Fatigue tests of heat-treated rails in the R350HT grade

Modern operational requirements related to increased track reliability and the constant increase in train speeds along with increasing axial loads observed for many years place high demands on the metallurgical process, and in particular, the rail heat treatment process. A modern technological process of rail head hardening should ensure a different level of properties depending on the rail cross-section, i.e. high hardness of the rail head, a ductile web and a hardened foot [1]. ArcelorMittal Poland S.A. Branch in Dąbrowa Górnicza launched a rail head heat treatment line implemented as part of the research and development project POIR.01.01.01-00-0438/18 co-financed by the National Centre for Research and Development Poland. The demonstration line enables heat treatment of the rail head by accelerated cooling of the head in a water–polymer mixture while blowing the water–air mixture onto the rail foot to compensate for stresses and obtain the straightest possible rail after the technological heat treatment operation. Photos of the heat treatment line are presented in Figure 1. It shows the process of accelerated cooling of the rail head, the final length of which after all technological operations is 120 m. The line for heat treatment of rail heads consists of two tanks with a water–polymer mixture, collectors for controlled cooling of the rail foot, mechanical grippers for proper positioning of the rail during immersion, and measuring equipment. The key process parameters that are monitored during the heat treatment operation are the temperature of the rail strip before immersion in the water–polymer mixture, the polymer concentration, the temperature of the water-polymer mixture, the coolant flow and the temperature of the rail strip after the heat treatment process. The above parameters are individually selected for a given type of rail and steel grade, essential for obtaining repeatable properties and appropriate microstructure of the rails. An important element of the process is also the sequential water–air spray applied to the rail foot during the immersion of the rail head in the cooling medium, which allows for minimizing the temperature difference between the rail head and the rail foot after the heat treatment process, and thus, contributes to the compensation of residual stresses.

Figure 1

Rail heat treatment line.

Railway rails are characterized by a pearlitic structure, which determines their properties. The aim of the heat treatment process is to modify the morphology of perlite in such a way that, as a result of accelerated cooling, the structure of very fine pearlite is obtained. Both not heat-treated and heat-treated rails have a pearlitic structure, but the difference between them is the significant fragmentation of the pearlite in heat-treated rails. As a result of heat treatment, we obtain a very fine pearlite structure in the area of the rail subjected to accelerated cooling, i.e. in the rail head. Pearlite consists of alternating lamellas of ferrite and cementite with an interlamellar distance ranging from 0.2 to 0.25 μm for untreated rails and 0.05 to 0.15 μm for heat-treated rails [2]. Reducing the distance between the lamellas not only increases the strength of the rails (increase in tensile strength, increase in hardness, increase in fatigue strength) but also improves the operational properties of the rails by increasing wear resistance, increasing resistance to rolling-contact fatigue (RCF) defects and increasing the level of the stress intensity factor K _Ic.

