A time-domain cycle counting method for turbine disk creep-fatigue loading

Abstract

This invention relates to the field of life assessment and load spectrum compilation technology for high-temperature components of aero-engines, and provides a time-domain cyclic counting method for creep-fatigue loads on turbine disks. The method includes: preprocessing the measured load-time history; simplifying the preprocessed load history using trapezoidal wave equivalence; dividing the load history based on the rainflow counting method; extracting the load-holding parameters for each cycle based on the division results; and outputting all parameters of the time-domain load sequence of creep-fatigue interaction effects. This invention enables the output cyclic sequence to accurately reflect the creep-fatigue interaction effects experienced by the turbine disk in actual service; it has clear logical definitions and repeatable operation procedures, significantly improving the processing efficiency and engineering applicability of complex load spectra for high-temperature components such as turbine disks.

Description

Technical Field

[0001] This invention relates to the field of life assessment and load spectrum compilation technology for high-temperature components of aero-engines, and particularly to a time-domain cyclic counting method for turbine disk creep-fatigue loads. Background Technology

[0002] With the continuous improvement of aero-engine performance, key hot-end components such as turbine disks are subjected to complex cyclic loading environments of extreme high temperature and high stress for extended periods. Their failure modes mainly stem from two mechanisms: fatigue and creep. Furthermore, these two mechanisms exhibit significant interaction and synergistic effects during actual service, significantly increasing the complexity of life prediction. Against this backdrop, accurately characterizing the measured load-time history using the cyclic counting method has become a crucial prerequisite for constructing life prediction models, reliability design, and experimental verification.

[0003] Cyclic counting methods aim to transform a continuous random load spectrum into a series of discrete load cycle events that can be used for quantitative damage analysis. Currently, rainflow counting is widely used as a standard method in the engineering field. To meet the needs of creep-fatigue life analysis of high-temperature components, various improved methods have been developed by the industry to further introduce time-related parameters.

[0004] However, while the commonly used rainflow counting method and its improved versions are widely used in engineering, their core algorithms fundamentally contradict the needs of creep-fatigue interaction analysis for high-temperature components. Traditional rainflow counting methods are primarily designed for fatigue-driven loads, often compressing or ignoring the load holding phase during the counting process, making it difficult to effectively characterize time-related creep damage mechanisms. Subsequent improved methods, developed to incorporate time parameters, generally employ a two-step architecture: "first identifying fatigue cycles, then adding the holding time." This is essentially a back-insertion of the holding segment within the fatigue analysis framework, failing to extract the complete time-domain waveform consisting of "rise-hold-fall" as an integrated whole. This fragmented approach not only disrupts the time-domain interaction between fatigue and creep events in the original load but also makes it difficult to faithfully reproduce the true location, duration, and morphological characteristics of the holding segment. Ultimately, the output discrete cyclic event set has inherent defects in information completeness and time-series fidelity, failing to provide a reliable data foundation for high-precision creep-fatigue life prediction.

[0005] In summary, existing cyclic counting methods still have limitations in preserving complete time-domain load information. Therefore, this invention proposes a time-domain cyclic counting method for turbine disk creep-fatigue loads to better support life analysis and prediction of high-temperature components such as turbine disks under creep-fatigue effects. Summary of the Invention

[0006] The purpose of this invention is to provide a time-domain cycle counting method for turbine disk creep-fatigue loads, which can directly retain the complete time-domain waveform and extract the load-preserving segment as an integral part of the cycle.

[0007] This invention is implemented as follows: a time-domain cyclic counting method for turbine disk creep-fatigue loads, the method comprising:

[0008] Measured load-time history preprocessing;

[0009] The preprocessed load history is simplified by trapezoidal wave equivalence: the load history is simplified according to the load change trend over time, and the continuously changing load path is equivalently represented as a trapezoidal wave form composed of a combination of ramp loading section and stationary load holding section. Thus, while keeping the physical meaning of the original load path unchanged, the random load time history is transformed into a structurally regular trapezoidal load history with clear parameters.

[0010] Load history is determined based on rainflow counting method;

[0011] Based on the partitioning results, extract the load-preserving parameters for each cycle sequentially;

[0012] Output all parameters of the time-domain load sequence of creep-fatigue interaction effect.

[0013] Preferably, the measured load-time history preprocessing specifically includes:

[0014] Outlier data removal: Verify the monotonicity of the time series, remove outlier timestamps, calculate the load change rate of adjacent sampling points, and identify and remove load abrupt changes based on the load change rate threshold.

