability. The efficiency of the gas in the BF shaft was stable during the different alternatives and there was a good gas distribution as shown by the measurements of the in-burden shaft probes. However, there were a few equipment failures during the campaigns due to which the EBF had to be stopped during the operation under the new process conditions. Some long stoppages needed coming back to the operation step of the ‘conventional operation’ with artificial hot blast of cold O2 and hot N2 operation. The small stoppages required only stopping the gas injection and the addition of some extra coke and/or coal.

After start up with N2, when there is no top BF gas available, the product gas could be recycled again within around 1 hour. All problems were solved as predicted and the thermal stability of the BF was never seriously in danger. Along the campaign, the gained experience and increased confidence allowed progressively to make the EBF recovery faster (at production level). Each time during change of alternatives (3 to 4, 3 to 1), the BF was stopped for around 8 hours for making the necessary gas connections to the shaft tuyeres and the change of hearth tuyeres.

The experiments carried out at the laboratory level had shown that the conventional burden material would not be a problem for the new process. This was confirmed during the campaigns. No particular process problems were related to the properties of the burden materials. The results of both probes and excavation samples had shown the reduction profile of a centre working furnace which is a low reduction level at the wall and a higher reduction level at the furnace centre for both sinter and pellets samples. Tumbling tests on the excavation samples had shown disintegration behaviour similar to the one of a conventional BF process which corresponded to the laboratory tests. From the burden testing work, it could be concluded that the burden properties as used in present day conventional BF seem to have no problem for the ULCOS TGR-BF process.

The results achieved during the ULCOS TGR-BF campaigns were very encouraging with regards to saving of the C (coal and coke). The trials of all the three campaigns had shown a substantial decrease in the rate of the reductant which was achieved by the injection of the de-carbonated top BF gas. During the three campaigns the coal and coke input dropped from around 530 kg/tHM to 400 kg/tHM, which represent a considerable saving of the C. The carbon input was reduced from 470 kg/tHM to around 350 kg/tHM, resulting in a C saving of around 25 %.

Although alternative 1 could not be fully explored because of the early stoppage of the second campaign the maximum reduction in C input via coke was 21 % compared to the reference period under conventional BF operation. For this alternative, a new tuyere technology was developed. The tuyere design consisted of co-axial pipes with the inner pipe used for the injection of the pulverized coal and the outer pipe for the injection of O2. The three installed tuyeres worked very well and after disassembling neither damages nor wear was observed. The VPSA was able to recycle upto 88 % of the BF top gas.

As regards alternative 3 the C consumption could be reduced upto 15 % in the first campaign with a top gas recycling ratio of around 72 %. The results of this alternative were lower (around 15 %) than expected from the heat and mass balance calculations (24 %) as this was the first experience with top gas recycling mode and the process was not optimized. In the second campaign the results of this alternative were much better, when the maximum reduction in C input of around 25 % was achieved with a top gas recycling ratio upto 90 %.

In case of alternative 4, a C saving of 24 % was achieved with a top gas recycling ratio of 90 %. In terms of coke and coal consumption, there was saving of upto 123 kg/tHM in the new process (alternative 4) compared to the reference operation period. From these results a good correlation between the quantity of injected gas (CO+H2) and the reduction in reductant rate could be determined. The C input via coke and coal could be reduced by 17 kg in average per 100 N cum of gas (CO+H2) injected.

The campaigns of EBF have proved that it is possible to run a BF process at a much lower fossil C consumption level when compared with the consumption level of the present day BF. A C saving upto 25 % was proven by the injection of the reducing de-carbonated top gas. This is a significant drop compared to present day best operated BF process. As a matter of fact, the application of BF-TGR technology on modern BF is expected to lead to reduce the C consumption from a present level of around 405 kgs C/tHM to a level of around 295 kgs C/tHM.

VPSA unit has operated stably. It had been noticed that the VPSA unit could treat 97 % of top gas from the BF. The average volume fraction of CO2 in injected gas was around 2.67 % and the CO recovery rate was 88 %, meeting the requirements of quantity and quality. Combined with VPSA and CCS units, the CO2 emissions reduced by TGR-BF process could reach to 1270 kg/tHM which accounts for 76 % of the total CO2 emissions in ironmaking process. 24 % of reduced CO2 was by gas recycling and the other 52 % was transported and stored underground by CCS.

From the process point of view, it can be stated that the ULCOS operation is more stable than the conventional BF operation, as far as the temperature and quality of the HM is concerned. This seems essentially due to the lower influence of the solution-loss reaction related to the much lower levels of direct reduction rate (DRR). The lowest observed value of this DRR is 5 %. There was no indication in the operating results that this was actually the minimum value which could be reached in ULCOS TGR-BF. The quality of HM was largely impacted by the operation of ULCOS TGR-BF. Especially, a substantial decrease of the silicon content (greater than 1 % absolute) and a correlative increase of the C content were observed. It must however be pointed out that the silicon content in conventional BF operation is much lower (around0.5 % in conventional BF against around 2.0 % in the EBF), and hence no such a big change should be expected during the application of the ULCOS BF-TGR process at the industrial scale.

The test campaigns of ULCOS TGR-BF have shown that the new TGR-BF process is feasible and easy to operate. It can be operated with good safety, high efficiency and strong stability. The test campaigns have proved that it is possible to shift the EBF operation between 4 modes of operation (conventional, alternatives 1, 3 and 4). It also proved that it is possible to operate the BF process and the gas separation plant VPSA in a closed loop. The obtained C savings were consistent with the predictions from the flow sheet calculations. The tests also indicated that conventional burden materials sinter, pellet and coke are suitable for the ULCOS TGR-BF process.

Based on the experience from EBF campaigns the items which requires attention for safe and sustainable industrial application of the ULCOS TGR-BF process are (i) risk for gas leakage at tuyere level may demand that the tuyere level is physically separated from the cast house for fire protection, (ii) risk of leakage due to bending of flanges at high temperatures, (iii) elongation of the chimney from hot stoves/heater to avoid poisonous gas in the surroundings, (iv) for avoiding the failures at start-up phase, operational personnel are to be trained with artificial gas during functioning test, (v) optimum injection parameters are important to avoid build-up/clogging at tuyere nose, (vi) importance of individual blast flow control to the tuyeres to avoid build-up of excessive O2 and pulverized coal in the tuyere in case of clogging, (vii) selection of refractory and manufacturing of refractory items are to be modified and the use of C steel anchors are to be if possible avoided, alternatively other materials than C steel are to be used, or steel coated with alumina, and (viii) it is important to avoid metal dusting by means of a careful selection of metal parts.

The test results have shown that the alternative 4 had the best effect of emission reduction and was selected as the first choice for the trial on the industrial-scale BF during the next stage.