Heat-treated rails must have a fully pearlitic structure over the entire cross-section, without any martensite or bainite precipitation. It is assumed that hardness is a measure of the abrasion resistance of the pearlitic structure and each increase in the rail hardness by 40 HBW increases the abrasion resistance by an index of 1.75 [3]. The benefits of higher rail hardness include higher permissible axial loads, lower lateral and vertical head wear, higher transferred total gross loads/year and subsequent generation of operational defects on the rail running surface. The minimum hardness of the rail on the running surface in the standard, not heat-treated R260 grade is 260 HBW, according to the European standard EN 13674-1:2011 + A1:2017 [4], while the minimum hardness of heat-treated rails in the R350HT grade is defined at 350 HBW, which means that heat-treated rails have almost four times higher abrasion resistance compared to untreated rails. This extends operation, especially in applications such as track bends, intensive use (high frequency of trains, e.g. metro), high unit axle loads (tracks for transporting natural resources, e.g. iron ore), high speeds. The production of rails with a fine-lamellar perlite structure is an important factor improving the operational durability and safety of railway traffic. However, before heat-treated rails are installed in tracks, they must undergo very rigorous tests in the field of qualifying tests described in point 8 of the standard [4]. The main element of qualifying tests is fatigue tests, for which the standard in question requires three types of tests, i.e. determining the stress intensity factor, fatigue crack growth rate and fatigue strength. For each of these tests, high parameters are specified to be met in a strictly defined test procedure. Only rails with appropriate macro- and microstructure parameters and other properties can pass the above tests with a positive result. An important element of the qualifying tests is also the determination of the level of residual stresses in the rail foot. The importance of residual stresses in rails is related to their direct impact on the behaviour of the rail on the track and, consequently, on the safety of railway traffic. The lower the residual stresses in the rail, the lower the average operational stresses, which consequently leads to slower crack development. If the tensile stresses reach a critical value of the stress intensity factor K _Ic (determining the resistance to brittle fracture), the crack develops in an uncontrolled manner, leading to a rail fracture [5]. Therefore, rail manufacturers try to use various methods to reduce the level of residual stresses in rails inherited from the production process to ensure their high operational durability [6,7]. The available literature includes studies broadly analysing the impact of heat treatment parameters on the obtained level of rail properties [8,9,10,11,12], analyses of microstructure parameters of heat-treated rails, taking into account the prediction of their level on the operational properties of rails [13,14,15,16,17], as well as studies devoted to determining stress intensity factor in rail steel [18,19,20,21]. However, no comprehensive summary of all types of tests in the field of qualifying tests is provided in the standard [4] along with a discussion of their mutual correlation in terms of the operational properties of rails. This work describes the test procedure for the entire scope of qualifying tests, presents the results of tests of mechanical properties, assessment of macro and microstructure, residual stresses, and presents the level of fatigue tests obtained on a 60E1 rail profile in the R350HT steel grade. The scientific objective of the work is to demonstrate the stability of the technological line for heat treatment of the rail head by analysing samples taken from different rails coming from different melts and conducting tests in independent research laboratories of a number of tests within the scope of qualification tests described in the standard [4]. Such an approach will enable the demonstration of the stability of the heat treatment process by confirming the fulfilment of the qualification criteria and the repeatability of the measured parameters, which will be direct evidence of obtaining the proper microstructure of the heat-treated rails.

The European standard EN13674-1 [4] defines four grades of rail steel (R350HT, R350LHT, R370CrHT, R400HT) with different properties dedicated to the heat treatment of rails, which are presented in Table 1. All these grades must be characterized by a fully pearlitic structure, while the level of increased mechanical properties compared to the standard, untreated grade R260 is obtained by modifying the morphology of pearlite [22], especially by reducing the distance between the cementite lamellas in pearlite and changing the thickness of the cementite lamellas [13].

The above table shows that the difference in chemical composition between the basic heat-treatable rail steel grade R350HT and grade R260 is small, the only difference being that grade R350HT has a higher minimum carbon content than grade R260. The difference in chemical composition between the R350HT and R350LHT grades is twice the permissible chromium content in the R350LHT grade, with the same requirements for the level of mechanical properties and hardness. The R370CrHT grade is characterized primarily by the highest permissible level of chromium of all grades for heat treatment and an increased mass fraction of silicon, which ensures a hardness in the range of 370–410 HBW and a higher level of tensile strength of at least 1,280 MPa. The highest grade of rail steel currently used in terms of hardness is the R400HT grade, which is a hypereutectic steel with a reference hardness in the range of 400–440 HBW. The use of such high hardness on railway rails results from the constant pressure to improve railway efficiency by increasing axial loads, train speeds and rail frequency [23]. The increase in requirements was due to higher loads on the rail–wheel system and greater degradation of track elements, and thus increased rail wear and increasing problems caused by RCF defects. Currently occurring fatigue processes at the wheel–rail contact are one of the biggest problems of railway companies, causing degradation of the running surface, manifesting itself primarily in the form of defects such as head checking, squat, and spalling [23,24,25]. The remedy for excessive rail wear, especially in the area of ​​the rail head corner, and for the development of RCF defects is the use of harder steel grades, mainly rail steels intended for heat treatment. The research results presented in the article by Girsch et al. [23] indicate that the higher the grade of rail steel, the later the appearance of head checking defects and, more importantly, the smaller the degradation caused by the development of cracks. Additionally, these studies showed a significant improvement in steels with high hardness levels in terms of their resistance to the formation of corrugated wear.