[0015] High-frequency noise filtering: To further reduce the number of inflection points in subsequent processing, the load history is filtered. An amplitude threshold is set, and adjacent peaks and valleys with amplitude differences less than this threshold are merged, thereby simplifying the history while preserving the main load fluctuation pattern; the amplitude threshold... Calculate using the following formula:

[0016] ;

[0017] in, and These represent the maximum and minimum values ​​of the pre-filter load history, respectively. These are empirical constants;

[0018] Small load cycle deletion: To reduce data volume and focus on the main cycles that significantly contribute to damage, a round of small load cycle pre-deletion based on engineering experience is performed on the filtered load history before performing fine creep-fatigue cycle counting; a simplified threshold determination method is used, where the load range is less than a threshold. The tiny fluctuations in the cycle are removed during this stage, and their threshold is... Calculate using the following formula:

[0019] ;

[0020] in, and These are the maximum and minimum load amplitudes for the current load block or the entire load history, respectively. This is an empirical coefficient.

[0021] Preferably, the division of load history based on rainflow counting specifically includes:

[0022] Store complete original data segments: Scan the trapezoidal load history point by point, and store the complete load-time original data segments to maintain the continuity and integrity of each trapezoidal load unit in time sequence, providing basic data support for subsequent inflection point identification and cyclic division;

[0023] Identify all local peak-valley sequences: Based on stored load-time history data, detect nodes where the load change direction reverses, i.e., identify all local maxima and minima, and store them as peak sequences. Valley value sequence Finally, the load inflection points are recorded in chronological order as an ordered sequence. ;

[0024] Matching similar inflection points: Based on the temporal structure of the load history, and in accordance with the closure principle of the rainflow method, similar inflection points are matched to obtain matching results. That is, peak values ​​are associated with peak values ​​and valley values ​​with valley values ​​according to the temporal sequence and the cycle closure principle to form potential cycle boundary pairs, thereby determining the start and end range of each independent cycle.

[0025] Dividing the Cycle Intervals: Based on the matching results above, the original trapezoidal wave load history is divided into continuous intervals on the time axis, with each interval corresponding to a preliminarily defined load cycle. The It is a complete, uncut trapezoidal wave load segment that runs from the start point to the end point.

[0026] Preferably, the step of sequentially extracting each cyclic load-preserving parameter based on the partitioning result specifically includes:

[0027] Identify multiple load segments within each cycle: Sequentially identify and determine each defined load cycle. During all load change phases; based on the physical definition of the load-maintaining segment, it is determined whether the derivative of the load with respect to time is greater than a set minimum rate threshold. At the same time, whether the duration of this state is greater than the minimum hold-up time threshold. Identify all load-preserving segments that meet the criteria. ,in For the first The first loop within the loop Each carrier sub-segment, Number the current cycle;

[0028] Record load values, load durations, and locations: Identify each load segment set. Record key parameters to fully characterize their state in the time domain and load domain, and provide a data foundation for subsequent high-fidelity reconstruction;

[0029] Identify and mark the creep-preserving segments: perform creep correlation judgment and marking on the creep-preserving segments, and set the creep-preserving sub-segments. Calculate the coefficient of variation and the maximum coefficient of variation separately;

[0030] The formulas for calculating the coefficient of variation and the maximum coefficient of variation are as follows:

[0031] ;

[0032] ;

[0033] in, The coefficient of variation; This represents the standard deviation of the load within the protection section; Its average value; The maximum coefficient of variation is the standard. and These are the maximum and minimum values ​​of the load data segment, respectively.

[0034] If a certain load-preserving sub-segment The corresponding original load data simultaneously satisfy Less than the first threshold If the load is less than the second threshold, the load segment is determined to be highly stable and is marked as "true" for creep load protection segment; otherwise, it is marked as "false" for non-creep load protection segment. The first and second thresholds are set based on engineering experience and material creep characteristics.

[0035] Preferably, the key parameters include:

[0036] Load Capacity The stable load level within this load-bearing section;

[0037] Load hold start and end time: Record the load hold segment during the current load cycle. Start time within and end time ;

[0038] Global time location: based on load cycle The offset on the global timeline determines the absolute time position of this payload-carrying segment within the complete flight mission profile.

[0039] Duration of service The calculation formula is: .

[0040] Preferably, all parameters of the output creep-fatigue interaction effect time-domain load sequence specifically include:

[0041] Extracting the load mean and load amplitude: for each load cycle Based on its boundary inflection point, i.e., the maximum load within that cycle. and minimum value Calculate the characteristic parameters of the load cycle, namely the load amplitude and the load mean.

[0042] Determine the loading direction: based on the load cycle The order of load changes determines the overall loading direction. If the load changes in the order of "first increasing and then decreasing", it is defined as a negative loading cycle, denoted as . If the load changes in the order of "first decreasing and then increasing", it is defined as a positive loading cycle, denoted as . This directional parameter is essential information for accurately reconstructing the load sequence, the stress-strain hysteresis response of the associated material, and distinguishing the tensile-compressive asymmetric damage mechanism.

