Method and system for evaluating performance of steel weld under deep peak shaving condition

Abstract

The application discloses a steel weld joint performance evaluation method and system under deep peak regulation conditions, relates to the technical field of welded joint performance evaluation, and comprises the following steps: obtaining a target application scene working condition and a steel structure to be measured, combining historical steel weld joint abnormal data sets to screen a plurality of key weld joint anchor point sets; simulating a deep peak regulation environment to sample a plurality of key weld joint anchor point test data sequence set and determine a plurality of key weld joint anchor point performance trend feature set through performance trend feature time sequence interaction; extracting failure test data to evaluate working condition results, and determining a target application scene steel weld joint performance evaluation result through weighting. The application solves the technical problems that a traditional steel weld joint performance detection method does not design dynamic working conditions for deep peak regulation, resulting in inaccurate detection data, missing working condition adaptability and insufficient data correlation, and achieves the technical effects that steel weld joint performance detection data are more accurate and more targeted, and the steel weld joint performance precise evaluation demand under deep peak regulation conditions is met.

Description

Technical Field

[0001] This invention relates to the field of welded joint performance evaluation technology, and in particular to a method and system for evaluating the performance of steel welds under deep peak shaving conditions. Background Technology

[0002] Against the backdrop of national energy structure transformation, power plant units, in addition to meeting normal power supply demands, also need to undertake deep peak shaving tasks. The steel welds on their heated surfaces are subjected to harsh conditions of creep-fatigue interaction for extended periods. The performance of these welds directly affects the safe operation of the unit, making weld performance testing a crucial link in ensuring reliable equipment service. Current technologies for weld performance testing primarily employ traditional static tests and single-data evaluation methods, playing a certain role in testing under normal stable operating conditions. However, with the increasing safety requirements of deep peak shaving, the limitations of traditional testing technologies are becoming apparent. Due to the dynamic operating characteristics of deep peak shaving and the complexity of stress on the welds, traditional testing methods cannot accurately capture the dynamic performance changes of key welds, resulting in data with poor correlation and insufficient specificity, failing to meet the needs for accurate assessment of weld performance and safe management of the unit under deep peak shaving conditions. Summary of the Invention

[0003] This application provides a method and system for evaluating the performance of steel welds under deep peak shaving conditions, which solves the technical problems of traditional steel weld performance testing methods that do not have dynamic working condition design for deep peak shaving, resulting in inaccurate test data, lack of working condition adaptability and insufficient data correlation.

[0004] The first aspect of this application provides a method for evaluating the performance of steel welds under deep peak shaving conditions. The method includes: acquiring multiple typical working conditions and the steel structure to be tested for a target application scenario, and filtering multiple sets of key weld anchor points of the steel structure under multiple typical working conditions by combining historical steel weld anomaly data sets; acquiring the load protection method and duration, traversing the multiple typical working conditions to perform deep peak shaving environment simulation, and mapping the multiple sets of key weld anchor points to time-series sampling according to preset sampling indicators to obtain multiple sets of key weld anchor point test data sequences; performing time-series interaction of performance trend characteristics on the multiple sets of key weld anchor point test data sequences to determine multiple sets of key weld anchor point performance trend characteristics; extracting multiple sets of key weld anchor point failure test data at the failure time from the multiple sets of key weld anchor point test data sequences, and combining the multiple sets of key weld anchor point performance trend characteristics to evaluate the steel weld performance, obtaining performance evaluation results for steel welds under multiple typical working conditions; and performing weighted calculations on the performance evaluation results for steel welds under multiple typical working conditions to determine the performance evaluation results for steel welds in the target application scenario.

[0005] A second aspect of this application provides a performance evaluation system for steel welds under deep peak shaving conditions. The system includes: a key weld anchor point acquisition module, used to acquire multiple typical working conditions and the steel structure to be tested in a target application scenario, and to filter multiple key weld anchor point sets of the steel structure under multiple typical working conditions by combining historical steel weld anomaly data sets; a test data sequence acquisition module, used to acquire the load holding method and duration, traverse the multiple typical working conditions to perform deep peak shaving environment simulation, and perform time-series sampling of the multiple key weld anchor point sets according to preset sampling indicators to obtain multiple key weld anchor point test data sequence sets; and a performance trend feature set acquisition module. The module is used to perform time-series interaction of performance trend characteristics on the multiple key weld anchor point test data sequence sets to determine the performance trend characteristic set of multiple key weld anchor points; the steel weld performance evaluation result acquisition module is used to extract the failure test data sets of multiple key weld anchor points at the failure time of the multiple key weld anchor point test data sequence sets, and combine them with the performance trend characteristic set of multiple key weld anchor points to perform steel weld performance evaluation to obtain steel weld performance evaluation results under multiple typical working conditions; the steel weld performance evaluation result determination module is used to perform weighted calculation on the steel weld performance evaluation results under multiple typical working conditions to determine the steel weld performance evaluation result for the target application scenario.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0007] This application selects a set of key weld anchor points by combining typical operating conditions of the target application scenario, the steel structure under test, and historical abnormal data of steel welds; obtains load protection methods and durations to simulate deep peak shaving environments under various operating conditions, and samples the anchor points in a time sequence according to preset sampling indicators to obtain test data sequences; performs time-series interaction of performance trend characteristics on the data sequences to determine performance trend characteristics, extracts failure time data, and evaluates weld performance under single operating conditions in combination with trend characteristics; and performs weighted calculations on the evaluation results of multiple operating conditions to accurately evaluate the performance of steel welds under deep peak shaving conditions, making the steel weld performance evaluation results more accurate and reliable, meeting the unit's safety management and control requirements, and achieving the technical effect of making steel weld performance testing data more accurate and targeted, meeting the need for accurate evaluation of steel weld performance under deep peak shaving conditions. Attached Figure Description

[0008] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1This is a flowchart illustrating the method for evaluating the performance of steel welds under deep peak shaving conditions provided in the embodiments of this application.

[0010] Figure 2 This is a schematic diagram of the steel weld performance evaluation system under deep peak shaving conditions provided in the embodiments of this application.

[0011] Figure labeling: Module 1 for obtaining key weld anchor point set, Module 2 for obtaining test data sequence set, Module 3 for obtaining performance trend characteristic set, Module 4 for obtaining steel weld performance evaluation results, and Module 5 for determining steel weld performance evaluation results. Detailed Implementation

[0012] This application provides a method and system for evaluating the performance of steel welds under deep peak shaving conditions, which solves the technical problems of traditional steel weld performance testing methods that do not have dynamic working condition design for deep peak shaving, resulting in inaccurate test data, lack of working condition adaptability and insufficient data correlation.