The test material was R350HT rail steel with the chemical composition, according to the melt analysis, given in Table 2. The test samples came from heat-treated rails according to the developed technology and produced in the heavy section mill of ArcelorMittal Poland S.A. on a demonstration line for heat treatment using accelerated cooling of the rail head in a water–polymer mixture. The following heat treatment parameters were used: –

immersion start temperature 810°C,

For such selected parameters of the heat treatment process, fine lamellar perlite was obtained, characterized by an average interlamellar distance of 95 nm in the perlite measured at the corner of the head. From the CTPc diagram developed for the chemical composition corresponding to the range for the R350HT grade, it is known that the heat dissipation rate should be in the range of 1–2°C/s. This ensures the avoidance of bainite precipitation and the achievement of HV pearlite hardness in the range of 370–400. An example image of the microstructure on the cross-section obtained after industrial hardening tests of the corner of the rail head is shown in Figure 2. The structure of the tested rails was fully pearlitic, without bainite or martensite precipitation. The size of the pearlite colony in the head was in the range of 3.4–9.28 µm. Average size of perlite colony in the rail head was 6.34 µm. The measurements of the perlite colony size were carried out using specialist μgrain computer software. The program allows the determination of the perlite colony size based on measurement lines drawn through the boundaries of the perlite colony. In order to determine the size of the perlite colony, the following relationship was used: d p = ( d 1 + d 2 ) / 2 , dp=(d1+d2)/2, where dp is the perlite colony size, μm, d1 is the maximum colony length, μm, and d2 is the maximum colony width, μm.

Figure 2

Image of the microstructure at the corner of the rail heads after heat treatment: (a) rail from melt no. 1 and (b) rail from melt no. 2.

The correctness of the structure was checked on all discussed samples numbered 1–7, which were the subject of the evaluation of the qualification tests, in no case were undesirable bainite or martensite precipitations recorded, which indicates the consistency and effectiveness of the heat treatment process and its repeatability leading to obtaining fine lamellar pearlite on the entire cross-section of the rail head. Due to the fact that the range of chemical composition for each melt was maintained in a very narrow range for all elements, it can be assumed that the material was quite homogeneous.

The tests used rails made of seven different melts, which are characterized by the properties given in Table 3. The parameters R _m, R _p0.2, A and hardness on the running surface were determined. The predestined method for measuring hardness in rails is the Brinell method using a ball with a diameter of 2.5 mm and a pressure force of 187.5 N according to the EN ISO 6506-1 standard [26]. Samples for qualification tests in terms of stress intensity factor, fatigue crack growth rate and fatigue strength according to the standard [4] must be taken from different rails from at least three different melts. For the qualification tests, it was decided to take samples from seven different melts with the chemical composition given in Table 2 in order to check a wider spectrum of material with different chemical compositions resulting from the actual technological process carried out in production conditions. However, all samples came from rails that were subjected to heat treatment with constant process parameters. This approach allowed for the verification of the stability of the heat treatment line operating parameters in a wider range than required by the standard [4].

Samples were taken on rails from three melts to evaluate Baumann’s prints, which determine sulphur segregation on the cross-section of the rail. None of Baumann’s prints exceeded the permissible patterns of the sulphur segregation assessment scale. All prints were graded according to the D2 pattern, except one, which was graded according to the D4 pattern. Examples of Baumann prints for the tested rails are presented in Figures 3–5.

Figure 3

Baumann’s print: (a) Sample 5 – pattern D2 and (b) sample 5 – pattern D2. The area of slight negative and positive sulphur segregation in the rail web is marked.

Figure 4

Baumann’s print: (a) Sample 6 – pattern D4 and (b) sample 6 – pattern D2. The area of slight negative sulphur segregation in the rail web is marked.

