Ironmaking
Premature Failure of Copper Staves and Applied Results for New Designed Staves
2021 年 61 巻 10 号 p. 2507-2512
詳細
2021 年 61 巻 10 号 p. 2507-2512
In a blast furnace employing a copper stave as a cooling system, premature stave wear is experienced in many steel mills. To confirm this failure mechanism, an online ultrasonic thickness measuring device was installed in the Pohang 4 blast furnace from blow-in to acquire data. Through this, the refractory damage in hot face and real-time wear data were obtained, and the correlation with the operational factors was analyzed. Through the CFD simulation, the effect of temperature due to the refractory was confirmed, and the possibility of acceleration of the copper body was confirmed in the absence of the refractory in hot face. To improve the life of the entire refractory as well as the stave body, a refractory stress test was conducted in the laboratory, and a stable groove structure was derived using the structure program DEFORM using the data obtained from cold tests. The new designed long-life copper stave was applied to the Pohang 4 blast furnace in November 2015, and it has been proved that the life of the stave has been significantly improved since wear has not been measured until now.
Copper staves had been first applied in 1979 in blast furnace No. 4 at the Hamborn plant in Thyssenkrupp (former Thyssen Stahl AG). The applied results of two staves for 9 years were successful without any severe damage in cooling channel. Thereafter it has been applied in numerous blast furnaces due to extension of service life as well as ease to inner volume increase and shortening of relining period. Due to these advantages, some blast furnaces began to adopt copper staves in the 1990s, and in the 2000s, plenty of blast furnaces, including POSCO, began to adopt copper staves. However, premature stave abrasion occurred in some blast furnaces, and WSA (World Steel Association) surveyed the blast furnaces in the membership companies.1) The operational factors including productivity and the design factors of the blast furnace were analyzed, and the blast furnace profile was mentioned as a major factor in stave wear. However, the report did not cover detailed stave wear mechanisms as it was only a macro data analysis.
After applying copper staves to the Gwangyang No. 1 blast furnace in 2002 for the first time, POSCO adopted them in 6 of 10 blast furnaces as their main cooling system. The initially applied copper staves were designed to be about 3 m in height with 5 cooling channels and they were prematurely damaged by thermal deformation due to the so-called banana effect. After applying staves of about 2 m in height in the following blast furnaces, this problem was solved. However, in the later blast furnaces, premature damage to the stave was caused by abrasion of the inner surface of the blast furnace.
Prior to the development of new staves, before 2015, damaged staves in POSCO were repaired in the form of inserting cooling plates from outside by cutting the furnace shell on which the damaged stave is located. This shortened the campaign life of some blast furnaces.
In this paper, the stave wear behavior was measured using the online wear measuring device installed from the beginning of blow-in, and the stave damage mechanism was studied through experimental, operational and computational analysis as well as measured data.
Pohang No. 4 blast furnace was blown in Oct. 2010. The cooling system of the blast furnace is a combination of staves and cooling plates, and 14 rows cooling plates are applied in the bosh to maintain bosh angle, copper staves from belly to middle shaft and cast-iron staves in the rest.
The common design of previous old copper staves in POSCO is shown in Fig. 1, and the hot face design has dove-tail grooves, which are combined with wear resistance refractory bricks, SiC, and the whole face is covered with castable refractory material which is based on Al2O3 and SiO2. The thickness of stave in belly is 150 mm in total, where the depth of grooves (or thickness of rib) is 40 mm, the thickness between groove and cooling channel edges 30 mm, the diameter of cooling channel 50 mm and the total thickness of SiC brick is 60 mm (20 mm extruded on the surface of old copper stave).
Typical design of old copper stave; numerical unit is in millimeter. (Online version in color.)
Despite the bosh angle was maintained by installed cooling plates, wear started first from belly and failure in the cooling channel began on June 20. 2014 as shown in Fig. 2. After detection of wear on belly, which was measured with the ultrasonic devices to be later described with on-line as well as off-line, in order to protect the facility and adjust the production, productivity was slightly lowered.
Cumulative cooling channel failures in belly (B1) and from first row shaft staves adjacent to belly (S1) to third row shaft staves (S3). (Online version in color.)
The productivity of Pohang No. 4 blast furnace after blow-in is shown in Fig. 3. At the initial stage after blow-in, a high productivity operation of around 2.65 was performed, and stave wear on the belly had been measured at that time. The productivity was lowered to an average of about 2.5 from 2014 by adjusting the production, and stave wear on lower shaft was accelerated in Fig. 2. because in the cohesive zone, the stress to the wall can be maximized by the load of the charged material and the lifting force of the gas, and the high FeO component can be reacted with the primary protective layer, especially castable based on Al2O3 and SiO2. For these reasons, although failure to the belly stave started in advance at a high productivity of more than 2.6 during the initial operation after blow-in, the rate of damage to staves in the lower shaft was faster.