[0013] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0014] It should be noted that the terms "first," "second," etc., in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules not explicitly listed or inherent to such processes, methods, products, or devices.

[0015] Example 1, as Figure 1 As shown, a method for evaluating the performance of steel welds under deep peak-shaving conditions, wherein the method includes:

[0016] Step A100: Obtain multiple typical working conditions and the steel structure to be tested for the target application scenario, and combine them with the historical steel weld anomaly data set to filter multiple key weld anchor points of the steel structure to be tested under multiple typical working conditions.

[0017] Specifically, firstly, the historical steel weld abnormality data set is retrieved for multiple typical working conditions and the steel structure to be tested to obtain the corresponding abnormal data set. Then, the historical steel weld abnormality dataset of the first typical working condition is extracted and merged. The historical steel weld location set of the first typical working condition is obtained by retrieving according to the weld location. After filtering this set to determine the first key weld anchor point set, multiple key weld anchor point sets are added. The specific steps are explained in detail in A110-A140.

[0018] Step A200: Obtain the load protection method and duration, traverse the multiple typical working conditions to perform in-depth peak-shaving environment simulation, and perform time-series sampling on the multiple key weld anchor point sets according to the preset sampling index to obtain a set of test data sequences for multiple key weld anchor points.

[0019] Optionally, firstly, the load holding method and holding time are obtained: From the high-temperature creep test at 650℃ and the high-temperature endurance test at 650℃, the load holding method is determined to be cumulative timing, with a total holding time of 18,000 hours / point, using a combination of target isochronous samples and repeat samples; From the high-temperature low-cycle fatigue test at 650℃ and the creep-fatigue interaction test at 650℃, the load holding method is determined to be distributed under multiple strain amplitudes and load holding conditions, with a total holding time of 800 hours for each condition; From the high-temperature aging accelerated aging test at 700℃±4℃, the load holding method is determined to be sampling according to time nodes, with a total holding time of 14,000 hours, and sampling time nodes of 100 hours, 500 hours, 1000 hours, 3000 hours, 7000 hours, and 14000 hours.

[0020] Next, deep peak-shaving environment simulations were conducted for 30%, 50%, and 70% rated load fluctuations under multiple typical operating conditions. Taking the 30% rated load fluctuation condition as an example, the cumulative time-based load maintenance method at 650℃ was used for 18,000 hours / point, combined with target isochronous samples and repeat samples. At the same time, high-temperature low-cycle fatigue at 650℃ and creep-fatigue interaction at 650℃ were carried out, that is, 800 hours were carried out under multiple strain amplitudes and load maintenance conditions, and a high-temperature aging accelerated aging test at 700℃±4℃ was carried out for a total of 14,000 hours according to the above time nodes. For the 50% and 70% rated load fluctuation conditions, the deep peak-shaving environment was reproduced according to the load maintenance methods and durations recorded in the above steps.

[0021] Then, according to the preset sampling indicators, time-strain curves, fracture elongation, fracture morphology, and stress-strain cycle curves, multiple key weld anchor point sets were mapped and time-series sampled: during the environmental simulation of each typical working condition, at different load holding time nodes, such as 100 hours and 500 hours of high-temperature aging, and cumulative timing nodes of high-temperature creep, time-strain curves of key weld anchor points were collected to understand the change of strain over time, fracture elongation was measured to understand fracture elongation performance, fracture morphology was observed to analyze micro-fracture characteristics, and stress-strain cycle curves were collected to understand the strain response under cyclic stress. The specific acquisition method is explained in detail in A210. Finally, a set of test data sequences of multiple key weld anchor points was obtained.

[0022] By acquiring the load maintenance method and duration of the deep peak shaving test, traversing the typical working condition simulation environment, and sampling the key weld anchor points in sequence according to multiple indicators, the system achieved the effect of comprehensively and accurately acquiring the performance evolution data of key weld anchor points under the deep peak shaving condition, providing a detailed data foundation for subsequent performance evaluation.

[0023] Step A300: Perform time-series interaction of performance trend characteristics on the test data sequence set of the multiple key weld anchor points to determine the performance trend characteristic set of the multiple key weld anchor points.

[0024] In one embodiment of this application, a multi-scale performance feature recognizer is invoked to analyze a set of test data sequences of multiple key weld anchor points to obtain a set of performance feature groups of multiple key weld anchor points. Then, the performance feature groups are subjected to time-series interaction of performance trend features to determine a set of performance trend features of multiple key weld anchor points. The specific steps are described in detail in A310-A320.

[0025] Step A400: Extract the test data sets of multiple key weld anchor points at the failure time, and combine them with the performance trend characteristics set of multiple key weld anchor points to evaluate the performance of the steel weld and obtain the performance evaluation results of steel welds under multiple typical working conditions.

[0026] Specifically, firstly, when extracting the failure test data sets of multiple key weld anchor points at the failure time, it is necessary to first determine the failure time of each key weld anchor point test data sequence, that is, the time node when the weld exhibits failure phenomena such as fracture or leakage, corresponding to the last data record in the data sequence. Then, for each key weld anchor point, all preset sampling index data for the failure time are extracted from its corresponding test data sequence, including the final strain value of the time-strain curve, the final measured value of the fracture elongation, the final microscopic characteristics of the fracture morphology, and the final cyclic parameters of the stress-strain cycle curve. These data are integrated into a single key weld anchor point failure test data set. After summing up the failure test data sets of multiple key weld anchor points, multiple key weld anchor point failure test data sets are formed.

[0027] For example, in a TP347H steel weld anchor point, the failure time was 15200h during a 18000h high-temperature creep test at 650℃. The failure data extracted at this time included a final strain of 0.78% on the time-strain curve, a fracture elongation of 10.5%, a fracture morphology showing a reduction in the number of dimples and signs of intergranular cracking, and a final cyclic peak stress of 470MPa on the stress-strain cycle curve. These data together constitute the failure test data set of this key weld anchor point.

