Experimental investigation of erosion rate for gas-solid two-phase flow in 304 stainless/L245 carbon steel

Bingyun Hong ,Yno Li ,Xioping Li ,*,Gn Li ,Anong Hung ,Shuipng Ji ,Wiong Li ,Jing Gong ,**,Jin Guo

a National-Local Joint Engineering Laboratory of Harbor Oil ＆ Gas Storage and Transportation Technology,Zhejiang Provincial Key Laboratory of Petrochemical Pollution Control,Zhejiang Ocean University,Zhoushan,316022,PR China

b National Engineering Laboratory for Pipeline Safety,MOE Key Laboratory of Petroleum Engineering,Beijing Key Laboratory of Urban Oil and Gas Distribution Technology,China University of Petroleum-Beijing,Fuxue Road No.18,Changping District,Beijing,102249,PR China

c CNOOC Research Institute Co.Ltd,Beijing,100028,PR China

d PipeChina North Pipeline Company,LangFang,065000,PR China

e College of Chemical Engineering,Fuzhou University,Fuzhou,350108,PR China

Keywords:

ABSTRACT

1.Introductio n

Technological advances in hydraulic fracturing and horizontal drilling have driven the worldwide exploitation of shale gas(Hong et al.,2020).However,a large amount of particle will be carried in the process of shale gas exploitation(Hong et al.,2021a).During the flow of gas carrying particle in the pipeline,solid particles continuously impact the inner surface of the pipeline(Wang et al.,2021),and the material will gradually be removed from the inner wall of the pipeline and equipment(Jia et al.,2021).The erosion can easily lead to wall thinning,perforation leakage and other problems,and cause significant safety risks to the safe operation of pipelines(Hong et al.,2021b;Peng et al.,2021).

In recent years,the erosion problem in the oil and gas industry has attracted extensive attention of researchers,and three methods are employed to investigate erosion,including theoretical analysis(Parsi et al.,2014),numerical simulation(Abduljabbar et al.,2021)and experimental research(Wang et al.,2021).In theoretical analysis methods,the Lagrange approach is used to predict the particle trajectory,and then the first and second collision velocities are calculated(Kang and Liu,2020).Using the trajectory probability of a single particle,the probabilistic behavior of a large number of particles in the bend is derived,and the erosion distribution caused by the first and second collisions in the bend can be calculated(Kang and Liu,2020;Wang et al.,2017).As for the numerical simulation of erosion,it is mainly based on Computational Fluid Dynamics software to model the fluid flow process,turbulence models are used for fluid and particle flow simulation(Bilal et al.,2021),and empirical erosion models are used for erosion calculation(Abduljabbar et al.,2021;Alghurabi et al.,2021).Finally,the results of particle concentration,particle flow rate,wall collision velocity,wall erosion amount and so on can be obtained(Tang et al.,2021).

Many researchers have conducted erosion experiments to obtain more direct and intuitive results.The measurement of gas velocity and particle velocity is mainly through the particle image velocimetry(PIV),which uses the tracer particles in the flow field without interfering with the flow field(Novelletto Ricardo and Sommerfeld,2020;Wang et al.,2019).A.Mansouri et al.(2015)employed PIV to determine the slip velocity between gas and particle in horizontal and inclined pipes,and measured the local erosion depth with contour scanner to study the effects of erosion angle and gas velocity on erosion results.By comparing PIV and PTV,Lin et al.(2018)concluded that the PIV device has a relatively reliable calculation accuracy when measuring the flow field with high velocity and small particle size(Lin et al.,2015).The erosion degree is mainly evaluated by weight loss and wall thickness loss.However,Solnordal et al.(2015)believed that it difficult to obtain specific erosion conditions at specific locations by using weight loss evaluation.Therefore,they used a surface profilometer to record surface loss conditions of target materials.The experimental results showed that the maximum erosion location occurs between 55?and 60?of the elbows,and the depth of erosion decreases uniformly in all directions from this point.The three-dimensional scanner is used to accurately measure the erosion wall thickness changes of vulnerable parts such as elbows and ball valves(Zhang et al.,2015;Zheng et al.,2019).Wong et al.(2013)designed an experimental device to explore the erosion law of the cavity,and used the inner coating technology to qualitatively observe the erosion under the impact of sand and gravel.Vieira et al.(2016)used ultrasonic technology to measure the erosion at different points by marking 16 different points on the outside of the elbow.