Figure 5

Baumann’s print: (a) Sample 7 – pattern D2 and (b) sample 7 – pattern D2. The area of positive sulphur segregation with a small amount of negative sulphur segregation in the rail web is marked.

Residual stresses in the rail foot were also measured on six samples from different rails. Residual stress tests were carried out in accordance with the procedure described in Annex C of the standard [4]. The measured level of residual stresses for all tests was in the range of 66–116 MPa, and the average value was 97 MPa. This is a very low level of residual stress and a good symptom of operational properties because the level of residual stress inherited from the production process has a direct impact on the behaviour of the rail on the track, in particular on the maximum size of a critical defect, and thus affects the safety of railway traffic. For heat-treated rails, in addition to measuring the hardness on the running surface, it is also necessary to measure the hardness on the cross-section of the rail head at points precisely defined by the standard, as shown in Figure 6(a). This is due to the need to determine the gradient of hardness drop in the inner area of the rail head and the requirement to ensure a minimum level of hardness at defined points for operational reasons. The obtained hardness level on the cross-section of the head of sample number 4 as an example is shown in Figure 6(b). As can be seen, this rail is characterized by a slight decrease in hardness at a distance of 10 and 20 mm from the running surface and a symmetrical hardness distribution relative to the vertical axis of symmetry of the rail. This hardness distribution will ensure even wear of the rail during operation because the rail head is not only hardened in the surface layer, but also has a high level of hardness throughout its entire volume.

Figure 6

Hardness distribution on the cross-section of the rail head for the R350HT grade according to the standard [4] (a), and hardness distribution on the cross-section of the rail head – sample 4 (b).

Fatigue tests belong to the group of qualifying tests according to the standard [4] and include determining the stress intensity factor K _Ic, the fatigue crack growth rate and testing the fatigue strength of the rail material. In addition to the above-mentioned tests, the scope of qualifying tests also includes tests of residual stresses in the rail foot, tests of hardness, tensile strength and elongation, and sulphur segregation distribution. Qualifying tests should be carried out at least once every 5 years, and also at each significant change in the production process for all steel grades. These tests are carried out on a 60E1 type rail profile or the heaviest produced one. Test sections must come from at least three different melts in a given steel grade and from different strands of a continuous steel casting machine. Fatigue tests should be taken from three different rails, at a distance of not less than 3 m from the ends of the rails.

The eligibility criteria for individual fatigue tests are as follows.

4.1

Fatigue crack growth rate

The test is performed in accordance with the general requirements contained in the ISO 12108 standard [27]. The fatigue crack growth rate (m/Gc) for the parameter ΔK = 10 MPa m^1/2 should not exceed 17 m/Gc for R350HT steel, while for the parameter ΔK = 13.5 MPa m^1/2, it should not be greater than 55 m/Gc. A minimum of three tests were performed for each rail under the following conditions: –

test temperature range +15 to +25°C,

–

ratio of minimum cyclic load to maximum cyclic load R = 0.5,

–

the load cycle frequency should be in the range of 15–40 Hz,

–

the spacing of supports for the three-point bending of the sample is to be 4 W.

Samples with dimensions shown in Figure 7 were used for the test.

Figure 7

Dimensions of the sample for testing the fatigue crack growth rate.

4.2

Fracture toughness K _Ic

For steel grade R350HT, the minimum single K _Ic value should be greater than 30 MPa m^1/2 and the minimum average value should exceed 32 MPa m^1/2. The determination of the stress intensity factor shall be performed in accordance with the procedure described in ASTM E399 [28], except that those requirements are superseded by those specified in Annex B of the standard [4] and apply only to the determination of the plane strain fracture toughness of rail grade steels defined in this standard. The initial fatigue crack of the sample should be generated in the temperature range of +15 to +25°C using a stress factor in the range >0 <+0.1. Initial fatigue cracking occurs at a cycle frequency ranging from 15 to 120 Hz. The ratio of the total crack length to the sample width should be from 0.45 to 0.55, and during the last 1.25 mm of crack growth, K _max is required to be in the range of 18–22 MPa m^1/2. A bending specimen with a single notch on the edge is loaded with displacement control using a three-point bending method. The support spacing is defined as four times the width (W) of the sample. The test is carried out at a temperature of −20°C ± 2°C. The dimensions of the samples for determining the stress intensity factor are presented in Figure 8.

