Manabu NOGUCHI*
Eiji ISHIKAWA**
Takashi KOGIN***
Suzue YONEDA****
Shigenari HAYASHI****
*
Technologies, R&D and Intellectual Property Division
**
Ebara Environmental Plant Co., Ltd
***
Dai-ichi High Frequency Co., Ltd
****
Faculty of Engineering, Hokkaido University
The self-fusing alloy spraying, which is performed as a countermeasure to the thinning of the in-bed tube of the fluidized bed boiler, requires periodic maintenance due to erosion-corrosion caused by the synergistic effects of corrosion caused by fuel-derived chlorine and erosion caused by the continuous impingement of the flowing medium. In order to reduce maintenance costs, we have developed a self-fusing alloy with excellent erosion-corrosion resistance. We have developed a small fluidized bed test apparatus that can simulate actual equipment, investigated the effects of constituent elements of SFNi4, a conventional material, developed an improved alloy, and proposed a erosion-corrosion mechanism.
Fluidized bed boilers are typical in biomass-burning power plants. The inside of the furnace of an internally circulating fluidized bed boiler provided by EBARA Environmental Plant Co., Ltd., is divided into a combustion chamber and a heat recovery chamber with a heat transfer tube (hereinafter referred to as an in-bed tube) which is the prominent characteristic of the internally circulating fluidized bed boiler (Fig. 1). The heat recovery chamber is capable of controlling the flow conditions independently to adjust the amount of heat recovery and has the function of maintaining and controlling the furnace bed temperature at an appropriate temperature. It also has the advantages of a heat exchanger with superior heat transfer efficiency compared to that in combustion gas, milder flow conditions than those in a combustion chamber, and an in-bed tube less subject to erosion.
Fig. 1 Conceptual scheme of internally circulating fluidized bed boilers
Fuel-derived corrosive elements such as chlorine become gas and are discharged from the furnace, but a part of them form chloride and remain in the fluidized bed. Corrosion promoted by the chloride and erosion caused by impinge of the fluidized medium damage the in-bed tube due to erosion-corrosion. The self-fusing alloy spraying performed on the surface of the in-bed tube to extend the service life of the fluidized bed boiler requires periodic maintenance and maintenance costs that may present an excessive burden plant operation.
In order to reduce maintenance costs, we have developed high-temperature erosion-corrosion-resistant materials and report the results.
Fig. 1 shows a photo of the appearance of an in-bed tube actually used in actual equipment. Ni base self-fusing alloy (JIS SFNi4) is sprayed on the surface. The white circle in the photo is where the sprayed coating disappeared to expose the carbon steel substrate. The reddish-brown appearance is due to the rust of carbon steel exposed during outdoor storage. As shown in the Figure, no corrosion products are observed on the in-bed tube due to the effect of the fluidized medium, resulting in a very smooth appearance similar to that of erosion thinning.
On the other hand, Fig. 2 shows the results of the analysis of adhered materials collected from areas with mild flow conditions. The deposits mainly consist of oxides such as Ni and Cr, which are the main components of thermal spray materials. The chloride, mainly composed of KCl, adhered to the surface. A small amount of Cl was observed in the corrosion products, confirming the occurrence of chlorination.
Fig. 2 Results of in-bed tube adhered materials analysis
Since the chloride is derived from fuel, it is concluded that a part of the fuel charged into the combustion chamber was carried to the heat recovery chamber. Fig. 3 shows the analysis results of the fluidized medium. The Cl content existed in the entire fluidized medium was as small as 0.1 mass% but concentrated to about 3 mass% in the periphery. The fluidized medium coated with such chloride works as a carrier and supplies chloride to the heat recovery chamber.