The influencing factors of erosion mainly include gas velocity(Islam and Farhat,2014),particle size(Nguyen et al.,2016),material and shape of particles(Arabnejad et al.,2015;Xie et al.,2021),the impact angle(Nandre and Desale,2018;Nguyen et al.,2014)and so on(Javaheri et al.,2018).Oka and Yoshida(2005)believed that the degree of erosion damage is not only related to the flow field and the properties of particles,but also related to the target material(Oka et al.,2009).The erosion law is studied by the impact of SiC,SiO2and glass beads materials on aluminum,copper,carbon steel and stainless steel(Oka et al.,2005).Meanwhile,the experiments also investigated the degree of damage caused by the angle of impact on different materials,and the conclusion showed that the hardness of target material is an indispensable factor.Finally,a general erosion prediction equation which can be used in any impact condition and any type of material was proposed(Oka et al.,2005).Liu et al.(2015)showed that the erosion rate could be effectively reduced by increasing the bending radius of the elbow.When actual operating space is limited,a tee can be used instead of an elbow to reduce the erosion rate.Recently,Wang et al.(2021)performed an experimental and numerical study of slurry erosion behavior in a horizontal elbow and elbows in series.It was found that the maximum erosion location changes to the extrados of the elbow when velocity increases.The above researches have promoted the understanding and quantitative calculation of the gas-solid two-phase erosion law.However,the composition of particle in the study area is complex,and its erosion has not been explored in the existing model and experimental research.A vital knowledge gap exists in understanding the erosion between particles and pipelines of different materials in the gas field.

In this work,the self-designed experimental device is employed to investigate gas-solid two-phase erosion according to the on-site working conditions,including the on-site particle,304 stainless steel and L245 carbon steel.The effects of gas velocity,impact angle,erosion time and target material on the erosion rate are explored,and the apparent morphology is observed by Scanning Electron Microscope(SEM).By comparing the erosion experimental results with the existing erosion models,a modified model for different specimen materials is proposed to improve the accuracy of the model for meeting the engineering practical application standards.

The structure of this paper is as follows.In section 2,the materials and methods is explained in detail,including the apparatus and materials,experimental procedures and the definition of erosion speed and erosion rate.section 3 presents the results of the impact of typical factors on erosion rate,such as gas velocity,impact angle,erosion time,particle material and target material.Finally,the conclusion and future research prospects are drawn in section 4.

2.Materials and methods

2.1.Apparatus and materials

The experiment was carried out on a gas-solid two-phase erosion device,as shown in Fig.1.The device consists of four parts:compression system,feeding system,erosion system and particle recovery system.The air enters the pipe after being compressed by the twin-screw air compressor,and the flow rate of air is controlled by adjusting the opening of the ball valve.Then,the gas passes through the feeding system,mixes with solid particles,accelerates together,and then is ejected from the nozzle to erode the test specimen.Finally,the eroded particles are collected by the cyclone separator.The whole device is built of 1 inch pipe with inner diameter of 25 mm.

The compression system(Fig.2a)mainly includes two twinscrew air compressors(SA22A),buffer tanks,ball valves,flow meters(MF5712),etc.The maximum supply pressure is 0.85 MPa,and the air flow rate is 3.4 Nm3/min.The twin-screw air compressors are capable of supplying air continuously with tiny pressure fluctuation.They can run empty after the pressure reaches the maximum set value,and start automatically when the pressure drops below the minimum value.The gas is compressed into five gas buffer tanks that can withstand a pressure of 1 MPa,which can reduce the pressure fluctuation and store the air to ensure that the erosion device is supplied with sufficient and stable airflow.

The feeding system(Fig.2b)consists of feeding funnels,screws,motors and other structures.Particles are transported to the feeding port by the screw rod at the bottom of the feeding funnel.The speed of the motor can be controlled by changing the frequency of the motor,thereby changing the rotation speed of the screw.The higher the frequency,the faster the rotation speed and particle feeding speed,and the greater the mass flow rate of the particles.Moreover,the air flows in the pipe at low pressure,and the particles in the hopper are pushed into the pipe from the feed port through the hose under the action of the pressure differential to mix with the air.

The erosion system(Fig.2c),housed in a plexiglass container,consists of a nozzle,an eroded specimen,a specimen base,a glass cover,etc.The nozzle outlet diameter is 8 mm,a metal cylindrical specimen with a diameter of 30 mm and a thickness of 3 mm is fixed on a base in the erosion system.The impact angle of particles can be changed by adjusting the installation angle of the specimens between 0?and 90?.In addition,the glass cover outside the erosion system can effectively prevent particles from diffusing into the laboratory and polluting the environment of the lab.