Figure 8

Dimensions of the sample for determining the K _Ic factor.

The process of generating the initial fatigue crack (before the fracture mechanics study) was carried out at room temperature in the stress intensity factor (K) control mode via special CTOD Fracture Toughness (INSTRON) control software. The pre-crack generation procedure ensures that the maximum stress intensity factor (K _max) remains below the permissible limits. The amplitude for initial crack generation was sinusoidal with a frequency of 15 Hz and an index R = P _min/P _max = 0.09. Pre-cracking was stopped when the crack length (notch + fatigue crack) reached a value between 0.45W and 0.55W (W = sample width). A strain gauge was used to measure the crack length in the notch hole, and the crack length was obtained using the compliance method. The K _Ic test was performed in a temperature-controlled chamber using liquid nitrogen as a coolant. The test temperature was measured on the front surface of the sample, near the fracture site, using a contact thermocouple, waiting for temperature stabilization for at least 45 min for each sample when the temperature reached −20°C ± 2°C, and ensuring a stable temperature at all times during the test. All tests were performed using special control software that recorded the load (P) and the crack displacement opening. The test plots showed linear elastic behaviour, consistent with the requirements of linear-elastic fracture mechanics for brittle fracture of the specimens. After the sample cracked, a direct measurement of the crack front was performed according to the ASTM E399 standard [28], and the result was entered into the testing software to calculate K _Ic and all validations required by the standard.

The determination of the fatigue crack growth rate and the stress intensity factor K _Ic was performed by the Idonial Institute in Spain. The fatigue strength test was conducted by the Railway Institute in Warsaw. Due to the volume of the work, only selected results were presented.

5.1

Fatigue crack growth rate

The fatigue crack growth rate test was performed in accordance with the ISO 12108 standard [27]. The tests were carried out at room temperature. During the test, the load factor (R = P _min/P _max) was 0.5 in accordance with the standard [4]. The equation of Paris’s law in zone II was determined using the following data pairs: crack growth rate da/dN (mm/cycle) in relation to the range of stress intensity factor ΔK (MPa m^1/2), generated after applying a controlled sinusoidal loads with a frequency of 15 Hz. The initial fatigue crack was generated at room temperature by recording the following parameters: load frequency, ΔK, K _max, crack size and total number of cycles. A strain gauge was used to measure the crack. Data pairs da/dN (mm/cycle) with respect to ΔK (MPa m^1/2) were recorded every 0.08 mm of crack growth during the test. A double logarithmic zone plot was constructed from the recorded data and the line graph and fitted to the potential equation by calculating the parameters “C” and “m” in accordance with Paris’ law da/dN = C (ΔK) ^m (Figure 9).

Figure 9

Fatigue crack growth rate – Paris’ law.

Photos of the fracture surfaces for samples 2a, 2b, 2c are presented in Figure 10. Based on the observations of the fractures of the discussed samples, it can be concluded that the fatigue fracture is ductile, and the fatigue scrap zone for the discussed samples is approximately 15–17 mm.

Figure 10

Sample fractures for determining the fatigue crack growth rate representing. The photos represent the following samples: (a) sample 2a, (b) sample 2b and (c) sample 3c.

The data used to determine the fatigue crack growth rate for the example rail 321037 A201 (sample 2) are given in Table 4.

Table 4

Conditions for determining the fatigue crack growth rate.

