Experimental data from service-like creep-fatigue experiments on grade P92 steel

This article refers to the research article entitled “Creep-Fatigue of P92 in Service-Like Tests with Combined Stress- and Strain-Controlled Dwell Times” [1]. It presents experimental mechanical data from complex service-like creep-fatigue experiments performed isothermally at 620°C and a low strain amplitude of 0.2 % on tempered martensite-ferritic grade P92 steel. The datasets in text file format provide cyclic deformation (minimum and maximum stresses) and the total (hysteresis) data of all recorded fatigue cycles for three different creep-fatigue experiments: 1) a standard relaxation fatigue (RF) test with symmetrical dwell times of three minutes introduced at minimum and maximum strain, 2) a fully strain-controlled service-like relaxation (SLR) test combining these three-minute peak strain dwells with a 30-minute dwell in between at zero strain, and 3) a partly stress-controlled service-like creep (SLC) test combining the three-minute peak strain dwells with 30-minute dwells at constant stress. Such service-like (SL) tests with additional long-term stress- and strain-controlled dwell times are non-standard, rare, and expensive, making these data very valuable. They may be used to approximate cyclic softening in the technically relevant range, for the design of complex SL experiments, or for detailed analyses of stress-strain hystereses (e.g., for stress or strain partitioning methods, for the determination of hysteresis energies (work), inelastic strain components, etc.). In addition, the latter analyses may supply important input for advanced parametric lifetime modeling of components under creep-fatigue loading or model calibration parameters.


a b s t r a c t
This article refers to the research article entitled "Creep-Fatigue of P92 in Service-Like Tests with Combined Stress-and Strain-Controlled Dwell Times " [1] . It presents experimental mechanical data from complex service-like creep-fatigue experiments performed isothermally at 620 °C and a low strain amplitude of 0.2 % on tempered martensite-ferritic grade P92 steel. The datasets in text file format provide cyclic deformation (minimum and maximum stresses) and the total (hysteresis) data of all recorded fatigue cycles for three different creep-fatigue experiments: 1) a standard relaxation fatigue (RF) test with symmetrical dwell times of three minutes introduced at minimum and maximum strain, 2) a fully strain-controlled service-like relaxation (SLR) test combining these three-minute peak strain dwells with a 30-minute dwell in between at zero strain, and 3) a partly stresscontrolled service-like creep (SLC) test combining the threeminute peak strain dwells with 30-minute dwells at constant stress. Such service-like (SL) tests with additional long-term stress-and strain-controlled dwell times are non-standard, rare, and expensive, making these data very valuable. They may be used to approximate cyclic softening in the technically relevant range, for the design of complex SL experiments, or for detailed analyses of stress-strain hystereses (e.g., for stress or strain partitioning methods, for the determination of hysteresis energies (work), inelastic strain components, etc.

Value of the Data
• The data in this article relates to complex (non-standard) creep-fatigue experiments with combined stress-and strain-controlled dwell times and extraordinarily long dwells that make these tests very expensive and valuable. • They provide insight into the fundamentally different mechanical behavior during longterm stress-and strain-controlled dwell times in creep-fatigue experiments and demonstrate their effect on subsequent transient load shifts and dwell times. • Because the high-temperature tests were performed at a low mechanical strain amplitude (in the technically relevant elastic-plastic deformation range), the data are of interest to both materials science and industry (e.g., the nuclear or thermal power plant industry). The current experiments were performed on one widely used steel grade for steam pipelines. It is anticipated that the demonstrated impact of the different load profiles on the deformation and lifetime is qualitatively similar for the whole material class, i.e., all 9-12% chromium tempered martensite-ferritic steels. • The data can be used to approximate cyclic softening in the technically relevant range, for the design of complex SL experiments, or for detailed analyses of stress-strain hystereses (e.g., for stress or strain partitioning methods, for the determination of hysteresis energies (work), inelastic strain components, etc.). In addition, the latter analyses can supply important input for advanced parametric lifetime modeling of components under creepfatigue loading or model calibration parameters.

Objective
The objective of this article is two-fold: First, it highlights the relevance of the data for different interest groups, which is based on the special, non-standard loading paths that were applied in the related mechanical experiments (cf. previous section for further details). Second, it provides additional details on the test setup, data structure and applied analysis methods (cf. following section). Thereby, it guides possible users of the datasets through the provided files and ensures proper analysis/interpretation of the available data. This exceeds/extends the scope of the original research article that was limited to a description and phenomenological interpretation of the observed material behavior.

Data Description
One summarizing * .lis file is provided for each of the three dwell-fatigue experiments, namely relaxation fatigue (RF), service-like relaxation fatigue (SLR), and service-like creepfatigue tests (SLC), all of which are indicated in the file names. The file format * .lis is SQR output file format (tab-delimited) that can be opened like ASCII * .txt or * .dat files by e.g., text editors, MS Excel, or common data analysis software. Each data file consists of three parts in a broadto-specific order, as described in detail in the following:

File name and meta-data of experimental conditions (headers)
This first file section summarizes general information on the testing parameters/conditions, machine, and specimen geometry. The number of cycles to failure ( N f ) and the total number of fatigue cycles ( N) are also indicated in this upper file part.

Max-min data
This section starts with the sub-heading/entry "[MAX-MIN]". The peak (maximum and minimum) values of stress, strain, and temperature are listed for all analyzed fatigue cycles ("Cycle No") in this section. They were extracted from the total data given in the third file section and are marked for σ min , σ max and ε min , ε max in the stress-strain hystereses of the first cycles in Fig. 1 exemplarily. The provided mechanical stresses ( σ ) were determined from mechanical forces ( F ) and the initial cross-section ( A 0 ) of the samples at room temperature via σ = F / A 0 .