[0028] Next, the performance of the steel weld is evaluated by combining the performance trend characteristics of multiple key weld anchor points. During the evaluation process, it is necessary to correlate the failure test data of each key weld anchor point with its performance trend characteristics: on the one hand, to verify whether the failure test data is consistent with the final evolution direction of the performance trend characteristics. For example, if the performance trend characteristics show that the strain growth rate of a certain anchor point gradually accelerates with the test time, the final strain value at the time of failure must conform to this accelerated growth law. If they are consistent, it indicates that the performance evolution process is reasonable; if they are inconsistent, the validity of the test data needs to be investigated. On the other hand, based on the performance trend characteristics, the degree of performance degradation reflected by the failure test data is judged. For example, if the performance trend characteristics of a certain key weld anchor point show that its fracture elongation decreases from 15% at the beginning of the test at a rate of 0.2% per month, and the fracture elongation at the time of failure is 10.5%, combined with this downward trend, it can be deduced that its performance degradation has reached 85% of the design critical value, thus determining that the weld performance corresponding to this anchor point does not meet the long-term service requirements under this typical working condition.

[0029] Subsequently, the above evaluation process was completed for each typical working condition: For the 30% rated load fluctuation working condition, the failure test data set of all key weld anchor points under this working condition was extracted, and combined with the performance trend characteristic set of multiple key weld anchor points corresponding to this working condition, the overall performance status of the weld under this working condition was judged, such as whether there is general performance degradation or concentrated failure risk areas; the same operation was repeated for the 50% and 70% rated load fluctuation working conditions, and finally, an independent steel weld performance evaluation conclusion was generated for each typical working condition. These conclusions together constitute the steel weld performance evaluation results for multiple typical working conditions.

[0030] By first extracting full-dimensional test data at the failure moment of key weld anchor points, then combining performance trend characteristics for correlation analysis and evaluation, and finally generating evaluation results for each typical working condition, the system achieves the effect of accurately judging the performance status and failure risk of steel welds under typical working conditions of different depth peak shaving, and providing a reliable basis for subsequent weighted calculation of the overall evaluation results of the target application scenario.

[0031] Step A500: Perform a weighted calculation on the performance evaluation results of steel welds under multiple typical working conditions to determine the performance evaluation results of steel welds for the target application scenario.

[0032] Specifically, firstly, the performance evaluation results of steel welds under multiple typical working conditions are clarified. These results are the failure test data sets of multiple key weld anchor points extracted under each typical working condition of the target application scenario in the aforementioned steps. Combined with the performance trend characteristics set of key weld anchor points under the corresponding working conditions, independent evaluation conclusions are generated for each typical working condition through performance degradation degree analysis, failure risk judgment, and other methods. Each evaluation conclusion needs to be quantified into a calculable performance index, such as a performance score of 0-100 points. The higher the score, the better the performance of the steel weld and the lower the failure risk.

[0033] The process of quantifying the performance into calculable indicators is as follows: First, based on the failure test data and performance trend characteristics of key weld anchor points under each typical working condition, the evaluation dimensions are defined, such as the consistency between failure data and performance trends, the degree of performance degradation, and the correlation of failure risk. Then, corresponding scoring rules are set for each dimension. For example, if the failure data and trend are highly consistent, the performance degradation is slight, and the failure risk is low, a high score range is given, and vice versa. Subsequently, weights are assigned according to the importance of each dimension in the performance evaluation of steel welds. The scores are converted by combining the performance of each dimension. Finally, the scores are normalized to the range of 0-100 points to obtain the calculable performance indicators corresponding to the independent evaluation conclusions of each typical working condition.

[0034] Next, the core first step in weighted calculation is to determine the weight of each typical operating condition. The weight allocation needs to be based on the actual operating proportion of each typical operating condition in the target application scenario, that is, the proportion of the operating time of a certain typical operating condition in the target application scenario to the total peak-shaving operating time. The higher the operating proportion, the greater the impact of the operating condition on the long-term performance of the steel weld, and the higher the corresponding weight. The specific steps are as follows: First, calculate the actual operating time of each typical operating condition in the target application scenario over the past 1-3 operating cycles. For example, the annual operating time is 1200 hours for the 30% rated load fluctuation condition, 2500 hours for the 50% rated load fluctuation condition, and 1800 hours for the 70% rated load fluctuation condition, with a total peak shaving operating time of 5500 hours. Then, calculate the weight using the formula: Weight of a certain operating condition = Operating time of that operating condition / Total peak shaving operating time. The weights are approximately 1200 / 5500≈0.218 for the 30% operating condition, 2500 / 5500≈0.455 for the 50% operating condition, and 1800 / 5500≈0.327 for the 70% operating condition. This ensures that the weights objectively reflect the actual impact of each operating condition.

[0035] Then, the core second step of the weighted calculation is to calculate and sum the weighted contribution values ​​of the evaluation results for each working condition. The quantitative performance score for each typical working condition is multiplied by its corresponding weight to obtain the weighted contribution value for that working condition. The weighted contribution values ​​for all working conditions are then summed to obtain the comprehensive performance score of the steel weld in the target application scenario. For example, if the performance evaluation result for 30% of the working conditions is 92 points, it means the steel weld performance is good and the failure risk is low under that condition; 68 points for 50% of the working conditions indicates critical performance and a moderate failure risk; and 45 points for 70% of the working conditions indicates deteriorated performance and a high failure risk. Then the weighted contribution values ​​for each working condition are 92 × 0.218 ≈ 20.06, 68 × 0.455 ≈ 30.94, and 45 × 0.327 ≈ 14.72, respectively. The comprehensive score is approximately 20.06 + 30.94 + 14.72 ≈ 65.72 points.

[0036] Finally, the performance evaluation results of the steel welds in the target application scenario are determined based on the comprehensive score. A standard for corresponding comprehensive scores and performance levels should be established in advance. For example, a score of 80-100 indicates excellent performance, requiring no additional maintenance and only routine monitoring; a score of 60-79 indicates acceptable performance, suggesting additional specialized monitoring every 3 months, focusing on the weld condition under high-weight conditions; and a score of 0-59 indicates unacceptable performance, requiring immediate shutdown and maintenance, and replacement of high-risk weld sections. Taking the comprehensive score of 65.72 in the example above, the performance level of the steel welds in the target application scenario can be determined as acceptable. Furthermore, considering the weighting structure, the 50% condition with a weight of 0.455 has the greatest impact on the comprehensive result. Subsequent monitoring should focus on key weld anchor points under this condition to ensure that the evaluation results can directly guide actual operation and maintenance.