Fig.1.Schematic diagram of the gas-solid two-phase erosion device.

The particle recovery system(Fig.2d)consists of two cyclone separators,a blower,and hoses.The two-stage cyclone separation can store and recover the particles used in the experiment more efficiently,which saves the experimental cost and avoids polluting the laboratory environment.Before the experiment,the blower is turned on to suck air into the device,and particles enters the cyclone separator under the action of pressure differential.

304 stainless steel and L245 carbon steel are selected based on the usage on site.304 stainless steel is a common material in stainless steel,containing more than 18% of chromium and more than 8% of nickel.It can withstand high temperature of 800?C,and has the characteristics of good processability,high toughness and corrosion resistance.Therefore,it is widely used in pipelines that require corrosion protection in the shale gas development industry.As for L245 carbon steel,it is mainly used to make seamless steel pipes for the transportation of natural gas and oil.The significant difference between the two is that 304 stainless steel contains chromium and nickel.The specific chemical composition is shown in Table 1.

Table1 Chemical compositions of 304 stainless steel and L245 carbon steel(%).

The specimens of 304 stainless steel and L245 carbon steel with the diameter of 30 mm were polished smooth with sandpaper before each experiment to avoid experimental errors caused by different surface roughness.A high-precision balance made by Metter Toledo,model XS205 with the accuracy of 0.1×10-3g was used to evaluate erosion through weight loss.

Two types of particles were used in the experiment,type I particle and type II particle.Type I particle consists of 60 mesh(150μm)quartz sand whose SEM picture is shown in Fig.3a.Type II particle from the production site is composed of 70% pottery and 30% quartz sand.The pottery is mainly spherical and with a size between 40 mesh and 70 mesh,while the particle size of type II particle is mainly between 70 mesh and 140 mesh.It was observed that the type II particle is irregular ellipsoid with convex surface.

2.2.Experimental procedures

The weighing method was used to evaluate the degree of erosion.In order to ensure the stability of the experimental results,each group of experiments was repeated three times under the same conditions,and the effective experimental results were selected and the average value was calculated.

Step 1 Before the start of the experiment,use sandpaper to polish the surface of the specimen smoothly,clean the surface of the metal chips,weigh the specimen and clamp it on the base.

Step 2 Set gas velocity as predetermined,and then adjust the motor frequency of the feeding system to make the particle enter the pipe as the set mass flow rate,mix with the gas,thereby corroding the specimen.Each group of experiments consumes a certain quality of particles,then the gas and power supply of the compressor is cut off to end the experiment.

Step 3 Take off and weigh the specimen,then calculate the erosion rate of this group of experiments.The erosion speed E and the erosion rate ER are used as criteria for measuring the erosion loss of the specimen.Erosion speed E is the ratio of mass loss to the time,while erosion rate ER is the ratio of mass loss to the mass of particles(Arabnejad et al.,2015),which are defined as follows.

where ms1is the mass of the specimen before the experiment,ms2is the mass of the specimen after the experiment,t is the total time of each group of experiments,and mpis the total mass of the particle used in each group of experiments.

It can be seen from the definition formula that the erosion speed E characterizes the speed of erosion and is related to time;the erosion rate ER characterizes the severity of erosion and is related to the quality of the particles.

Fig.2.Gas-solid two-phase erosion device:(a)compression system,(b)feeding system,(c)erosion system and(d)particle recovery system.

Step 4 Wash the oil and sand off the specimen surface with acetone,and then observe the changes of specimen surface with SEM to explore the erosion mechanism.

3.Results and discussion

3.1.Effect of particle mass flow rate on E ＆ ER

The 304 stainless steel was selected and the erosion angle was set to 90?.50 kg of type I particle was used in each group of experiments,which was mixed uniformly with gas in the pipe through the feeding system.The gas flow rate was measured by the flowmeter in front of the nozzle,and the rotational speed of the motor of the feeding system was adjusted to change the grit mass flow rate.The results shown in Fig.4a indicates that,(a)In the case of fixed conditions such as erosion speed,angle and particle quality,the erosion speed E of 304 stainless steel increases uniformly with the increase of the particle mass flow rate.For example,when the gas flow rate is 26.62 m/s,the erosion speed E changes from 0.012 mg/s to 0.025 mg/s as the mass flow rate changes from 4.79 g/s to 11.11 g/s.(b)When other conditions are fixed,the greater the gas velocity in the pipe,the greater the erosion speed E.In addition,the greater the gas velocity,the greater the erosion speed E changes with the particle mass flow rate,indicating that the erosion speed E with the particle mass flow rate are more sensitive to large gas velocity.