Total data of individual fatigue cycles
This section starts with the sub-heading/entry "[Data]". Time-dependent stress-, strain-, and temperature data are provided in this final file section for a reduced number of complete fatigue cycles, as standard in fatigue data analysis.
The following cycles were analyzed: • all cycles for 1 < N ≤ 10, • every 10 th cycle for N > 10, • and, additionally, all cycles, in which variations of σ min , σ max , ε min , ε max of more than 1 % with respect to the preceding cycle were detected.
Data within individual cycles were recorded using combined acquisition rates. Data were recorded closely during the ramps in all experiments, i.e., using small time increments. The time increment was increased during the dwell times to keep the total data volume reasonably low. However, additional data were recorded mainly at the beginning of the dwell times at smaller time steps to be able to analyze nonlinear stress or strain changes with sufficient accuracy. In summary, data were collected at the following time increments (TI): RF: • during ramps: TI = 0.08 s • during dwell times: TI = 1.80 s • additionally, during the first six seconds of the compressive dwell (DT min ): TI = 0.08 s • additionally, during the last two seconds of the compressive dwell (DT min ): TI = 0.08 s • additionally, during the first four seconds of the tensile dwell (DT max ): TI = 0.08 s • additionally, during the last two seconds of the tensile dwell (DT max ): TI = 0.08 s SLR and SLC: • during ramps: TI = 0.08 s • during dwell times: TI = 5.98 s • additionally, during the first 10 seconds of the compressive dwell (DT min ), the SSO-dwell and the tensile dwell (DT max ): TI = 0.08 s • additionally, during the first 60 seconds of the compressive dwell (DT min ), the SSO-dwell and the tensile dwell (DT max ) TI = 2.00 s The data type "single" allowed recording the data at a time resolution of eight digits with variable decimal separator position.

Experimental Design, Materials and Methods
Experimental design: Three types of complex creep-fatigue experiments were conducted isothermally at 620 °C in air on grade P92 steel with a mechanical strain amplitude of 0.2 % and a strain ratio of R = -1. They led to the sophisticated, test-specific stress-strain relations shown in Fig. 1 : (1) Strain-controlled relaxation fatigue (RF) test with strain-controlled three-minute dwell times (DT) at minimum and maximum strain (referred to as DT min and DT max ) in each cycle [1 , 2] ; (2) service-like relaxation fatigue (SLR) with strain-controlled three-minute DT min and DT max and additional 30-minute dwells at constant strain ( ε = 0, which represent steady-state operation (therefore designated as SSO(SLR) periods) [1] , and (3) service-like creep-fatigue (SLC) with strain-controlled three-minute DT min and DT max and additional 30-minute dwells at constant stress [ σ ( ε = 0)] (designated as SSO(SLC) periods) [1] .
The designs of experiments are illustrated in Fig. 2 for the RF test and in Fig. 3 a,b for both SL tests, and specifications for all segments of the corresponding loading paths are listed in Tables 1-3 .  Loading path of fully strain-controlled relaxation fatigue (RF) test. Specific loading points (0 to 7) are numbered in analogy to those of the SL tests ( Fig. 3 ) for the sake of comparability, i.e., points 3 and 4 are missing in this scheme due to the absence of an SSO period for this type of test [2] . Specific information on the depicted loading segments of one standard RF cycle are summarized in Table 1 .

Table 2
Summary of SLR test design. Description of loading segments according to the numbering in Fig. 3   Material: The creep-fatigue tests were conducted for a grade P92 steel (X12CrMoWVNbN10-1) with a tempered martensite-ferritic microstructure and the chemical composition Fe-0.126C-0.114Si-0.446Mn-0.012P-0.005S-8.930Cr-0.167Ni-0.500Mo-0.019Co-0.092Nb-0.169V-1.95W (mass%). Cylindrical specimens of 18 mm gauge length and 6 mm gauge diameter were machined from a virgin pipe with a wall thickness of 47 mm in tangential direction. The specimens were tested with ground surface conditions (surface roughness, R z ≤ 1). For a technical drawing of the specimen geometry and for the material's microstructure, it is referred to [2] . The  to the slope of the stabilized deformation region (narrow parallel red lines). Plotted mean stresses and stress ranges can be determined from the "Max-min" stress data in the second sections of the respective * .lis files via ( σmax + σ min ) / 2 and σmax − σ min , respectively. original steam pipe was a fully heat treated section, provided by a commercial manufacturer (Vallourec, Germany). The well-controlled and reproducible industrial manufacturing process ensured representative material properties and high homogeneity of the material.
Methods: All tests were conducted in air, using a servo-hydraulic test machine with 100 kN force transducers (MTS Landmark, MTS, Eden Prairie, USA). Mechanical strains were measured by a water-cooled high-temperature extensometer (MTS-632.51F.04, MTS, Eden Prairie, USA) with an initial gauge length of 12 mm. All strain ramps (cf. Tables 1-3 ) were performed at a constant strain rate of 1.0 × 10 -3 1/s. Specimens were inductively heated to a target temperature of 620 °C, which was indicated and controlled by a type S thermocouple of 0.2 mm diameter, spot-welded to the center of the gauge length. A plug-in high stability temperature controller (Eurotherm 2604, Schneider Electric Systems, Limburg, Germany) was used to measure and control the temperature. After fatigue testing, the number of cycles to failure, N f , was determined by applying the failure criterion of 10% load drop of the maximum peak stress against the stabilized cyclic deformation region, as illustrated in Fig. 4 .

Ethics Statements
The present work did not involve human subjects or animal experiments. No sensitive personal data was used.

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.

Data Availability
Experimental Data from Service-Like Creep-Fatigue Experiments on Grade P92 Steel (Original data) (Zenodo).