[0037] By first determining the weights based on the proportion of operating conditions, then calculating and summing the weighted contribution values ​​of each operating condition, and finally determining the final evaluation result in combination with the scoring criteria, the system achieves the effect of comprehensively integrating the impact of different typical operating conditions on the performance of steel welds, objectively reflecting the overall performance status of steel welds in the target application scenario, and providing precise guidance for unit operation and maintenance.

[0038] Furthermore, step A100 in the method provided in this application embodiment includes:

[0039] A110: Using the multiple typical working conditions and the steel structure to be tested, the historical steel weld anomaly data set is retrieved to obtain multiple typical working condition historical steel weld anomaly data sets.

[0040] A120: Extract the first typical working condition historical steel weld abnormal data set from the multiple typical working condition historical steel weld abnormal data sets, and retrieve the first typical working condition historical steel weld abnormal data set by weld position to obtain the first typical working condition historical steel weld position set.

[0041] A130: Filter the set of historical steel weld locations under the first typical working condition to determine the set of anchor points for the first critical weld.

[0042] A140: Add the first set of critical weld anchor points to a set of multiple critical weld anchor points.

[0043] Specifically, the first step is to acquire a historical collection of abnormal steel weld data. This is achieved by those skilled in the art through collecting abnormal records of steel welds of various steel structures, such as TP347H steel heat-receiving tubes, T91 steel base materials, and dissimilar steel welded components, under different operating conditions during the past operation of the power plant unit. These abnormal records include relevant data on weld leakage, crack formation, and early failure. Specifically, they can be extracted from the unit's daily operation and maintenance logs, maintenance reports, fault analysis records, and other materials, and then categorized and organized according to dimensions such as operating conditions and steel structure type to form a searchable historical collection of abnormal steel weld data. This lays the foundation for subsequent data matching between typical operating conditions of the target application scenario and the steel structure to be tested.

[0044] Next, the association between the historical steel weld anomaly data set and the target application scenario is established. Specifically, the historical steel weld anomaly data set is precisely retrieved using multiple typical operating conditions of the target application scenario and the steel structure to be tested as search criteria. For example, if the target application scenario is deep peak shaving of power plant units, multiple typical operating conditions can be set as 30% rated load fluctuation, 50% rated load fluctuation, and 70% rated load fluctuation, and the steel structure to be tested is TP347H steel heat-receiving pipe. Using these two criteria, the historical steel weld anomaly data set is retrieved to find records of weld leakage, cracks, and other anomalies of this type of heat-receiving pipe under the corresponding peak shaving conditions over the past 5 years. Assuming 120 anomaly data points are obtained under 30% load fluctuation, 150 anomaly data points under 50% load fluctuation, and 90 anomaly data points under 70% load fluctuation, a historical steel weld anomaly data set for multiple typical operating conditions is formed. This step achieves a preliminary match between the historical steel weld anomaly data and the specific evaluation scenario, avoiding the problem of data generalization.

[0045] Next, it is necessary to further focus on the weld location information of a single typical working condition. From the historical steel weld anomaly data sets of multiple typical working conditions obtained above, one representative working condition is selected as the first typical working condition. For example, the working condition of 50% rated load fluctuation, which has a high frequency of anomalies, is selected, and the corresponding historical steel weld anomaly data set of the first typical working condition is extracted, that is, the 150 anomaly data under the 50% load fluctuation in the example above. Subsequently, using the weld location as a new search dimension, a second search is performed on the 150 anomaly data, distinguishing different locations such as butt welds, fillet welds, and lap welds of the heated surface pipe, and statistically obtaining the results of the corresponding number of anomalies in butt welds, fillet welds, and lap welds, thereby obtaining the historical steel weld location set of the first typical working condition.

[0046] Subsequently, based on the historical steel weld location set of the first typical working condition, weld anomaly identification is performed on the steel structure to be tested to obtain the identified steel structure to be tested. Then, the weld corresponding to the maximum number of weld anomaly identifications is extracted as the first critical weld anchor point. Finally, the anchor point is further filtered according to the preset first and second filtering bandwidths to determine the first critical weld anchor point set. The specific steps are explained in detail in A131-A133.

[0047] Finally, to ensure that subsequent performance evaluations cover the critical welds under this typical operating condition, the first set of critical weld anchor points needs to be added to multiple sets of critical weld anchor points. Specifically, all weld anchor point elements in the first set of critical weld anchor points are merged one by one into the multiple sets of critical weld anchor points originally used for performance evaluation. This allows the multiple sets of critical weld anchor points to include not only the critical weld anchor points from other operating conditions or other screening logics, but also the newly added first set of critical weld anchor points obtained from the screening of the first typical operating condition. This expands the range of critical weld anchor points used for subsequent test data sampling and performance analysis, ensuring the comprehensiveness of the evaluation.

[0048] Furthermore, step A130 in the method provided in this application embodiment includes:

[0049] A131: Based on the historical steel weld location set of the first typical working condition, the steel structure to be tested is marked with weld anomalies to obtain the marked steel structure to be tested.

[0050] A132: Extract the weld with the maximum number of weld anomaly markers from the steel structure to be tested, and use it as the first critical weld anchor point.

[0051] A133: According to the preset first screening bandwidth and the preset second screening bandwidth, the first key weld anchor points are further screened in the identified steel structure to be tested to obtain the set of the first key weld anchor points.

[0052] Optionally, based on the historical steel weld location set of the first typical operating condition, weld anomalies are first marked on the steel structure to be tested. Taking the first typical operating condition of 50% rated load fluctuation of power plant unit as an example, if the historical steel weld location set under this operating condition contains information on 20 weld locations where abnormalities such as leakage and cracks have occurred, then on the TP347H steel heat-receiving tube steel structure to be tested, these corresponding locations are marked one by one to obtain the marked steel structure to be tested.

[0053] Subsequently, the number of anomaly markers in each weld area of ​​the steel structure was statistically analyzed and compared. Specifically, the marked steel structure under test was divided into multiple independent weld segments, such as based on the spatial distribution of welds and structural type. Then, for each weld segment, the total number of historical anomaly markers under the first typical working condition was summarized. For example, if there are three main segments: weld segment X, weld segment Y, and weld segment Z: weld segment X has a cumulative total of 15 anomaly markers in the historical record, covering various anomalies such as leakage and cracks; weld segment Y has a cumulative total of 9; and weld segment Z has a cumulative total of 6. By comprehensively comparing the number of anomaly markers in all weld segments, the weld segment with the most anomaly markers, namely weld segment X, was selected because it had the highest frequency of anomalies and the most prominent risk level under this typical working condition. Therefore, the weld corresponding to weld segment X was extracted and identified as the first critical weld anchor point, providing precise guidance for subsequent performance evaluation focusing on high-risk areas.