Fig. 3 Results of fluidized medium XRF and EDS analysis
As described above, the damage condition of the in-bed tube appears to be erosion. Fig. 4 shows the results of erosion-corrosion tests conducted in the past. 1) The dependence of test piece rotation speed (erosion conditions) on the thinning amount was evaluated under different conditions of the corrosive environment. It was observed that the thinning hardly progressed in the atmosphere (Air), but that the thinning was drastically promoted by severe erosion conditions in a corrosive environment (HCl + Salt). Observation of the damage condition of the test piece at this time (750 rpm, HCl+Salt; SUS347H) showed a smooth appearance which appears to be the erosion, indicating that thinning was drastically promoted by superimposing corrosion and erosion compared with corrosion alone (≒ low rotation speed) and erosion alone (≒ atmospheric atmosphere). Meanwhile, it was clarified that the damage condition has an appearance that appears to be the result of erosion alone.
Fig. 4 Results of rotary erosion-corrosion tests
Accordingly, the thinning of the in-bed tube, which appears to be erosion-like damage in the actual equipment, was probably caused by erosion-corrosion. It was confirmed in the actual equipment that the thinning amount of the in-bed tube greatly changes depending on the fuel, clarifying that the corrosive environment had a significant effect. It is clear that simple erosion does not cause thinning, supporting the above determination.
Based on past knowledge, we undertook to develop thermal spray materials to reduce maintenance costs for in-bed tubes. Creating evaluation methods is a key element. In the results of thinning in a corrosive environment shown in Fig.4, the weight loss of SUS347H steel was smaller than that of carbon steel (C.S.) at up to about 200 rpm, when the effect of corrosion was stronger, but the weight loss of carbon steel was smaller from 500 to 750 rpm. The order of materials may change depending on the magnitude of corrosion and erosion effects, requiring the establishment of evaluation methods that accurately reproduce the conditions of the actual equipment.
The prominent characteristics of the environment of the in-bed tube in the actual equipment are: a fluidized medium coating chlorides fluidizes at a flow condition of about 2 Umf (Umf; minimum fluidization velocity), a 700 °C atmosphere, an approximate 300 °C metal temperature, and the existence of a temperature gradient. To reproduce these characteristics, we have developed a fluidized bed test apparatus equipped with a test piece holder having an internal cooling mechanism. Fig. 5 shows the schematic diagram of the apparatus The test apparatus has a double-tube structure of quartz and inside is equipped with a reaction tube made of quartz with an inner diameter of 55 mm having a glass filter attached to the bottom. A specified amount of silica sand is put into the reaction tube and air blown in from a glass filter fluidizes the silica sand, which is the fluidized medium.
Fig. 5 Schematic diagram of test apparatus
The test piece is attached to the tip of a sample holder inside of which a copper block capable of water cooling is placed to adjust the depth of the cooling part; this allows the thermocouple temperature at the tip of the holder to be maintained at 300 ℃. Then, the test piece temperature is controlled by adjusting the cooling water flow rate while monitoring the thermocouple during the experiment. Considering that the thinning occurs preferentially in the direction of 45 ° against the impingement angle of the fluidized medium in the actual equipment, the test piece attaching part is set at an angle of 45 ° against the vertical direction.
Small fluidized bed erosion-corrosion test apparatus having these functions have been developed.
The test conditions which greatly effect the reproduction of the actual equipment include temperature, fluidization conditions (gas flow), fluidized medium, and the corrosive environment (concentration of Cl in the fluidized medium).
The corrosive environment was the most difficult condition to set. The thinning amount was greatly changed by the difference in the magnitude of erosion and corrosion effects, requiring great efforts to reproduce the experimental conditions. Fig. 6 shows the erosion-corrosion phenomena which are arranged qualitatively. 2) When corrosion is mild and erosion is dominant (erosion-dominant region), the thinning rate under actual operating conditions is small. Under severe corrosion conditions, the thinning amount increased (erosion-corrosion region) although no remarkable adhered material was observed on the test piece. However, when the corrosive environment became more severe, corrosion products covered the surface of the metal. Conversely, the thinning amount was suppressed (corrosion-dominant region). For example, at 750 rpm, as shown in Fig. 4, SUS347H with excellent corrosion resistance and a slow growth rate of corrosion products was in the erosion-corrosion region, while carbon steel, whose surface was covered with corrosion products, was in the corrosion-dominant region. The thinning part of the actual equipment is considered to be exposed to the erosion-corrosion region, requiring experimental reproduction of this region.