Fatigue crack growth rate
Sample no.	2a	2b	2c
Load rate R	0.5	0.5	0.5
Amplitude type	Sinusoidal	Sinusoidal	Sinusoidal
Load frequency (Hz)	15	15	15
Initial ΔK (MPa m1/2)	9.32	9.39	9.27
Initial K max (MPa m1/2)	18.64	18.78	18.54
Initial ΔP (kN)	6605.00	6600.00	6608.00
Final crack length (mm)	27.26	26.93	27.75
Temperature (°C)	Room temp.	Room temp.	Room temp.
	Measurement of the crack front
Average a (mm)	25.24	24.47	24.69
Standard deviation (mm)	1.0041	1.1314	0.3606
Paris’ law (da/dN) = C(ΔK) m
C	2.0673 × 107	4.8451 × 108	8.5625 × 108
m	1.92	2.45	2.26
R m (MPa)	1,292
R p0,2 (MPa)	894

Logarithmic charts showing the dependence of da/dN on ΔK are shown in Figures 11–13.

Figure 11

Graph showing the dependence of da/dN on ΔK determined for the sample 2a.

Figure 12

Graph showing the dependence of da/dN on ΔK determined for the sample 2b.

Figure 13

Graph showing the dependence of da/dN on ΔK determined for the sample 2c.

The obtained results of the fatigue crack growth rate for all tested samples are presented in Table 5. Each sample meets the requirements of the standard [4].

Table 5

Results of testing the fatigue crack growth rate.

Sample no.	da/dN = C (ΔK) m		Indicator m/Gc ΔK = 10, MPa m1/2	Indicator m/Gc ΔK = 13.5, MPa m1/2
	C	m
2a	2.0673 × 107	1.9153	17.00	30.22
2b	4.8851 × 108	2.4538	13.78	28.77
2c	8.5625 × 108	2.2583	15.52	30.57
1a	1.4207 × 107	1.9498	12.66	22.72
1b	1.1649 × 107	2.0288	12.45	22.88
1c	2.8016 × 107	1.6677	13.03	21.50
4a	1.3571 × 107	2.1066	17.00	32.64
4b	1.4220 × 107	2.0735	16.84	31.38
4c	1.4038 × 107	2.0718	16.56	30.84
	Requirements of the EN13674-1 standard		Max 17 m/Gc	Max 55 m/Gc

5.2

Stress intensity factor

The determination of the stress intensity factor was carried out according to the procedure described in the ASTM E399 standard [28] taking into account the limitations imposed by the EN13674-1 standard [4]. Test results in accordance with the standard [4] should be presented as K _Ic, K Q ⁎ {K}_{\text{Q}}^{\ast } values or K _Q values, where K Q ⁎ {K}_{\text{Q}}^{\ast } values are those that do not meet the correctness criteria due to one or more of the following reasons:

P _max/P _Q > 1.1;

Exceeding the criterion 2.5(K _Q/R _p0.2);

Displacement-crack opening force relationship.

Mean values and standard deviations for both K _Ic and K Q ⁎ {K}_{\text{Q}}^{\ast } values should be recorded. The value used for the acceptance criteria is the average K _Ic value and should be based on at least five K _Ic values. If five K _Ic values are not obtained, any K Q ⁎ {K}_{\text{Q}}^{\ast } values should be included in the acceptance criterion along with the average K _Ic values. However, in this case, the measurement results should be based on a minimum of ten measurements. An exemplary K _Ic acceptance criterion for sample 3 is given in Tables 6 and 7.

Table 6

Validation requirements for 321038 A101 rail (samples 3).

Determining K Ic
Sample no.		3a	3b	3c	3d	3e
R m (MPa)		1,331
R p0.2 (MPa)		696
Test temperature (°C)		−20	−20	−20	−20	−20
P max (kN)		18.3443	18.5739	18.2744	18.7456	18.7456
P Q (kN)		17.81	18.39	17.24	18.54	18.56
Time test (s)		35.21	61.15	71.01	76.27	69.1
K Q (MPa m1/2)		39.2	40.27	37.59	41.27	40.97
Rate test (MPa m1/2/s)		1.11	0.66	0.53	0.54	0.59
Mean value of a		19.81	19.83	19.91	20.11	19.53
Standard dev. of a		1.1066	1.1580	0.9459	0.8739	1.2609
0.45 ≤ (a/W) ≤ 0.55		0.50	0.50	0.50	0.50	0.50
P max/P Q < 1.1		1.03	1.01	1.06	1.01	1.01
K max < 0.6 K Q	K max (MPa m1/2)	22.08	22.2	22.2	22.1	22.1
	0.6 K Q (MPa m1/2)	23.52	24.162	22.554	24.762	24.582

Table 7

Validation requirements for 321038 A101 rail (samples 3) – the criterion 2.5(K _Q/R _p0.2).