[0054] Finally, based on the two filtering bandwidths, the first and second derived key weld anchors are derived from the first key weld anchor. Then, it is determined whether the clustered neighborhood density of the two derived key weld anchors is greater than or equal to that of the first key weld anchor. If so, the first and second derived filtering key weld anchor sets are obtained by further derivation according to the corresponding bandwidth. Finally, the union of these two sets is calculated to obtain the first key weld anchor set. The specific steps are explained in detail in A133-1-A133-4.

[0055] Furthermore, step A133 in the method provided in this application embodiment includes:

[0056] A133-1: Based on the first screening bandwidth and the second screening bandwidth, the first critical weld anchor point is derived to obtain the first derived critical weld anchor point and the second derived critical weld anchor point.

[0057] A133-2: Determine whether the clustered neighborhood density of the first derived critical weld anchor point is greater than or equal to the clustered neighborhood density of the first critical weld anchor point. If so, continue to perform derivative filtering on the first derived critical weld anchor point according to the first filtering bandwidth to obtain the first derived filtered critical weld anchor point set.

[0058] A133-3: Determine whether the clustered neighborhood density of the second derived critical weld anchor point is greater than or equal to the clustered neighborhood density of the first critical weld anchor point. If so, continue to perform derivative screening on the second derived critical weld anchor point according to the second screening bandwidth to obtain the second derived screening critical weld anchor point set.

[0059] A133-4: Obtain the first set of key weld anchor points by performing a union operation on the first set of derived screening key weld anchor points and the second set of derived screening key weld anchor points.

[0060] In this embodiment, the screening bandwidth refers to the different physical distances used between anchor points when performing derivative screening on critical weld anchor points, which can cover different ranges to reduce anchor point omissions. The clustering neighborhood density is the ratio of the number of anchor points in a circular region constructed with the anchor point as the center and the corresponding screening radius to the area of ​​that region, used to characterize the degree of anchor point clustering.

[0061] Specifically, to more comprehensively screen high-risk weld anchor points, the first critical weld anchor point is first derived according to a first screening bandwidth and a second screening bandwidth with different distances. The first critical weld anchor point is the weld location with the most abnormal markings on the steel structure under test. Centered on this anchor point, first derived critical weld anchor points and second derived critical weld anchor points are derived along the corresponding physical distance. By extending with different bandwidths, different ranges of potential high-risk areas are covered, reducing the omission of anchor points.

[0062] Next, the clustered neighborhood density is calculated. Specifically, a spherical volume region is constructed with the anchor point as the center and the radius of the corresponding filtering bandwidth. The number of anchor points within the spherical region is counted and divided by the region volume to obtain the clustered neighborhood density, and it is then determined whether to continue the derivative filtering. Assuming that the clustered neighborhood density is calculated based on the spherical volume and number of anchor points of the first critical weld anchor point and the number of anchor points of the first derived critical weld anchor point within the same radius spherical volume region, if the clustered neighborhood density of the first derived critical weld anchor point is greater than or equal to that of the first critical weld anchor point, the derivative filtering is continued for the first derived critical weld anchor point according to the first filtering bandwidth. If there are multiple anchor points with equidistant clustered neighborhood densities, one is randomly selected. This process continues until the clustered neighborhood density of subsequent derived critical weld anchor points is less than that of the first critical weld anchor point, thus stopping the process. Finally, the set of first derived filtered critical weld anchor points is obtained.

[0063] Next, for the second derived critical weld anchor point, the corresponding clustering neighborhood density is calculated as follows: Construct a spherical volume region with the second screening bandwidth as the radius, and calculate the clustering neighborhood density of the first critical weld anchor point and the second derived critical weld anchor point in this region in the same way as the above steps. If the clustering neighborhood density of the second derived critical weld anchor point is greater than or equal to that of the first critical weld anchor point, continue to perform derivative screening on the second derived critical weld anchor point according to the second screening bandwidth to obtain the set of second derived key weld anchor points.

[0064] Finally, a union operation is performed on the first and second derived sets of key weld anchor points to remove redundant duplicate anchor points, thus obtaining the first set of key weld anchor points.

[0065] By first generating anchor points according to different screening bandwidths, then determining whether to continue generating them based on the density of the clustered neighborhood, and finally merging the generated sets, high-risk steel weld anchor points under deep peak shaving conditions are screened more comprehensively and thoroughly, reducing the omission of anchor points.

[0066] Furthermore, step A133-2 in the method provided in this application embodiment includes:

[0067] A133-2A: When the clustered neighborhood density of the first derived critical weld anchor point is less than the clustered neighborhood density of the first critical weld anchor point, calculate the difference between the clustered neighborhood density of the first derived critical weld anchor point and the clustered neighborhood density of the first critical weld anchor point. If the difference in clustered neighborhood density is less than or equal to a preset density difference threshold, then continue to perform derivative filtering on the first derived critical weld anchor point according to the first filtering bandwidth to obtain the first derived filtered critical weld anchor point set.

[0068] Specifically, in the derivative screening process for steel weld performance evaluation under deep peak shaving conditions, when the clustered neighborhood density of the first derivative critical weld anchor point is less than that of the first critical weld anchor point, it is necessary to first calculate the difference in their clustered neighborhood densities. Taking the first critical weld anchor point as the weld with the most abnormal indicators under 50% rated load fluctuation conditions of TP347H steel heat-receiving surface tube as an example, its clustered neighborhood density is calculated using the first screening bandwidth as a spherical volume region. Then, the first derivative critical weld anchor point is also constructed with the same first screening bandwidth as the screening radius to form a spherical volume region, and the number of anchor points in the region is obtained. The clustered neighborhood density of the derivative critical weld anchor point is calculated and compared with that of the first critical weld anchor point. If the clustered neighborhood density of the derivative critical weld anchor point is less than that of the first critical weld anchor point, it is necessary to further calculate the difference in their clustered neighborhood densities.