Productivity trends after blow-in in Pohang No. 4 blast furnace. (Online version in color.)
For stave replacement from November 2015, a new design of long-life staves, which will be introduced in this paper, has been applied. The best record is that 71 staves were replaced just in 58 hours in 4th repair, and it was carried out without banking such as for (semi-) relining, using the extended regular safety day period. In addition, there is an advantage in terms of economics as well as safety because workers do not need to enter into the blast furnace.
2.2. Measurement Results for On-line Stave ThicknessThe Pohang No. 4 blast furnace was installed with a unique embedded ultra-sonic sensor in Fig. 4 that can measure wear conditions in real time from the beginning of blow-in.2) These sensors were installed in each row and every 90 degrees from belly to 3rd shaft row, and a total of 16 were installed. From these sensors, noble data to understand the stave wear mechanism were obtained and stave wear started at first from belly and showed wear trends in four directions in Fig. 5. The first abrasion started around October 2011, which is 8 months after blow-in in October 2010, and wear is observed in all directions in 12 months. This means that the life of castable refractory and SiC brick to protect belly stave is less than 1 year. In the period from August 2011 to May 2012 after the initial stave body wear, the average productivity was about 2.67, and the wear rate was measured from at least 5 mm/a in the 360-degree direction to 10 mm/a in the 180-degree direction. However, there was no correlation between productivity and wear rate from the data. Since the damage timing of staves of S1 or higher rows is late, the wear rate is less than the above rate. The thickness of the stave hot face is 70 mm, at this wear rate, the belly stave can be used for 7 to 14 years. However, not only the belly part but also all staves up to S3 in Fig. 2 are damaged in 6 years. This phenomenon is well explained by data in circles from March to October 2012 in Fig. 5. In certain months of this period, sudden wears of about 5 mm in a few days are observed and this made the stave life much shorten.
Cross section image to show the embedded ultrasonic sensor position and the cooling line of the copper stave behind steel shell. (Online version in color.)
Stave body remaining and productivity trends which is measured by embedded ultra-sonic sensors in belly since blow-in. (Online version in color.)
The hardness of copper decreases to less than half at 200°C or more.3) This fact can be proved in 2013 by analyzing the heat-load of cooling water by zone at that time in Fig. 8. The stave coolant of our blast furnace is divided into 6 lower headers from the main header from the circumferential direction of 330° for zone 1, and the cumulative minutes of the zone in which the heat-load of more than 9000 Mcal/hr (or 10000 Mcal/hr) was recorded was accumulated for each zone. It was confirmed that the area where the long-time high heat-load was recorded in zones 4, 5, and 6, which is from 150° to 330°. As a result, stave failure in that area was faster than elsewhere and a higher heat-load of 10000 Mcal/hr had a high correlation with the stave damage.
Cumulative minutes of more than a specific heat-load (Mcal/hr) per zone for about 1 year in 2013; Zone1 (330–30°), Zone2 (30–90°), Zone3 (90–150°), Zone4 (150–210°). (Online version in color.)
If the profile of the blast furnace, especially the bosh angle is too steep, the rate of peripheral burden descent is too fast, accretion is suppressed, and stave wear can be accelerated. On the other hand, in some steel works, even if the profile is not properly designed, blast furnaces with a bell-like charging pattern in which the burden is heavily charged to the wall, are suppressed abrasion due to accretion on the stave surface lowering the peripheral burden descending speed as well as the decrease in the temperature of the wall.
However, premature wear of the staves occurred in Pohang 4 blast furnace, regardless of the blast furnace profile or operational factors, and research and development proceeded in the direction of replacing new designed staves which had extending the life of artificial accretion, abrasion-resistant refractory material.
A structure in which the entire hot face of the copper stave is specially protruded to hold burden material without refractory to form a stagnant burden layer on the wall was also conceived, but it was not adopted due to the strength of copper to be described later.
3.1. Surface Temperature Calculation of Copper Stave from CFDCFD simulation was performed using FLUENT to infer the temperature of the copper stave body. The stave of the POSCO blast furnace has five cooling channels and stave body thickness is 130–150 mm (130 mm for simulation). From the simulation experience so far, we know that the temperature at the edge of the stave is high, so as shown in the Fig. 9. 3 and 2 cooling channels of two adjacent staves were set as periodic conditions. Governing equations including the k-ε turbulent model and boundary conditions except the following values in Yeh et al.’s paper was referenced.4)
Periodic condition for CFD simulation; max temperature in red is 274°C. (Online version in color.)