[0043] Calculate the loading rate for each rising and falling segment: Calculate the load cycle. The geometric features of all non-horizontal segments are used to accurately reconstruct the load waveform;

[0044] The slopes of all the diagonal segments, arranged in chronological order, form the slope sequence of this cycle. , The total number of diagonal segments represents the complete shape of all linearly changing segments in the cyclic waveform.

[0045] Mark the load type: to specify each load cycle The underlying dominant damage mechanism, based on the identification and labeling of the creep properties of the load-holding segments at the peak and trough points within the cycle, determines the creep-fatigue waveform type of the cycle. A comprehensive assessment is conducted to provide a direct basis for mechanism classification in order to select the appropriate creep-fatigue interaction damage model.

[0046] Reconstructing the true load sequence: Cycle each load cycle based on the parameters output above. Generate a complete structured time-domain cyclic description vector The vector is constructed as follows:

[0047] ;

[0048] in, and Load cycles The maximum and minimum load values; and respectively according to and The calculated average load and load amplitude; For load cycles The overall loading direction; The load cycles were recorded in chronological order for this sequence. Calculate the slope values ​​of all oblique line segments within the range; To preserve the ordered parameter set for the load segment, load cycles were recorded in chronological order. Complete parameters for all load segments within the package.

[0049] Preferably, the calculation of the characteristic parameters of the load cycle, namely the load amplitude and the load mean, specifically includes:

[0050] Load amplitude The range of stress fluctuations characteristic of the cycle is calculated using the following formula:

[0051] ;

[0052] Average load The average stress level of the cycle is characterized by the following formula:

[0053] .

[0054] Preferably, the calculation of the loading rate of each rising segment and falling segment specifically includes:

[0055] For the first chronological order Calculate the slope of a diagonal line segment. The slope defines the rate at which the load changes over time in that segment:

[0056] ;

[0057] in, and These are the starting and ending load values ​​of the i-th oblique line segment, respectively; and The corresponding start and end times are respectively.

[0058] Preferably, the creep attribute marking results of the load-holding segments at the peak and trough values ​​within the cycle, based on the identified and marked load-holding segments where creep occurs, determine the creep-fatigue waveform type of the cycle. The comprehensive judgment specifically includes:

[0059] Based on the combination of creep / non-creep properties of the load-bearing sections at peak and trough values, they are divided into the following two categories:

[0060] PC type: There is no creep holding section at the peak of the cycle, but there is a creep holding section at the valley, which represents the combination of tensile plastic strain and reverse compressive creep strain.

[0061] CP type: There is a creep-holding segment at the peak of the cycle, but no creep-holding segment at the valley, which represents the combination of tensile creep strain and reverse compressive plastic strain.

[0062] Compared with the prior art, the beneficial effects of the present invention are:

[0063] This invention proposes an integrated time-domain waveform extraction method, which identifies and records the complete load waveform containing rising, holding, and falling phases as the basic cyclic unit. Based on the time sequence and morphological characteristics of the original load, the coupling relationship between creep holding and fatigue fluctuation is fully preserved, so that the output cyclic sequence can truly reflect the creep-fatigue interaction effect experienced by the turbine disk in actual service.

[0064] The output of this invention is a structured time-domain load sequence, which embeds multi-dimensional parameters such as amplitude, mean, load holding stress level, load holding duration and its temporal position in the cycle. It has clear logical definitions and repeatable operation procedures, which significantly improves the processing efficiency and engineering applicability of complex load spectra of high-temperature components such as turbine disks, and provides efficient and reliable technical support for creep-fatigue design and life assessment of key components of aero-engines. Attached Figure Description

[0065] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0066] Figure 1 This is a flowchart of the method of the present invention; Figure 2 This is a schematic diagram of the load-time history after preprocessing according to the present invention; Figure 3 This is a simplified load-time history diagram of the trapezoidal wave equivalent of the present invention; Figure 4 This is a schematic diagram of the cyclic load division of the present invention; Figure 5This is a schematic diagram of the 43rd cyclic reconstruction in the embodiment of the present invention. Detailed Implementation

[0067] To better understand the technical content of this invention, the technical solutions of this invention are further described and explained below with reference to specific embodiments, but are not limited thereto. The technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0068] refer to Figures 1 to 2 The time-domain cyclic counting method for turbine disk creep-fatigue loads includes:

[0069] S1: Measured load-time history preprocessing.

[0070] The measured load-time history preprocessing specifically includes:

[0071] Abnormal data removal: Verify the monotonicity of the time series, remove outlier timestamps, calculate the load change rate of adjacent sampling points, and identify and remove load abrupt change points based on the load change rate threshold.