Determining K Ic
Sample no.		3a	3b	3c	3d	3e
2.5 (K Q/R p0.2)2 < a 0	2.5 (K Q/R p0.2)2	0.0079	0.0084	0.0073	0.0088	0.0087
	a 0	19.81	19.83	19.91	20.11	19.93
2.5 (K Q/R p0.2)2 < B	2.5 (K Q/R p0.2)2	0.0079	0.0084	0.0073	0.0088	0.0087
	B	24.98	25.02	49.98	25.05	24.96
2.5 (K Q/R p0.2)2 < W − a	2.5 (K Q/R p0.2)2	0.0079	0.0084	0.0073	0.0088	0.0087
	W − a	20.13	20.20	20.17	19.93	20.03
K Ic results		39.2	40.27	37.59	41.27	40.97

The fatigue crack lengths (a _s), as well as the total crack length a, i.e. including the mechanical notch, were measured on the obtained sample fractures a _o (a = a _o + a _s),where a _o is the length of the mechanical notch, a _s is the fatigue crack length, and a is the total crack length.

All quantitative results of measurements of the depth of the fatigue zone, both at the edges and in the central zone, prove that one of the conditions for generating a fatigue crack is met [4]. The fracture images of the sample that gave the lowest K _Ic result of 37.59 MPa m^1/2 and the sample that gave the highest result of 41.27 MPa m^1/2 are shown in Figure 14. It can be seen that the fatigue fracture is ductile and the zone fatigue scrap is approximately 6 mm for both samples.

Figure 14

View of fractures for samples (a) 3c and (b) 3d.

The fracture toughness of the material in the plane deformation state K _Ic corresponds to the K _Q value if the conditions specified in Tables 8–10 are met. All criteria for meeting the K _Ic coefficient have been confirmed. For the tested sections of various rails, stress intensity factor values were obtained in the range of 39.86–42.97 MPa m^1/2. Table 11 presents a summary of the stress intensity factor test results. These are results similar to those presented in the literature [30,31], where the given ranges of stress intensity factor values determined on heat-treated rail samples are in the range of 37–42 MPa m^1/2. To calculate the K _Ic value, the values of the conventional yield strength R _p0.2 determined at a temperature of −20°C were adopted.

Table 8

Checking the validity criterion K _Ic.

Sample	Check the conditions
	a ≥ 2.5 K Q R p 0.2 2 \hspace{1em}{\left(\frac{{K}_{{\rm{Q}}}}{{R}_{{\rm{p}}0.2}}\right)}^{2} , [m]	B ≥ 2.5 K Q R p 0.2 2 \hspace{1em}{\left(\frac{{K}_{{\rm{Q}}}}{{R}_{{\rm{p}}0.2}}\right)}^{2} , [m]
	Steel – R350HT
3a	0.01981 > −0.0079	0.02498 > −0.0079
3b	0.01983 > −0.0084	0.02502 > −0.0084
3c	0.01991 > −0.0073	0.02498 > −0.0073
3d	0.02011 > −0.0088	0.02505 > −0.0088
3e	0.01993 > −0.0087	0.02496 > −0.0087

Table 9

Validity check K _Ic.

Sample	Check the conditions
	(W – a) > 2.5 K Q R p 0.2 2 {\left(\frac{{K}_{{\rm{Q}}}}{{R}_{{\rm{p}}0.2}}\right)}^{2} , [m]
	Steel – R350HT
3a	0.02013 > −0.0079
3b	0.02020 > −0.0084
3c	0.02017 > −0.0073
3d	0.01993 > −0.0088
3e	0.02003 > −0.0087

Table 10

Condition check P _max/P _Q ≤ 1.1.