But the temperature of gas at hot face was 1200°C, the heat transfer coefficient was 230 W/m2K, the air at the shell was 27°C, and the air flow rate of 5 m/s was assumed and calculated. So, the heat transfer coefficient at outside of furnace shell was set to 10 W/m2K and cooling water speed to 3 m/s. In addition, material parameters such as stave and refractories were used as the values shown in Table 2, which are the values of materials used in POSCO. Since the purpose of the simulation is the effect of the temperature of the stave body by refractory materials, the comparison was made assuming that the front surface refractories layer was normally present as its initial state, or the inserted SiC was completely eliminated and only the copper body was present.
| No. | Period | Row & Exchang. No. | Remark | |||
|---|---|---|---|---|---|---|
| B1 | S1 | S2 | S3 | |||
| 1 | 2.–5.Nov.2015 | 11 | ||||
| 2 | 26.–29.Apr.2016 | 30 | ||||
| 3 | 11.–15.Oct.2016 | 22 | 32 | |||
| 4 | 17.–22.Oct.2017 | 41 | 30 | Best Record | ||
| 5 | 8.–12.Apr.2019 | 52 | ||||
| Material | Density (kg/m3) | Th. Conductivity (W/mK) | Heat Capacity (J/kgK) | Th. Expansion (mm2/s) |
|---|---|---|---|---|
| Mortar | 1500 | 0.8 | 1170 | 0.46 |
| Copper Body | 8960 | 381 | 387.6 | 109.7 |
| Steel Shell | 7700 | 25 | 460 | 7.06 |
| Cooling Water | 995.6 | 0.618 | 4174 | 0.15 |
| Castable | 2500 | 2 | 1130 | 0.71 |
| SiC | 2650 | 26.1 | 1207 | 8.16 |
| Backfilling | 2600 | 0.73 | 1208 | 0.23 |
As shown in the simulation result in Fig. 9, the edge of the adjacent stave has the highest temperature, and in particular, the temperature at the lower edge, which is far from the coolant, is the highest. In the presence of refractory materials, the average temperature of the stave body is 42°C and the maximum temperature is 71°C. In the case without refractory materials, the average temperature of the stave hot face rises to 133°C and the maximum temperature to 274°C even the all side edges exceed 200°C. When the stave rib is increased by 50 mm and the total thickness is 180 mm stave, it is predicted that the maximum temperature exceeds 400°C without refractory protection. Since the strength and hardness of copper rapidly decrease above 200°C, if the stave hot face is not protected by wear-resistant refractory material, or if no accretion layer is formed, the stave body can be easily worn.
3.2. Wear Resistance of SiC BrickPin on disc experiments were conducted using the device shown in Fig. 10, in order to test the abrasion resistance of refractories. The device was maintained at 1200°C or at room temperature under an argon atmosphere, and the RPM was adjusted in consideration of the descending speed of the burden according to productivity ratio. The burden descending speed was tested by changing the RPM from 8 cm/min for 2 THM/m3/d to 12 cm/min for 3.0 THM/m3/d, and the load on the abrasion material was set to 8.76 kg considering the wall pressure.
Wearing testing device (Pin on disc). (Online version in color.)
Although lower hardness burden, coke or ore should be used, the abrasion material was damaged and abrasion on SiC did not proceed, so a significant result could not be derived. Even zirconia could not withstand the test environment, but when it was performed with SiC as an abrasive, it showed a weight loss of 0.2 to 0.29 g/cm2 at room temperature and 0.25 to 0.39 g/cm2 at 1200°C. If the extreme case, the burden descending speed of 3.0 t/d/m3 at 1200°C even with the SiC abrasive, is converted, the total wear thickness of SiC is less than 70 mm in 10 years. Since the charged material has lower hardness than the SiC abrasive and considering experiences applied SiC brick in POSCO, it is concluded that it is possible to support the blast furnace campaign life in terms of pure wear.
Accordingly, detailed analysis was conducted on the wear mechanism of the existing stave. The fact that stave wear started within 1 year means that the life of the refractory in hot face that protects the copper stave body was extremely short. Even if castable refractories are easily damaged in a year from experiences, the wear resistance of SiC has been proven through wear tests, the possibility of damage due to wear was excluded. However, due to the structure of the dove-tail type joint on the main body of the copper stave, SiC could be damaged by intensive stress or the deformation of the joint itself could be expected.
3.3. Strength of SiCThe SiC was supplied by Chosun Refractories Corporation, and the average compressive strength was 131.8 MPa. Assuming castable is disappeared and the SiC is exposed to burden, shear stress tests in Fig. 6 were performed by applying a load from the top of SiC in a mold, which is made of carbon steel in the dove-tail type of old stave groove. As a result of the test, the refractory was damaged under a load of about 15 tons (about 60 MPa considering the contact surface of SiC).
Mold for SiC brick shear test (left) and press device (right). (Online version in color.)