[0072] High-frequency noise filtering: To further reduce the number of inflection points in subsequent processing, the load history can be filtered. This filtering involves setting an amplitude threshold and merging adjacent peaks and valleys with amplitude differences less than the threshold, thereby simplifying the history while preserving the main load fluctuation patterns. The amplitude threshold... Calculate using the following formula:

[0073] ;

[0074] in, and These represent the maximum and minimum values ​​of the pre-filter load history, respectively. This is an empirical constant; for aero-engine loads, a value range of 8.0 to 12.5 is recommended.

[0075] Small load cycle deletion: To reduce data volume and focus on the main cycles that significantly contribute to damage, a round of small load cycle pre-deletion based on engineering experience can be performed on the filtered load history before performing fine creep-fatigue cycle counting. A simplified threshold determination method is used, where the load range (peak-to-valley difference) is less than a threshold. The tiny fluctuations in the cycle are removed during this stage, and their threshold is... Calculate using the following formula:

[0076] ;

[0077] in, and These are the maximum and minimum load amplitudes for the current load block or the entire load history, respectively. This is an empirical coefficient that can be determined based on material limits, load spectrum type, and engineering experience. For turbine disk loads in aero-engines, the recommended value range is usually 0.02 to 0.10.

[0078] S2: Simplify the preprocessed load history using trapezoidal wave equivalent.

[0079] The load history is simplified based on the load variation trend over time. The continuously changing load path is equivalently represented as a trapezoidal wave composed of a combination of a ramp loading section and a stationary load-preserving section. Thus, while keeping the physical meaning of the original load path unchanged, the random load time history is transformed into a structurally regular trapezoidal load history with well-defined parameters.

[0080] S3: Load history is divided based on rainflow counting method.

[0081] S31: Store the complete original data segment:

[0082] The trapezoidal load history obtained from S2 is scanned point by point, and the original load-time data segments are completely stored to maintain the continuity and integrity of each trapezoidal load unit in time sequence, providing basic data support for subsequent inflection point identification and cyclic division.

[0083] S32: Identify all local peak-valley sequences:

[0084] Based on the saved load-time history data, nodes where the load change direction reverses are detected, i.e., all local maxima (peaks) and local minima (valleys) are identified and stored as a peak sequence. Valley value sequence Finally, the load inflection points are recorded in chronological order as an ordered sequence. .

[0085] S33: Matching inflection points of the same type:

[0086] Based on the temporal structure of the load history, and following the closure principle of the rainflow method, similar inflection points are matched to obtain matching results. Specifically, peak values ​​are correlated with peak values ​​and valley values ​​with valley values ​​according to the temporal sequence and the cyclic closure principle, forming potential cyclic boundary pairs (e.g., Define a cyclic framework from valley to valley, thereby determining the start and end ranges of each independent cycle.

[0087] S34: Divide the cyclic interval:

[0088] Based on the matching results of S33, the original trapezoidal wave load history is divided into continuous intervals on the time axis, with each interval corresponding to a preliminarily defined load cycle. Load cycle here It is a complete, uncut trapezoidal wave load segment that runs from the start point to the end point, and may contain multiple stages such as rising, holding, and falling.

[0089] S4: Extract the load parameters for each cycle sequentially based on the partitioning results.

[0090] S41: Identify multiple load holds within each cycle:

[0091] Sequentially identify and determine each divided load cycle During all load variation phases, based on the physical definition of the load-maintaining phase (load value remains essentially constant), the system determines whether the derivative of the load with respect to time (i.e., the slope) is greater than a set minimum rate threshold. At the same time, whether the duration of this state is greater than the minimum hold-up time threshold. Identify all load-preserving segments that meet the criteria. ,in For the first The first loop within the loop Each carrier sub-segment, Number the cycle it belongs to.

[0092] S42: Record load holding value, load holding time, and location:

[0093] For each load-preserving segment set identified by S41 Record the following key parameters to fully characterize its state in the time domain and load domain, and provide a data foundation for subsequent high-fidelity reconstruction;

[0094] Load Capacity The stable load level within this load-bearing section;

[0095] Load hold start and end time: Record the load hold segment during the current load cycle. Start time within and end time ;

[0096] Global time location: based on load cycle The offset on the global timeline determines the absolute time position of this payload-carrying segment within the complete flight mission profile.

[0097] Duration of service The calculation formula is: .

[0098] S43: Identify and mark the creep-preserving sections:

[0099] Not all load-holding segments lead to significant creep damage. To focus on key damage mechanisms, creep correlation was assessed and marked for the load-holding segments recorded in S42. The assessment criteria referenced the load characteristics of aero-engines, using statistical indicators of load fluctuation stability as the criterion. For the set of load-holding sub-segments... Calculate the coefficient of variation separately. Measured by the maximum coefficient of variation .