Fig. 6 Relationship between corrosion and erosion
Cl contained in the fluidized medium greatly affects the corrosion of the actual equipment. Therefore, we conducted a test by adding a chloride equivalent to that used in the actual equipment into the fluidized medium, but the test piece was hardly thinned. Analysis of the concentration of Cl in the fluidized medium showed that Cl continued to decrease immediately after the start of the test and disappeared within a short period of time, suggesting that the test was conducted under test conditions in the erosion-dominant region. When the method was changed to add the chloride including that expected to be lost from the beginning, the surface was covered with corrosion products and changed to the damage of the corrosion-dominant region, resulting in a much smaller amount of thinning than that expected in the actual equipment. From the above, it was found that reproducing the experiment in the erosion-corrosion region requires finding a method to maintain a constant concentration of chloride. For this reason, a larger amount of chloride than the target value of the experiment was added into the fluidized medium and heated and fluidized in an experimental apparatus for a short time. The fluidized medium with a stabilized concentration of Cl in the fluidized medium was prepared to mix with the unused product containing no Cl, thus adjusting the Cl content used in the experiment. However, the fluidized medium was pulverized during the experiment and its amount was reduced as its fine particles were scattered from the fluidized bed and a decrease in the concentration of Cl was inevitable. The experiment was continued by changing the fluidized medium every 50 hours. As a countermeasure for decreasing Cl, a crucible containing NaCl-KCl-CaCl2 mixed salt (melting point: 546 ℃) was placed in the lower part of the reaction tube to keep the surroundings above the melting point of the salt, and thus supplying chloride as vapor.
The laboratory equipment is capable of keeping flow conditions mostly constant. However, a slight condition change in the corrosive environment changes the thinning mode, requiring a very high level of know-how to ensure reproducibility. The establishment of the experimental conditions was a big highlight in developing the alloy.
In other words, it is very difficult to make the conditions constant because both flow conditions and corrosive environment fluctuate in the actual equipment. Therefore, developing materials with excellent robustness which show high characteristics even in various conditions is required.
Table 1 shows major experimental conditions.
The fluidized bed temperature and the test piece temperature were set in accordance with the operating conditions of the actual equipment, respectively. Considering that the Cl content in the actual equipment is 1 mass% or less in the fluidized medium, NaCl-KCl mixed salt was added to the fluidized medium to make the concentration of Cl at the start of the experiment 0.5 mass% or less. 3) The fluidized aeration was set at 25 L/min equivalents to 2.5 Umf, which is slightly accelerated from the average air in the heat recovery chamber.
Based on the JIS SFNi4 (Composition Table 2), which is generally adopted as a surface treatment alloy for the in-bed tube of actual equipment, a model alloy mainly composed of Ni-Cr-Fe was prepared by arc melting. Then, after homogenization heat treatment, surface polishing up to SiC polishing paper #1200 was conducted and used for the experiment.
After the completion of the experiment, the thickness of the test piece and weight change was measured to evaluate the thinning amount. The appearance and surface observation, and cross-sectional analysis were conducted as necessary.
SFNi4 is composed of various alloy components based on Ni, and B and C that form precipitates. First, the examination was conducted by focusing on alloy components excluding precipitate constituent elements (B, C) to examine the basic composition of the alloy. However, the magnitude of the Cu effect was judged to be small from the perspective of high-temperature erosion-corrosion -resistance and excluded from the beginning. The magnitude of the Cr, Mo, and Si effects was investigated. Fig. 7 shows the results of the investigations. Cr is a corrosion resistance improving element and was expected to improve the erosion-corrosion resistance but no significant effect was observed. Similar evaluations were conducted for Mo and Si, and it was found that the addition of both elements increased the thinning amount.