Fig.3.Experimental particle:(a)type I particle,(b)pottery of type II particle and(c)quartz sand of type II particle.

Taking the ratio of the erosion mass to the particle mass flow rate in Fig.4a,the correlation between the erosion rate and the particle mass flow rate can be obtained,as shown in Fig.4b.Under the specified gas velocity,the particle mass flow rate has little effect on ER,and even weakens the effect.That is,ER decreases with the increase of particle mass flow rate when other conditions are fixed.In other words,when other factors are fixed,with the particle mass flow rate increasing,the mass loss of the target material is roughly the same as long as the total mass of particle is fixed.

The main reasons for this phenomenon are the following three processes.First,when the gas velocity increases,the particles in the pipe have greater velocity under the carrying of gas,and according to the momentum theorem,the greater stress is generated when the particles impact the target surface.Therefore,when the gas velocity is large,the ER is greater.Second,with the increase of the particle mass flow rate,the particle on the cross-sectional area of the pipe will also increase,which will aggravate the collision of particle and consume the kinetic energy of particle.Therefore,the ER begins to decrease with the mass flow rate.Last,at a certain gas velocity,the particle-carrying capacity of the gas in the pipe is limited.When the mass of particle gradually increases,the particle cannot reach the gas velocity or cannot be blown up.Hence,the particle accumulates at the bottom of the pipe.As a result,the ER cannot increase with the particle mass flow rate gradually,but even decreases.

3.2.Effect of gas velocity on ER

The 304 stainless steel was selected and the erosion angle was set to 90?.50 kg of type I particle was used in each group of experiments.The particle mass flow rates were 10 g/s and 12 g/s,and the gas velocities were 26.20 m/s,28.82 m/s,31.43 m/s and 36.67 m/s.The relationship between ER and gas velocity of 304 stainless steel under impact angle 90?is shown in Fig.5.With the increase of gas velocity,the ER increases gradually.For example,when the gas velocity is 26.62 m/s,the ER is about 2.3 mg/kg.When the gas velocity gradually increases to 28.82 m/s,31.43 m/s and 36.67 m/s,the ER of the target material gradually increases from 2.8 mg/kg to 5.8 mg/kg.In addition,the difference between ER at mass flow rate of 12 g/s and that at mass flow rate of 10 g/s are minimal,which is also consistent with Section 3.1 that ER shows little relationship with particle mass flow rate.

The L245 carbon steel was selected and the erosion angle was set to 90?.50 kg of type II particle was used in each group of experiments and the particle mass flow rates were 10 g/s.The experiments were conducted under the gas velocity of 20.96 m/s,26.20 m/s,31.43 m/s,34.03 m/s,36.68 m/s and 41.92 m/s.As shown in Fig.6,similar to the 304 stainless steels,the ER of L245 carbon steel has the same change law with the change of gas velocity.

Extensive literature shows that there is an exponential relationship between erosion rate ER and gas velocity v:

Where Kmis a constant coefficient,n is the speed index and is related to the type of the material.

The power function is used to regress the curve in Figs.5 and 6,and Eq.(4)and Eq.(5)can be obtained,respectively:

For different types of materials,the value range of n is different.For plastic materials,the velocity index n is generally between 2 and 3.For brittle materials,the velocity index n is generally between 4 and 6.304 stainless steel is plastic material,and the value of n in Eq.(4)is 2.789,which is consistent with the results obtained by Oka et al.(2005),Oka and Yoshida(2005).

Fig.4.Effect of particle mass flow rate on E＆ ER under different gas velocity.

3.3.Effect of impact angle on ER

For each group of experiments,50 kg particle was used.The effect of the erosion angle on the ER was investigated at 15?intervals between 0?and 90?,and the microscopic surface morphology of the target material was observed at each erosion angle.Then,the ER at each erosion angle was normalized,i.e.,the erosion rate at each erosion angle was divided by the erosion rate at the erosion angle of 90?.