[0100] coefficient of variation The calculation formula is as follows:

[0101] ;

[0102] in, This represents the standard deviation of the load within the protection section; Its average value;

[0103] Maximum coefficient of variation measurement standard The calculation formula is:

[0104] ;

[0105] in, and These are the maximum and minimum values ​​of the load data for that segment, respectively.

[0106] Based on engineering experience and material creep characteristics, a judgment threshold can be set. For turbine disks, this threshold can be set to... and If a certain carrier segment The corresponding original load data simultaneously satisfy Less than the first threshold If the load is less than the second threshold, the load segment is determined to be highly stable and is marked as "true" for creep load protection segment; otherwise, it is marked as "false" for non-creep load protection segment. The first and second thresholds are set based on engineering experience and material creep characteristics.

[0107] S5: Output all parameters of the time-domain load sequence of creep-fatigue interaction effect.

[0108] S51: Extract the load mean and load amplitude:

[0109] For each load cycle Based on its boundary inflection point, i.e., the maximum load within that cycle. and minimum value Calculate the characteristic parameters of the load cycle, namely the load amplitude and the load mean.

[0110] Load amplitude The range of stress fluctuations characteristic of the cycle is calculated using the following formula:

[0111] ;

[0112] Average load The average stress level of the cycle is characterized by the following formula:

[0113] .

[0114] S52: Determine loading direction:

[0115] To accurately reconstruct the temporal characteristics of the load history, based on the load cycle The order of load changes determines the overall loading direction. If the load changes in the order of "first increasing and then decreasing", it is defined as a negative loading cycle, denoted as . If the load changes in the order of "first decreasing and then increasing", it is defined as a positive loading cycle, denoted as . This directional parameter is essential information for accurately reconstructing the load sequence, relating the material stress-strain hysteresis response, and distinguishing between tension-compression asymmetric damage mechanisms.

[0116] S53: Calculate the loading rate of each rising and falling segment:

[0117] Accurate reconstruction of the load waveform requires calculation of the load cycle. Geometric features of all non-horizontal segments. For the first segment arranged in chronological order... Calculate the slope of a diagonal line segment (e.g., a line segment connecting the valley value to the start point of the first load-preserving segment, the start and end points of adjacent load-preserving segments, or the end point of the last load-preserving segment to the peak value). The slope defines the rate at which the load changes over time in that segment:

[0118] ;

[0119] in, and These are the starting and ending load values ​​of the i-th oblique line segment, respectively; and The corresponding start and end times are respectively.

[0120] The slopes of all the diagonal segments, arranged in chronological order, form the slope sequence of this cycle. , The total number of diagonal segments represents the complete shape of all linearly changing segments in the cyclic waveform.

[0121] S54: Marking the load type:

[0122] To clarify each load cycle The underlying dominant damage mechanism, based on the creep property labeling results of the load-holding segments (if present) at the peak and trough points within the cycle in S43, determines the creep-fatigue waveform type of the cycle. A comprehensive judgment is made. Primarily based on the combination of creep and non-creep properties in the load-bearing sections at peak and trough values, they are divided into the following two categories:

[0123] PC type: There is no creep holding section at the peak of the cycle, but there is a creep holding section at the valley, which represents the combination of tensile plastic strain and reverse compressive creep strain.

[0124] CP type: There is a creep-holding segment at the peak of the cycle, but no creep-holding segment at the valley, which represents the combination of tensile creep strain and reverse compressive plastic strain.

[0125] Waveform type The determination of this can provide a direct basis for the mechanism classification of the subsequent selection of the appropriate creep-fatigue interaction damage model.

[0126] S55: Restore the true load sequence:

[0127] Based on the parameters of all the above outputs, each load cycle... Generate a complete structured time-domain cyclic description vector This vector is an integration of all feature information for the cycle, containing all the geometric and physical parameters required for high-fidelity reconstruction of the original trapezoidal wave load sequence, while also providing key feature labels for damage mechanism analysis and lifetime prediction. Its structure is as follows:

[0128] ;

[0129] in, and Load cycles The maximum and minimum load values; and respectively according to and The calculated average load and load amplitude; For load cycles The overall loading direction ("+" or "-"); The load cycles were recorded in chronological order for this sequence. Calculate the slope values ​​of all oblique line segments within the range; To preserve the ordered parameter set for the load segment, load cycles were recorded in chronological order. Complete parameters of all load sub-segments within the first sub-segment, for the first sub-segment Each load segment is recorded as follows: This includes, in turn, the start and end times, the load value, the load duration, and the creep attribute markers (creep / non-creep).