In the new stave, it was intended to protect the entire hot face with wear-resistant SiC without weak castable. Since stress is concentrated at the edge of the dove-tail joint in the old stave, the rib part was designed in a round shape in the new stave in Fig. 11 to distribute stress. In order to maintain the effect of SiC for a long time, instead of reducing the thickness of the stave body from 150 mm to 130 mm, the thickness of the SiC was enlarged to 70 mm to maintain the same overall thickness. Because only 20 mm of SiC brick protruded in the old stave, it was tested again under the same conditions with 70 mm for comparison under the same conditions. SiC brick were destroyed at 15 tons (28 MPa; despite a similar load with old stave test, the cross-sectional area became larger, resulting in a smaller stress per unit area.) in old stave design, but it was improved to 18 tons (33 MPa) in new one.
New stave design protected by SiC bricks with stress dissipation structure. (Online version in color.)
Since the shear stress test was performed on a carbon steel mold with higher strength than SiC, the results of the copper joint were simulated through DEFRORM which is a representative plastic deformation simulation tool. As a result of the first and second simulations with 60 MPa and 33 MPa applied respectively, all of the old dove-tail design with 20 mm extruded SiC in Fig. 7 were deformed. As a result of the simulation while decreasing by 5 MPa from 30 MPa, the joint was deformed even at 15 MPa, which is a much lower stress than the fracture strength of SiC from shear stress test. It can be said that the stave itself is deformed rather than the SiC refractory material being damaged or fractured, and the SiC can fall out. For reference, there was no deformation under the load of 10 MPa. In the new stave, there was no deformation even at the SiC fracture strength of 33 MPa on the contrary. In other words, the problem of premature wear of staves can be solved because SiC does not fall out as easily as before. In addition, the SiC bricks in Fig. 11 that protect the entire hot face are interconnected each other and the stress is distributed, so the load applied to the joint would be much smaller and deformation of the joint is minimized.
Simulation results for old design joint of copper stave by DEFORM: Load condition a) 33 MPa, b) 15 MPa, c) 10 MPa. (Online version in color.)
The newly developed 11 long-life staves were replaced on the belly where the first wear began in the Pohang No. 4 blast furnace in November 2015. In 2017, however, large scaffold was formed on the shaft and there was long operational problem for about 2 months from March to May due to hanging, slip and drop. The cause of this would be that yard sinter of low quality was charged about 8300 tons as the repair of the No. 4 sinter plant was delayed and scaffold was formed where the thickness difference between the second row of shafts with new staves and the old staves with wear was about 120 mm. In order to solve this problem, a number of rapid drops were performed to drop the scaffold by suddenly reducing blast volume, and harsh operation was performed in which the heat-load was increased to melt down the remaining scaffold. Therefore, operationally, it can be said that the new stave was exposed to more difficult condition, although the productivity in Fig. 3 has been decreased compared to the period with old staves. Moreover, the stave wear rate has relationship not with productivity but with heat-load as mentioned before.
Even when measuring the thickness using ultrasonic waves on the regular safety day, the wear of the body was not measured, and as a result of internal observation when additional staves were replaced, the front refractory was still maintained. It can be said that the stave life has been significantly improved compared to the previous one because the entire SiC bricks have been protecting the stave until now, more than 5 years later.
The old stave was found out through DEFORM analysis, in which SiC was removed by deformation of the copper body joint even with a relatively small stress of 15 MPa when SiC was exposed to the burden after the castable refractory material was removed, because the stress is concentrated at the edge of dove-tail joint structure. When a copper stave is exposed to the high temperature environment without protecting layer, either accretion or refractory, the surface temperature of copper stave rises above 200°C from CFD analysis, when gas temperature is 1200°C assuming concentrated heat-load in a specific direction. At this temperature, the hardness of the copper decreases by less than half, accelerating the wear. In general, the wear of the stave body which was exposed to burden material was about 5–10 mm in a year from continuous wear measurement data, although it was different in each direction. There was no correlation between productivity and wear rate from the data. However, there were cases of about 5 mm wear in a few days, when the copper hardness was decreased rapidly as the surface temperature of the stave would have exceeded 200°C where the heat load is concentrated. The copper stave, which is weaker than the burden, will inevitably wear out without accretion or protection of the refractory material.
In the new stave, SiC bricks with high abrasion resistance and high strength are designed to protect the entire hot face in order to improve the service life of the refractory material on the hot face. The thickness was increased to 70 mm so that the SiC could be maintained as long as possible. In particular, the rib, where stress is concentrated and easily released in the old stave, was designed in a round shape to allow the stress to be dispersed so that it did not deform even at 33 MPa, where SiC brick was destroyed. The stave reflecting this design was first applied to the Pohang 4 blast furnace in November 2015, and it has been proved that the stave life has been significantly extended since it has been in operation without damage to date, despite severe operating conditions that could be fatal for stave damage in 2017.