As shown in Fig.7,the ERs of 304 stainless steel and L245 carbon steel have the same change law under different impact angles.The ER of both materials reach the maximum at about 30?under the impacts of type I or type II particle.When the impact angle is less than 30?,the ER increases as the impact angle increases,while when the impact angle is greater than 30?,the ER decreases as the impact angle increases.In addition,under the same conditions,the ER caused by type I particle is higher than that by type II particle.By comparing the erosion morphology of the two types of particles at the same erosion angle,it can be found that the shape of type I is sharper,and the groove and the contour of erosion pit formed by type I are more obvious than type II,so the ER caused by type I is also larger.

Fig.5.Effect of gas velocity on ER of 304 stainless steel.

304 stainless steel and L245 carbon steel are ductile materials.According to Finnie's erosion theory(Finnie,1960),the erosion of ductile materials is mainly based on cutting at low angles.The wall surface morphology of 304 stainless steel and L245 carbon steel was observed by the SEM.The results show that 304 stainless steel and L245 carbon steel appear to be cut at low angles,and to be impacted at high angles to form erosion pits.As shown in Fig.8a and Fig.8b,when the particle washes the surface of the target material at the angle of 30?,the particle will cut the surface like a sharp knife.At the same time,grooves and lip edges will be formed on the surface washed away by the particle.According to Bitter's erosion theory(Bitter,1963a,1963b),when the particle impacts the surface at 90?,the speed of the particles hitting the surface of the material is fast,so the stress generated by the impact on the wall is greater than the yield strength of 304 stainless steel.Therefore,as shown in Fig.8(c)(d),the materials will gradually fall off the surface of the target material and erosion pits are formed.

Fig.6.Effect of gas velocity on ER of L245 carbon steel.

Fig.7.Effect of impact angle on ER.

The surface morphology of L245 carbon steel eroded by type I and II particle at different angles is provided in Appendix.Under low angles of erosion(impact angles less than 30?),grooves will be formed on the target material surface in the shape of strips cut by particles,meanwhile a lip is extruded around the grooves.When the impact angle gradually increases(impact angles greater than 30?),the grooves and lips edges gradually disappear and are replaced by erosion pits.

3.4.Effect of erosion time on ER

Fig.8.Surface morphology of 304 stainless steel,(a)(b)impact angle of 30?,(c)(d)impact angle of 90?.

Fig.9.Effect of erosion time on ER.

Set a specified sand mass flow rate,and the erosion time can be judged by the particle weight used in erosion experiment.The heavier the mass of particle used,the longer the erosion time.The experimental results of 304 and L245 carbon steel specimen are shown in Fig.9.The longer the erosion time,the greater the quality loss of the specimen and the more serious the erosion.However,the erosion rate is a relative ratio,which is the ratio of mass loss to the mass of eroded particles.It is found through experiments that the erosion time has no significant effect on the erosion rate.Previous studies have shown that the ER will gradually increase from 0 and then gradually rise with the increase of sand weight,until it finally reaches a state where ER tends to be stable.The former stage is called the erosion latent period,and the latter stage is called the erosion stable period.When a small amount of particle erodes the surface of the ductile material specimen,some of the material on the specimen surface does not fall off immediately,but forms a lip around the groove.A specified gas velocity,the particle-carrying capacity of the gas in the pipe is limited.Therefore,the ER in the initial stage of erosion,i.e.,erosion latent period is low.As the particle keeps hitting the surface of the specimen,the lip comes off and the erosion reaches a stable stage.Since the time of“erosion latent period”is very short for this experiment and the actual field operation,the ER is not affected by the erosion time when the erosion conditions remain stable.

Fig.10.Comparison between experimental data and calculated results.

Fig.11.Comparison between experimental data and calculated results.

3.5.Modification of erosion model

Three conventional erosion models E/CRC model(Zhang et al.,2007),Ahlert model(Ahlert,1994),DNV model(Peng and Cao,2016;Song et al.,2013;Veritas,2007)were conducted under experimental conditions,the mass flow rate of 0.01 kg/s,gas velocity of 26.62 m/s,erosion angle of 90?,and 304 stainless steel specimens.The comparison between experimental data and calculated results are shown in Fig.10.The maximum impact angle is between 40?and 50?,55?-65?and 25?-35?obtained by the E/CRC model,DNV model and Ahlert model,respectively.The predicted accuracy of DNV model is the lowest among the three models.There is a large deviation between the predicted value and the experimental data under various conditions,where the maximum relative error reaches 61.18%.Therefore,it's obvious that the DNV model cannot meet the prediction criteria.In addition,the average relative tolerance rates of E/CRC model and Ahlert model are 19.68%and 22.41%,respectively.The above three models fail to meet the industrial criteria.Hence,the E/CRC model with the smallest relative tolerance was selected for modification.