[0130] The following are specific examples:

[0131] This embodiment presents a time-domain cyclic counting method for turbine disk loads oriented towards creep-fatigue interaction effects. Its core lies in fully preserving the time-domain information of the load waveform and achieving integrated extraction of the load-holding segments. The method mainly consists of the following three parts: trapezoidal waveform simplification and preprocessing of the load history, load-holding cycle division and identification based on the complete waveform, and quantization and structured output of multi-dimensional characteristic parameters of the cycles.

[0132] Step 1) Preprocessing of measured load-time history;

[0133] Using measured data from a complete flight mission profile of the high-pressure turbine disk of a certain type of aero-engine as the raw load spectrum, firstly, timestamp anomalies and load abrupt changes were removed. Then, low-pass filtering was performed to suppress high-frequency noise, and small-amplitude fluctuation cycles with minimal damage contribution were removed based on the thresholds described above. After the above preprocessing, a continuous and smooth load-time history was obtained, as shown below. Figure 2 As shown.

[0134] Step 2) Simplify the preprocessed load history using trapezoidal wave equivalent;

[0135] right Figure 2 The preprocessed load history shown is then simplified into a trapezoidal wave. The continuous rise and fall of the load are fitted with straight line segments, and the segments that meet the load-maintaining conditions are simplified into horizontal segments. Finally, the original load history is equivalent to a piecewise linear trapezoidal wave sequence composed of successively connected oblique and horizontal segments, as shown in the simplified result. Figure 3 As shown.

[0136] Step 3) Divide the load history based on the rainflow counting method;

[0137] according to Figure 3 The trapezoidal wave sequence shown identifies all local peaks and valleys to construct a sequence of inflection points. Based on the temporal and amplitude relationships, similar inflection points are matched to determine the boundaries of each cycle. According to the matching results, the trapezoidal wave load history is divided into a series of continuous load cycle intervals on the time axis. A schematic diagram of the division process is shown below. Figure 4 As shown in the figure (this figure is a partial enlarged view of the load history in this embodiment, exemplarily demonstrating the removal of abnormal loads, simplification of trapezoidal waves, and the division results of 5 consecutive cycles). Unlike the classic rainflow method, which removes the load-preserving segment during counting, this method always retains the complete original trapezoidal wave data segment corresponding to each cycle, ensuring that the load waveform is not destroyed. In this embodiment, the entire flight profile is ultimately divided into 75 cycles, saving cycle information and completely recording the start and end times of each cycle. Examples of some division results are shown in the table below:

[0138] Table 1. Load Cyclic Division Results in the Examples

[0139]

[0140] Step 4) Extract the load-preserving parameters for each cycle based on the partitioning results;

[0141] Based on the load cycles obtained in step 3, all load-holding segments are identified from the complete waveform data of each cycle. In this embodiment, a total of 259 load-holding segments are identified across all 75 cycles. Subsequently, for each load-holding segment, its load value, start and end times within the cycle, absolute time position in the flight profile, and load-holding duration are recorded. Finally, creep properties are determined based on the above criteria, and load-holding segments that meet the conditions are marked as creep load-holding segments ("true"), while the rest are marked as non-creep load-holding segments ("false"). Examples of parameter recordings and creep marking results for some load-holding segments are shown in the table below:

[0142] Table 2. Results of load-carrying parameter extraction in the examples

[0143]

[0144] Step 5) Output all parameters of the creep-fatigue interaction effect time-domain load sequence.

[0145] For the 75 cycles defined in this embodiment, firstly, based on each cycle... Maximum load and minimum value The load amplitude is calculated according to equations (5) and (6). with load mean Subsequently, the overall loading direction was determined based on the sequence of load changes within the cycle. In this embodiment, 38 positive loading cycles ("+") and 37 negative loading cycles ("-") were determined. To accurately reconstruct the waveform, the slope of the oblique line segment of all connecting feature points (such as valleys, start and end points of load-preserving segments, and peak values) within each cycle was calculated according to equation (7), forming the slope sequence of that cycle. Determine the creep-fatigue dominant type for each cycle. In this embodiment, the 75 loops are classified into: 38 PC-type loops and 37 CP-type loops. Examples of key parameters for the description vectors of some loops are shown in the table below:

[0146] Table 3. Results of Loop Parameter Extraction in the Examples

[0147]

[0148] Finally, for ease of representation, all the above calculation and judgment results are integrated, and each cycle is calculated according to equation (8). Generate a structured temporal cyclic description vector This embodiment generates a total of 75 vectors. Taking the description vector of one of the loops as an example, its specific structure is as follows:

[0149] ;

[0150] in, ;

[0151] ;

[0152] The loop can be reconstructed using the vectors described above; that is, the loop consists of 5 segments:

[0153] (a) P0-P13: Hold the load at a value of 749.12 for 13 seconds, during the 853s-866s section;

[0154] (b) P13-P36: Reach the next load-bearing section with a slope of -3.23;

[0155] (c) P36-P58: The load is maintained at 674.80 for 22 seconds, and the valley value is between 889s and 911s in the profile.