[0069] Next, a preset density difference threshold is obtained. This step requires those skilled in the art to base their understanding on historical test data of the target application scenario and the performance failure patterns of steel welds, combined with the creep rupture test, fatigue test, microstructure evolution, and corresponding mechanical property test results of in-service TP347H and T91 steel base materials and dissimilar steel welds. Specifically, it is necessary to collect the clustered neighborhood density data of critical anchor points and derived critical weld anchor points of TP347H and T91 steel welds under deep peak shaving conditions over the past 5 years, as well as records of anomalies such as leakage and cracks that occurred in the subsequent actual service of these anchor points. Assuming that statistical analysis reveals that when the clustered neighborhood density difference between the derived critical weld anchor point and the first critical weld anchor point is ≤0.015 points / mm², the probability of subsequent performance failure of the derived critical weld anchor point deviates from the failure probability of the first critical weld anchor point by less than 10%, meaning that their risk levels are similar; if the difference exceeds 0.015 points / mm², the failure probability of the derived critical weld anchor point will drop sharply and the risk will be significantly reduced. Based on this statistical result, 0.015 per square millimeter was set as the preset density difference threshold to ensure that the selected derivative critical weld anchor points still have high-risk characteristics and avoid missing potential failure areas.

[0070] Next, after obtaining the density difference of the clustered neighborhood and the preset density difference threshold, a comparison and judgment step is performed. On the one hand, if the density difference of the clustered neighborhood is less than or equal to the preset density difference threshold, it means that the clustered neighborhood density of the first derived critical weld anchor point is similar to that of the first critical weld anchor point, and it still belongs to the high-risk area category. It can be included in the screening range, and the first derived critical weld anchor point is further screened according to the first screening bandwidth, gradually expanding the anchor point range, and finally forming the first derived screening critical weld anchor point set.

[0071] On the other hand, if the density difference of the clustered neighborhood is greater than the preset density difference threshold, it means that the clustered neighborhood density of the derived critical weld anchor point is significantly different from that of the first critical weld anchor point, and the risk level is significantly lower than that of the first critical weld anchor point. It does not meet the requirements of a high-risk anchor point, so it is necessary to re-select the first derived critical weld anchor point. For example, new anchor points can be derived from other surrounding positions of the first critical weld anchor point according to the first screening bandwidth, and the density calculation and difference judgment steps can be repeated until the first derived critical weld anchor point that meets the requirements is obtained.

[0072] By first calculating the density difference between the derived anchor points with a density less than that of the first critical weld anchor point and the former, and then making a judgment based on a preset density difference threshold determined by historical test and performance data, the decision is made on whether to continue derivation or re-screen based on the judgment result. This achieves the effect of avoiding the omission of derived anchor points with similar risk to the first critical weld anchor point, and ensuring that the first set of derived critical weld anchor points only includes high-risk anchor points.

[0073] Furthermore, step A200 in the method provided in this application embodiment includes:

[0074] A210: The preset sampling indicators include time-strain curve, fracture elongation, fracture morphology and stress-strain cycle curve.

[0075] In one embodiment, the time-strain curve is used to record the relationship between strain and time in steel welds under load during deep peak-shaving conditions. It can intuitively reflect the strain evolution law of the weld during creep or fatigue. To obtain this curve, strain sensors need to be installed at multiple key weld anchor points in a deep peak-shaving environment simulation test, such as a 650℃ high-temperature creep test or a 650℃ high-temperature endurance test. A fixed time sampling interval is set, for example, data is collected every 10 hours, continuously collecting strain data throughout the entire test cycle. Taking a 18000h 650℃ high-temperature creep test as an example, the strain at a key weld anchor point is 0.18% at the 1000th hour, 0.35% at the 5000th hour, 0.52% at the 10000th hour, and 0.75% at the 18000th hour. By mapping these strain values ​​at different time points one by one, the time-strain curve of that key weld anchor point can be plotted, thus forming a set of time-strain curves for multiple key weld anchor points.

[0076] Next, regarding the elongation at break, it is an important indicator for measuring the plasticity of steel welds. It refers to the percentage of the weld specimen's elongation at fracture relative to its original gauge length; a higher value indicates better weld plasticity. This indicator needs to be obtained after the test specimen fails, such as after a high-temperature creep test or a high-temperature low-cycle fatigue test. First, determine the original gauge length of the specimen, for example, set it to 50 mm. After the specimen fractures, use a vernier caliper with an accuracy of 0.01 mm to measure the gauge length of the fractured specimen. Taking a TP347H steel weld specimen after a 18,000-hour high-temperature creep test at 650℃ as an example, the original gauge length was 50 mm, and the gauge length after fracture was 57.5 mm. According to the formula of fracture elongation = (gauge length after fracture - original gauge length) / original gauge length × 100%, the fracture elongation of the specimen is calculated to be (57.5-50) / 50×100%=15%. By performing the same measurement and calculation on specimens corresponding to multiple key weld anchor points, the fracture elongation data of multiple key weld anchor points can be obtained.

[0077] Then, the fracture morphology refers to the microscopic or macroscopic morphological characteristics of the fracture surface after the steel weld fractures. It can reflect the fracture mechanism of the weld, including ductile fracture and brittle fracture, and is an important basis for analyzing the cause of weld failure. To obtain the fracture morphology, after the weld sample fails, the fracture surface must first be cleaned. Anhydrous ethanol can be used to remove surface impurities, and then microscopic observation and analysis can be performed using an electron backscatter microscope or a transmission electron microscope. Taking a T91 steel dissimilar steel weld sample after a 650℃ high-temperature creep-fatigue interaction test for 800 hours as an example, transmission electron microscopy revealed a large number of dimple structures with a diameter of about 2-5 μm on the fracture surface, and a small number of second-phase particles were distributed in the dimples. These characteristics indicate that the weld is a ductile fracture. At the same time, the macroscopic characteristics of the fracture surface, including the flatness of the fracture surface and the presence of crack propagation traces, were recorded. The observed microscopic and macroscopic characteristics were organized to form the fracture morphology data of the key weld anchor point, and then the fracture morphology of multiple key weld anchor points was collected.