Fig.12.Comparison between experimental data and calculated results.

The E/CRC model is modified based on the experimental data of 304 stainless steel as follows:

where BH is Brinell hardness,Fsis gravel shape factor,vpis particle impact velocity,αis particle impact angle,and

The calculated results of modified E/CRC model are shown in Fig.11.The maximum relative tolerance is 1.78%under this working condition,which is much smaller than that of the original E/CRC model.The simulation results of the modified E/CRC model have smaller tolerances and can make accurate prediction of the erosion results under this experimental condition.

Similarly,the E/CRC model is modified based on the experimental data of L245 carbon steel as follows:

where

The calculated results of E/CRC model after modification are shown in Fig.12.The maximum relative error is 3.95% at this experimental condition.The relative error in this set of experimental conditions is much smaller than that of the original E/CRC model.The simulation results of the modified E/CRC model have smaller errors and can make accurate prediction of the erosion results under this experimental condition.

4.Conclusions

(1)A gas-solid two-phase erosion device is designed and manufactured,which consists of air supply system,feeding system,erosion system and particle recovery system.The device can flexibly control the gas flow rate and particle mass rate,and it is convenient to alternate the specimen and adjust the angle.

(2)Due to the smaller hardness of L245 carbon steel,the erosion of L245 carbon steel is severer than that of 304 stainless steels under the same conditions.With the particle mass flow rate increasing,as long as the total mass of particle is fixed,the mass loss of the target material is roughly the same.The ER of both materials reach the maximum at about 30?under the impacts of type I or type II particle.304 stainless steel and L245 carbon steel appear to be cut at low angles,and to be impacted at high angles to form erosion pits.In particular,under low angles of erosion(impact angles less than 30?),grooves will be formed on the target material surface in the shape of strips cut by particle,meanwhile a lip is extruded around the grooves.When the impact angle gradually increases(impact angles greater than 30?),the grooves and lips edges gradually disappear and are replaced by erosion pits.Since the time of“erosion latent period”is very short for this experiment and the actual field operation,the ER is not affected by the erosion time in the steady operational state.

(3)Bycomparing the calculation results of the E/CRCmodel,Ahlert model and DNV model to the experimental data,the tolerance of each model is large and cannot meet the industry criteria.Therefore,the modified E/CRC model is proposed for 304 stainless steel and L245 carbon steel,respectively.The average relative error of the modified models for the two target materials is within 3%,which can make accurate prediction of the erosion results and meet the industrial requirements.

(4)Shale gas in gathering and transmission pipelines usually contains small solid particles and acid liquids such as dissolved CO2.Therefore,the coexistence of solid particle erosion and acid corrosion is a common phenomenon in shale gas gathering and transmission pipeline.In the future,we will further transform the experimental device to study the synergistic mechanism of erosion and corrosion.Nevertheless,our findings provide a theoretical support for the erosion of two-phase gas-solid flow in 304 stainless/L245 carbon steel to solve the flow assurance issue in the oil＆ gas pipelines.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This work was supported by the Zhejiang Province Key Research and Development Plan(2021C03152),Zhoushan Science and Technology Project(2021C21011),Industrial Project of Public Technology Research of Zhejiang Province Science and Technology Department(LGG18E040001),and Scientific Research Project of Zhejiang Province Education Department(Y20173854),all of which are gratefully acknowledged.Thanks to Zhouyin He and Yulin Liu for English language editing.

Appendix

Fig.A1.Surface morphology of L245 carbon steel at 15?(type I particle).

Fig.A2.Surface morphology of L245 carbon steel at 30?(type I particle).

Fig.A3.Surface morphology of L245 carbon steel at 45?(type I particle).

Fig.A5.Surface morphology of L245 carbon steel at 75?(type I particle).

Fig.A6.Surface morphology of L245 carbon steel at 90?(type I particle).

Fig.A7.Surface morphology of L245 carbon steel at 15?(type II particle).

Fig.A8.Surface morphology of L245 carbon steel at 30?(type II particle).

Fig.A9.Surface morphology of L245 carbon steel at 45?(type II particle).

Fig.A10.Surface morphology of L245 carbon steel at 60?(type II particle).

Fig.A11.Surface morphology of L245 carbon steel at 75?(type II particle).