Fig. 7 Dependence of constituent elements on erosion-corrosion characteristics
Based on the above results, we developed alloys with the addition of B and C. Table 3 shows the compositions of the developed alloys. Cr with lower effects was set as high as 20% in consideration of corrosion resistance and Mo and Si were reduced. The results of corrosion tests conducted separately showed that Mo contributes to improving corrosion resistance. Therefore, the additive amount of Mo was stipulated at 1 %. Remelting treatment was applied to the alloy after thermal spraying to use it as a self-fusing alloy. Excessive melting points in the remelting treatment makes the process difficult. Si and B adjust the melting point and ensure workability. In this alloy, Si was reduced to the limit of production. Instead, the additive amount of B was increased and C was determined as the composition of the Ni-base developed alloy as equivalent to the conventional alloy.
Fig. 8 shows a photo of compositions of Ni-base developed alloys. The alloy consisted of a light-colored base metal phase (Ni-Fe-Cr+Ni3B phase) and a precipitate phase, consisting mainly of Cr carbides and borides. The photo of the surface after the erosion-corrosion test is also shown in the figure. After the test, the surface became an uneven shape, proving that the base metal phase was preferentially thinned.
Fig. 8 Results of SEM observation of Ni-base developed alloys
Conventional alloys are mainly composed of expensive Ni and the cost of raw material is one of the problems. To reduce the cost without lowering the performance, we examined replacing the expensive Ni with Fe for a Ni-base developed alloy. Similarly, Ni-Cr-Fe model alloys were prepared to conduct erosion-corrosion tests. However, the structural analysis shown in Fig. 8 proved that Cr exists as precipitate and the content in the base metal is reduced. Therefore, the concentration of Cr in the model alloy was also reduced to 5%.
Fig. 9 shows the results of the erosion-corrosion tests. Contrary to the assumption, it was proven that increasing Fe content drastically decreases the amount of erosion-corrosion. The presence or absence of Fe greatly changes the thickness of corrosion products. The thick corrosion products mainly consisting of Fe oxides grew in the 40 Fe alloy.
Fig. 9 Dependence of Fe contents on erosion-corrosion characteristics
On the other hand, increasing Fe deteriorates the manufacturability due to an increase in the melting point. To improve this, the alloy (Ni-Fe base; see Table 3) with 30% Fe content was developed by removing Mo, which is a metal with a high melting point.
Fig.10 shows the results of erosion-corrosion tests on the developed alloys. All of the developed alloys showed excellent erosion-corrosion resistance compared with the conventional material SFNi4. A large amount of precipitate existed in the newly developed Ni-Fe base as well as in the Ni-base, and its proportion was larger than that of the Ni-base.
Fig. 10 Evaluation results of developed alloys
Ni is generally known to be superior to Fe in corrosion resistance and shows excellent corrosion resistance, especially in chlorine-containing atmospheres.4), 5) The results of the experiment showed that increasing Fe improves erosion-corrosion resistance. We considered this cause in terms of both precipitates and Fe effects.
As shown in Fig. 8, the developed alloy is composed of base metal and a precipitate phase, and the base metal was preferentially thinned. The precipitate is a compound consisting of carbide and boride. It is less susceptible to corrosion compared with the base metal, which is a metal and is less susceptible to erosion because it is very hard. Therefore, the base metal is thought to have suppressed thinning. Ni-Fe alloys have a high ratio of precipitate, according to the image analysis based on histological metallographic observation. A rough ratio of SFNi4 and Ni-base was about 20 % of the proportion of the surface area while the ratio of Ni-Fe base increased to about 40%. This is one of the probable contributing factors to the improvement in erosion-corrosion resistance.
When the base metal is thinned and the surface becomes uneven, the fluidized medium is thought to preferentially impinge on the precipitates in the convex part. Then, the impingement of particles on the base metal is eased and the erosion conditions become mild, resulting in a decrease in the thinning rate of the base metal itself.
However, when the amount of precipitate is excessively increased, a large amount of cracks are generated in the precipitate by the test, confirming that the thinning amount rather increases. This means that an optimum ratio of the amount of precipitate to base metal is thought to exist.
Next, the detailed status of thinning was observed by TEM to examine the effect of Fe (Fig. 11). The cross-section of the Ni-base shows that the base metal part is concave and thinned preferentially, compared to the precipitate. Although the precipitate showed little corrosion, thin Ni-rich oxides were observed on the surface of the base metal part. The oxide which is a corrosion product is continuously destroyed by erosion, allowing only thin oxide to exist. As a result, the thin oxide scale could not keep sufficient protection against corrosion and is thought to maintain a high corrosion rate.