[0156] (d) P58-P85: Reach the next load with a slope of 4.40;

[0157] (e) P85-P111: Hold the load at a value of 749.67 for 26 seconds, in the section from 938s to 964s;

[0158] Where Pj represents the time point within the loop, and the reconstruction diagram is as follows: Figure 5 As shown.

[0159] It should be noted that the above embodiments are only some embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the technical scope disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention. Parts not covered in the present invention are the same as or can be implemented using existing technology.

Claims

1. A time-domain cyclic counting method for turbine disk creep-fatigue loads, characterized in that, The method includes: Measured load-time history preprocessing; The preprocessed load history is simplified by trapezoidal wave equivalence: the load history is simplified according to the load change trend over time, and the continuously changing load path is equivalently represented as a trapezoidal wave form composed of a combination of ramp loading section and stationary load holding section. Thus, while keeping the physical meaning of the original load path unchanged, the random load time history is transformed into a structurally regular trapezoidal load history with clear parameters. Load history is determined based on rainflow counting method; Based on the partitioning results, extract the load-preserving parameters for each cycle sequentially; Output all parameters of the time-domain load sequence of creep-fatigue interaction effect.

2. The time-domain cyclic counting method for turbine disk creep-fatigue load as described in claim 1, characterized in that, The measured load-time history preprocessing specifically includes: Outlier data removal: Verify the monotonicity of the time series, remove outlier timestamps, calculate the load change rate of adjacent sampling points, and identify and remove load abrupt changes based on the load change rate threshold. High-frequency noise filtering: To further reduce the number of inflection points in subsequent processing, the load history is filtered. An amplitude threshold is set, and adjacent peaks and valleys with amplitude differences less than this threshold are merged, thereby simplifying the history while preserving the main load fluctuation pattern; the amplitude threshold... Calculate using the following formula: ； in, and These represent the maximum and minimum values ​​of the pre-filter load history, respectively. These are empirical constants; Small load cycle deletion: To reduce data volume and focus on the main cycles that significantly contribute to damage, a round of small load cycle pre-deletion based on engineering experience is performed on the filtered load history before performing fine creep-fatigue cycle counting; a simplified threshold determination method is used, where the load range is less than a threshold. The tiny fluctuations in the cycle are removed during this stage, and their threshold is... Calculate using the following formula: ； in, and These are the maximum and minimum load amplitudes for the current load block or the entire load history, respectively. This is an empirical coefficient.

3. The time-domain cyclic counting method for turbine disk creep-fatigue load as described in claim 1, characterized in that, The specific steps of dividing the load history based on the rainflow counting method include: Store complete original data segments: Scan the trapezoidal load history point by point, and store the complete load-time original data segments to maintain the continuity and integrity of each trapezoidal load unit in time sequence, providing basic data support for subsequent inflection point identification and cyclic division; Identify all local peak-valley sequences: Based on stored load-time history data, detect nodes where the load change direction reverses, i.e., identify all local maxima and minima, and store them as peak sequences. Valley value sequence Finally, the load inflection points are recorded in chronological order as an ordered sequence. ; Matching similar inflection points: Based on the temporal structure of the load history, and in accordance with the closure principle of the rainflow method, similar inflection points are matched to obtain matching results. That is, peak values ​​are associated with peak values ​​and valley values ​​with valley values ​​according to the temporal sequence and the cycle closure principle to form potential cycle boundary pairs, thereby determining the start and end range of each independent cycle. Dividing the Cycle Intervals: Based on the matching results above, the original trapezoidal wave load history is divided into continuous intervals on the time axis, with each interval corresponding to a preliminarily defined load cycle. The It is a complete, uncut trapezoidal wave load segment that runs from the start point to the end point.

4. The time-domain cyclic counting method for turbine disk creep-fatigue load as described in claim 1, characterized in that, The step of extracting each cycle load preservation parameter sequentially based on the partitioning results specifically includes: Identify multiple load segments within each cycle: Sequentially identify and determine each defined load cycle. During all load change phases; based on the physical definition of the load-maintaining segment, it is determined whether the derivative of the load with respect to time is greater than a set minimum rate threshold. At the same time, whether the duration of this state is greater than the minimum hold-up time threshold. Identify all load-preserving segments that meet the criteria. ,in For the first The first loop within the loop Each carrier sub-segment, Number the current cycle; Record load values, load durations, and locations: Identify each load segment set. Record key parameters to fully characterize their state in the time domain and load domain, and provide a data foundation for subsequent high-fidelity reconstruction; Identify and mark the creep-preserving segments: perform creep correlation judgment and marking on the creep-preserving segments, and set the creep-preserving sub-segments. Calculate the coefficient of variation and the maximum coefficient of variation separately; The formulas for calculating the coefficient of variation and the maximum coefficient of variation are as follows: ； ； in, The coefficient of variation; This represents the standard deviation of the load within the protection section; Its average value; The maximum coefficient of variation is the standard. and These are the maximum and minimum values ​​of the load data segment, respectively. If a certain load-preserving sub-segment The corresponding original load data simultaneously satisfy Less than the first threshold If the load is less than the second threshold, the load segment is determined to be highly stable and is marked as "true" for creep load protection segment; otherwise, it is marked as "false" for non-creep load protection segment. The first and second thresholds are set based on engineering experience and material creep characteristics.