[0078] Finally, the stress-strain cyclic curve records the relationship between stress and strain in a steel weld under cyclic loading conditions, reflecting the stress-strain response and cyclic hardening / softening characteristics during fatigue. Obtaining this curve requires conducting high-temperature, low-cycle fatigue tests. A preset cyclic strain load is applied to key weld anchor points, for example, setting the strain amplitude to ±0.6% and the cycle frequency to 0.1Hz. Stress data is collected every 10 cycles. Taking a key weld anchor point as an example, the maximum stress is 560MPa and the minimum stress is 120MPa in the first cycle; the maximum stress is 530MPa and the minimum stress is 115MPa in the 100th cycle; and the maximum stress is 500MPa and the minimum stress is 110MPa in the 500th cycle. After 800 hours of testing, the stress and strain values ​​corresponding to each cycle are mapped one-to-one to plot the stress-strain cyclic curve for that key weld anchor point. Similarly, a set of stress-strain cyclic curves for multiple key weld anchor points can be obtained.

[0079] By clearly defining the preset sampling indicators for time-strain curves, fracture elongation, fracture morphology, and stress-strain cycle curves, and combining them with specific acquisition steps in deep peak shaving-related experimental design and experimental data, the system achieved comprehensive and accurate acquisition of multi-dimensional performance data of steel welds under deep peak shaving conditions. This provides detailed basic data support for the subsequent construction of key weld anchor point test data sequence sets and the performance evaluation of steel welds.

[0080] Furthermore, step A300 in the method provided in this application embodiment includes:

[0081] A310: Call the multi-scale performance feature recognizer to analyze the test data sequence set of the multiple key weld anchor points to obtain a set of multiple key weld anchor point performance feature groups.

[0082] A320: Perform time-series interaction of performance trend features on the set of multiple key weld anchor points to determine the set of performance trend features for the multiple key weld anchor points.

[0083] In this embodiment, the multi-scale performance feature recognizer is a tool constructed from a neural network model for analyzing a set of test data sequences of multiple key weld anchor points of steel welds under deep peaking conditions.

[0084] Specifically, firstly, a multi-scale performance feature recognizer is invoked to analyze a set of test data sequences for multiple key weld anchor points. The core function of the multi-scale performance feature recognizer is to extract key features from two dimensions: time scale and performance index scale. On the time scale, for the time-strain data sequence of high-temperature creep tests, the recognizer divides it into three sub-scales: short-term (0-100h), medium-term (1000-5000h), and long-term (10000-18000h), extracting features such as strain growth rate and strain fluctuation amplitude for each stage. On the performance index scale, for the stress-strain cycle curve, features such as the maximum and minimum values ​​of cyclic peak stress and the cyclic hardening coefficient are extracted; for the fracture elongation, the numerical trend of change throughout the test is extracted; and for the fracture morphology, microstructural feature parameters such as dimple density are extracted.

[0085] Next, for each key weld anchor point's test data sequence, the multi-scale performance feature recognizer repeats the above multi-dimensional extraction process, integrating the extracted time-scale features and performance index-scale features into a structured performance feature set. For example, the performance feature set of a certain key weld anchor point includes features such as the short-term strain growth rate of high-temperature creep (0.0005% / h), the medium-term strain fluctuation (±0.02%), the long-term strain growth slope (0.00002875% / h), the peak cyclic stress of high-temperature low-cycle fatigue (560MPa / 120MPa), the cyclic hardening coefficient (0.02), the fracture elongation variation (15%-12%), and the dimple density of the fracture surface (3 dimples / μm²). After the test data sequences of all key weld anchor points have been analyzed, the corresponding structured performance feature sets together constitute multiple sets of key weld anchor point performance feature sets.

[0086] Finally, the execution of the time-series interaction steps for multiple key weld anchor point performance feature sets involves randomly extracting the first key weld anchor point performance feature set, and after feature combination, similarity recognition, time-series interaction, and mean calculation, obtaining the first key weld anchor point performance trend features and adding them to the set, thereby determining multiple key weld anchor point performance trend feature sets. The specific steps are explained in detail in A321-A326.

[0087] Furthermore, step A320 in the method provided in this application embodiment includes:

[0088] A321: Randomly extract the first key weld anchor point performance feature group from multiple sets of key weld anchor point performance feature groups.

[0089] A322: Perform pairwise enumeration and combination of the performance feature groups of the first critical weld anchor points to obtain the set of performance feature combinations of the first critical weld anchor points.

[0090] A323: Perform feature similarity identification on each combination in the first key weld anchor point performance feature combination set to obtain the first feature similarity set.

[0091] A324: Based on the first feature similarity set, perform time-series interaction on the first key weld anchor point performance feature combination set to obtain the first interactive key weld anchor point performance feature combination set.

[0092] A325: Based on the performance feature group of the first critical weld anchor point, the mean value of the performance feature combination set of the first interactive critical weld anchor point is calculated to obtain the performance trend feature group of the first interactive critical weld anchor point.

[0093] A326: Calculate the mean of the first interactive critical weld anchor point performance trend feature group to obtain the first critical weld anchor point performance trend feature, and add the first critical weld anchor point performance trend feature to a set of multiple critical weld anchor point performance trend features.

[0094] Optionally, firstly, after clarifying the basic source of multiple key weld anchor point performance characteristic groups, each performance characteristic group contains multi-dimensional characteristics of the corresponding anchor point, such as the short-term strain growth rate of the time-strain curve, the cyclic peak stress of the stress-strain cycle curve, the range of fracture elongation variation, and the density of dimples on the fracture surface. Based on this, the first key weld anchor point performance characteristic group is randomly extracted from this set. For example, the characteristic group of the 5th anchor point is randomly selected from multiple characteristic groups, which includes: 0.0005% / h strain growth rate in the short term (0-100h) of 650℃ high-temperature creep, 0.00008% / h strain growth rate in the medium term (1000-5000h), and 0.00002875% / h strain growth rate in the long term (10000-18000h); peak cyclic stress of 650℃ high-temperature low-cycle fatigue, with a maximum of 560MPa and a minimum of 120MPa, and a cyclic hardening coefficient of 0.02; fracture elongation decreasing from the initial 15% to 12% in the later stage of the test; and dimple density of 3 dimples / μm². This is used as the basic characteristic group for subsequent interactive analysis.