Fig. 11 Results of cross-sectional observation of developed alloys by TEM after erosion-corrosion
Increasing Fe content formed thick oxides on the surface of the model alloy (Fig. 9). Similarly, thicker oxides were formed in the Ni-Fe alloy than in the Ni-base. The growth rate of corrosion products composed of Fe, which is inferior to Ni in corrosion resistance, is fast. A thicker oxide scale was thought to have formed. As a result, the oxide layer protected the base metal, probably resulting in a decrease in the thinning rate.
Fig. 12 shows the erosion-corrosion mechanism summarized based on the above. Corrosion products grew on the surface of the base metal, which is more corrosive than the precipitate. However, the corrosion products were destroyed due to the continuous impingement by the fluidized medium. As a result, the corrosion rate increased (Stage 1). When the thinning of the base metal progressed precipitates protruded and formed unevenness on the surface. This resulted in the impingement of the fluidized medium with the protruding precipitate, and the eased erosion condition of the base metal (Stage 2). Easing the erosion condition grew corrosion products on the surface of the base metal, resulting in a decrease in the corrosion and thinning rates (Stage 3).
Fig. 12 Schematic diagram of the erosion-corrosion mechanism
However, when the corrosion resistance of the base metal was excellent, the growth rate of corrosion products was very slow. For this reason, it did not sufficiently function as a protective scale when the fluidized medium impinged on the base material. Therefore, a case in which the base material itself is scraped together with the corrosion product is thought to exist. The scratch marks on the surface of the Ni-base alloy shown in Fig. 8 are judged to be marks of direct scraping of the base material.
As described above, the erosion-corrosion rate are thought to be reduced by the existence of precipitate with a different thinning rate from that of the base material and the base material with a corrosion rate that plays a sufficient role as a protective scale.
We developed an erosion-corrosion resistant thermal spray material by laboratory test for the purpose of reducing the maintenance cost of the in-bed tube.
The key to the development is establishing evaluation methods that can qualitatively reproduce actual equipment. We developed a small fluidized bed test apparatus that reproduced the temperature gradient of the actual equipment. However, slight conditions can greatly change the results, leaving room for improvement in reproducibility. We used the apparatus and developed Ni-base alloys that exceed SFNi4, which is a conventional alloy, and Ni-Fe-base alloys. As a factor of thinning, it is concluded that the formation of an uneven shape by the precipitate and the protection of the base metal due to the formation of Fe oxide had large effects.
The demonstration experiment of the developed alloy is being continuously conducted and the results will be introduced in the second report.
1） M. Noguchi, H. Yakuwa, M. Miyasaka, H. Sakamoto, S. Kosugi: High temperature erosionHcorrosion behavior of boiler tube materials in fluidized‐bed waste incinerator conditions, In. Proceedings of HTCP2000, P.573-578. Hokkaido, (2000).
2） Manabu Noguchi, Hiroshi Yakuwa: Lecture on Fundamental Aspects of High Temperature Corrosion and Corrosion Protection, Part 4: (High Temperature Corrosion and Corrosion Protection in Incineration Plants), Ebara Engineering Review, No. 254, p. 29-40, (2017).
3） Mohammad Emami, Shigenari Hayashi, Takashi Kogin, Manabu Noguchi: Errosion-Corrosion Behavior of Metals in Chlorine Containing Oxidative Atmospheres, Proceedings of the 63rd Japan Conference on Materials and Environments，A-301，65 (2016).
4）Y Sato, M Hara, Y Shinada: Effect of a Small Amount of Hydrogen Chloride on High Temperature Oxidation of Pure Iron, J. Japan Inst. Metals, Vol.58, No.6, P. 654-659, (1994).
5）Y Sato, M Hara, Y Shinada: Effect of a Small Amount of Hydrogen Chloride on High Temperature Oxidation of Nickel, J. Japan Inst. Metals, Vol.60, No.2, P. 192-197, (1996).