Sample	P Q (N)	P max (N)	P max/P Q
3a	17,810	18,344	1.03
3b	18,390	18,574	1.01
3c	17,240	18,274	1.06
3d	18,560	18,746	1.01
3e	18,560	18,746	1.01

Table 11

Results of determining the stress intensity factor.

Sample no.	K Ic (MPa m1/2)	Average value K Ic (MPa m1/2)	Standard deviation K Ic (MPa m1/2)
3a	39.20	39.86	1.50
3b	40.27
3c	37.59
3d	41.27
3e	40.39
2a	39.95	40.12	1.09
2b	40.77
2c	40.38
2 d	38.35
2e	41.16
1a	46.24	42.97	2.70
1b	45.36
1c	40.32
1d	40.67
1e	42.28
Requirements according to EN13674-1 for the R350HT grade	Single min. value 30 MPa m1/2	Average value min 32 MPa m1/2

5.3

Fatigue strength

The results of the fatigue strength test are presented in Table 12. Figure 15 shows a view of the sample marked as number 2a after the fatigue test.

Table 12

Fatigue strength test results.

Sample no.	Amplitude	Force (kN)	Result
1a	0.00135	10.78	No cracks
1b	0.00135	10.88	No cracks
1c	0.00135	10.75	No cracks
2a	0.00135	10.83	No cracks
2b	0.00135	10.83	No cracks
2c	0.00135	10.88	No cracks
3a	0.00135	10.83	No cracks
3b	0.00135	10.76	No cracks
3c	0.00135	10.57	No cracks

Figure 15

Sample number 2a in the holders after fatigue testing.

The presented test results confirm the high properties of heat-treated rails in the R350HT steel grade, in terms of basic mechanical properties, hardness and fatigue tests. The tested rails were also characterized by an appropriate microstructure, which was fully pearlitic in the entire rail area without undesirable martensite or bainite precipitation. The average value of the fatigue crack growth rate from all nine tests for the parameter ΔK = 10 MPa m^1/2 was 14.98 m/Gc, while for the parameter ΔK = 13.5 MPa m^1/2, it was 27.95 m/Gc. In terms of the stress intensity factor K _Ic, the average result obtained from all 15 tests was 40.95 MPa m^1/2, which is 28% higher than the requirements of the standard [4] for this grade of steel. All results met the acceptance criterion for determining the K _Ic factor, none of the results was classified as K _Q. Each of the nine tests subjected to fatigue tests passed the 5 million cycles required by the standard [4] without cracking. Obtaining such good results in the entire spectrum of various tests is the result of obtaining the appropriate pearlite morphology, in particular its fragmentation, as a result of the controlled, accelerated cooling of the rail head in a water-polymer mixture with appropriately selected parameters of this process.

The scientific objective of the work was achieved by confirming the stability of the rail head heat treatment line in fatigue tests, residual stress tests and analysis of the microstructure and other properties, which confirm the high properties of steel in the R350HT grade. Based on the analysis of the obtained qualification test results, it can be stated that the heat treatment process is carried out correctly, ensuring the maintenance of key process parameters, i.e. processing time, temperature, properties of the water–polymer mixture, coolant flow conditions and rail foot cooling, within the reference ranges, which translates into meeting high and rigorous requirements regarding the properties of heat-treated rails and their microstructure.

The study was supported by the National Centre for Research and Development, Poland as part of the research and development project titled “Reliable and durable in operation, modern railway rails with a length of 120 m, characterized by high mechanical properties, high resistance to cracking and a modified material microstructure as a result of modernization of the cooling process after rolling.” – POIR.01.01.01-00-0438/18.

Author states no funding involved.

The author confirms the sole responsibility for the conception of the study, presented results and manuscript preparation.

Author states no conflict of interest.