5. The time-domain cyclic counting method for turbine disk creep-fatigue load as described in claim 4, characterized in that, The key parameters include: Load Capacity The stable load level within this load-bearing section; Load hold start and end time: Record the load hold segment during the current load cycle. Start time within and end time ; Global time location: based on load cycle The offset on the global timeline determines the absolute time position of this payload-carrying segment within the complete flight mission profile. Duration of service The calculation formula is: .

6. The time-domain cyclic counting method for turbine disk creep-fatigue load as described in claim 5, characterized in that, The complete parameters of the output creep-fatigue interaction effect time-domain load sequence specifically include: Extracting the load mean and load amplitude: for each load cycle Based on its boundary inflection point, i.e., the maximum load within that cycle. and minimum value Calculate the characteristic parameters of the load cycle, namely the load amplitude and the load mean. Determine the loading direction: based on the load cycle The order of load changes determines the overall loading direction. If the load changes in the order of "first increasing and then decreasing", it is defined as a negative loading cycle, denoted as . If the load change sequence is "first decrease then increase", it is defined as a positive loading cycle, denoted as . This directional parameter is essential information for accurately reconstructing the load sequence, the stress-strain hysteresis response of the associated material, and distinguishing the tensile-compressive asymmetric damage mechanism. Calculate the loading rate for each rising and falling segment: Calculate the load cycle. The geometric features of all non-horizontal segments are used to accurately reconstruct the load waveform; The slopes of all the diagonal segments, arranged in chronological order, form the slope sequence of this cycle. , The total number of diagonal segments represents the complete shape of all linearly changing segments in the cyclic waveform. Mark the load type: to specify each load cycle The underlying dominant damage mechanism, based on the identification and labeling of the creep properties of the load-holding segments at the peak and trough points within the cycle, determines the creep-fatigue waveform type of the cycle. A comprehensive assessment is conducted to provide a direct basis for mechanism classification in order to select the appropriate creep-fatigue interaction damage model. Reconstructing the true load sequence: Cycle each load cycle based on the parameters output above. Generate a complete structured time-domain cyclic description vector The vector is constructed as follows: ； in, and Load cycles The maximum and minimum load values; and respectively according to and The calculated average load and load amplitude; For load cycles The overall loading direction; The load cycles were recorded in chronological order for this sequence. Calculate the slope values ​​of all oblique line segments within the range; To preserve the ordered parameter set for the load segment, load cycles were recorded in chronological order. Complete parameters for all load segments within the package.

7. The time-domain cyclic counting method for turbine disk creep-fatigue load as described in claim 6, characterized in that, The calculation of the characteristic parameters of the load cycle, namely the load amplitude and the load mean, specifically includes: Load amplitude The range of stress fluctuations characteristic of the cycle is calculated using the following formula: ； Average load The average stress level of the cycle is characterized by the following formula: 。 8. The time-domain cyclic counting method for turbine disk creep-fatigue load as described in claim 6, characterized in that, The calculation of the loading rate for each rising and falling segment specifically includes: For the first chronological order Calculate the slope of a diagonal line segment. The slope defines the rate at which the load changes over time in that segment: ； in, and These are the starting and ending load values ​​of the i-th oblique line segment, respectively; and The corresponding start and end times are respectively.

9. The time-domain cyclic counting method for turbine disk creep-fatigue load as described in claim 8, characterized in that, The creep attribute marking results of the holding load segments at the peak and trough values ​​within the cycle, based on the identified and marked holding load segments where creep occurred, are used to classify the creep-fatigue waveform type of the cycle. The comprehensive judgment specifically includes: Based on the combination of creep / non-creep properties of the load-bearing sections at peak and trough values, they are divided into the following two categories: PC type: There is no creep holding section at the peak of the cycle, but there is a creep holding section at the valley, which represents the combination of tensile plastic strain and reverse compressive creep strain. CP type: There is a creep-holding segment at the peak of the cycle, but no creep-holding segment at the valley, which represents the combination of tensile creep strain and reverse compressive plastic strain.

Images