[0095] Next, the performance characteristic group of the first critical weld anchor point is enumerated and combined in pairs. If the performance characteristic group contains 7 independent characteristics, they can be denoted as A: short-term strain growth rate, B: medium-term strain growth rate, C: long-term strain growth rate, D: cyclic peak stress, E: cyclic hardening coefficient, F: fracture elongation change, and G: fracture dimple density. Then, according to the principle of pairwise non-repetition, 21 characteristic combinations can be obtained: AB, AC, AD, AE, AF, AG, BC, BD, BE, BF, BG, CD, CE, CF, CG, DE, DF, DG, EF, EG, and FG. Each combination represents the relationship between two different performance characteristics, which together constitute the set of performance characteristic combinations of the first critical weld anchor point, laying the foundation for subsequent analysis of the correlation between characteristics.

[0096] Subsequently, feature similarity identification is performed on each combination in the set of performance feature combinations for the first critical weld anchor point. The cosine similarity algorithm can be used to calculate the correlation between two features in each combination: for example, the short-term and medium-term strain growth rates of combination AB both reflect the strain growth pattern over time during creep testing, and their data trends are highly consistent; the assumed similarity calculation result is 0.85. The short-term strain growth rate and cyclic peak stress of combination AD, one being a creep time dimension feature and the other a fatigue stress dimension feature, have a weak correlation; the assumed similarity calculation result is 0.32. The short-term strain growth rate and fracture elongation change of combination AF show a certain correlation; the faster the short-term strain growth, the more pronounced the decrease in fracture elongation, and the assumed similarity calculation result is 0.68. After calculating for each of the above 21 combinations, a first feature similarity set containing 21 similarity values ​​is obtained.

[0097] Next, based on the first feature similarity set, the first key weld anchor point performance feature combination set is subjected to temporal interaction to obtain the first interactive key weld anchor point performance feature combination set. The specific steps are explained in detail in A324-1 below.

[0098] Subsequently, based on the performance feature group of the first critical weld anchor point, the mean value of the performance feature combination set of the first interactive critical weld anchor point is calculated. For each interactive combination, the basic features in its corresponding original feature group are retrieved, and the mean value of these features in the interactive combination is calculated. For example, for the A, B, and C features of the original group corresponding to the AB-AC interactive combination, the mean value of feature A after the interaction is (0.0005 + 0.00051) / 2 = 0.000505% / h, where 0.00 051 represents the adjusted value of feature A after time-series interaction. The mean value of feature B is (0.00008 + 0.000085) / 2 = 0.0000825% / h, and the mean value of feature C is (0.00002875 + 0.00003) / 2 = 0.000029375% / h. The AF interaction combination corresponds to the original group's features A and F, with mean values ​​of A = 0.00049% / h and F = (15% + 12% + 13.5%) / 3 = 13.5%. The mean values ​​of all interaction combinations are integrated to form the first interactive key weld anchor point performance trend feature group, which includes the mean values ​​of features A, B, C, D, E, F, and G.

[0099] Finally, the mean of the performance trend feature group of the first interactive key weld anchor point is calculated, and the overall mean is calculated again for the mean of each feature in the performance trend feature group: for example, the mean of feature A is 0.000505% / h, feature B is 0.0000825% / h, and feature C is 0.000029375% / h, and the mean of the three is (0.000505+0.0000825+0.000029375) / 3≈0.0002056% / h; the mean of feature F is 13.5%, and the mean of feature G is 2.8 pieces / μm², and the original values ​​are directly retained. These final averages are integrated into the performance trend characteristics of the first critical weld anchor point. Examples include the average strain growth rate during the creep stage (0.0002056% / h), the average fracture elongation (13.5%), the average dimple density on the fracture surface (2.8 dimples / μm²), and the average cyclic peak stress (340 MPa). This is then added to the performance trend characteristic set of multiple critical weld anchor points. The above steps are repeated for the performance characteristic groups of other anchor points to gradually improve the set.

[0100] By following a series of steps—from randomly extracting feature groups, enumerating combined features, identifying feature similarity, to temporal interactive integration, calculating the mean, and adding it to the set—the system accurately uncovers the performance evolution patterns of key weld anchor points, providing comprehensive and reliable performance trend characteristics to support subsequent performance evaluation of steel welds based on failure data.

[0101] Furthermore, step A324 in the method provided in this application embodiment includes:

[0102] A324-1: Traverse the first feature similarity set and perform normalization processing to construct a time-series interaction matrix set. Then, use the time-series interaction matrix set to perform time-series interaction on the first key weld anchor point performance feature combination set to obtain the first interactive key weld anchor point performance feature combination set.

[0103] In one embodiment, firstly, it is necessary to clarify the source of the first feature similarity set. This set is obtained by performing feature similarity identification on each combination of the performance feature combination of the first key weld anchor point in the aforementioned steps, for example, containing 21 feature combinations, corresponding to the time-strain curve, stress-strain cycle curve, and other multi-dimensional feature associations of the steel weld. The original set contains 21 similarity values, for example, 0.85, 0.78, 0.32, 0.45, 0.68, 0.52, 0.91, 0.55, 0.48, 0.72, 0.61, 0.88, 0.59, 0.75, 0.63, 0.41, 0.53, 0.67, 0.79, 0.58, and 0.48. Each value corresponds to the degree of association between two feature combinations, with a larger value indicating a stronger association.

[0104] Next, the first feature similarity set is traversed and normalized. Using the Min-Max normalization method, the maximum and minimum values ​​in the set are calculated first. In the example above, after statistical analysis, the maximum value is 0.91, corresponding to the feature combination of medium- and long-term strain growth rates; the minimum value is 0.32, corresponding to the feature combination of short-term strain growth rate and cyclic peak stress. Then, the normalization result is calculated one by one according to the formula: Normalized value = (Original value - Minimum value) / (Maximum value - Minimum value). For example, the normalized value of the original value 0.85 is (0.85 - 0.32) / (0.91 - 0.32) ≈ 0.898, the normalized value of the original value 0.32 is 0, and the normalized value of the original value 0.91 is 1. Finally, the normalized feature similarity set is obtained, ensuring that all similarity values ​​are unified to the 0-1 range, eliminating the impact of differences in the original value range on subsequent interactive calculations.

[0105] Subsequently, a set of temporal interaction matrices is constructed based on the normalized feature similarity set. Since the first key weld anchor point performance feature combination set contains 21 feature combinations, a 21-row, 21-column temporal interaction matrix is ​​constructed. A single temporal dimension results in a single matrix, while multiple temporal dimensions form a set of matrices. Here, the core temporal dimension of deep peak tuning is taken as an example. The rows and columns of the matrix correspond to the 21 feature combinations, and the matrix elements... The rule for determining the value is: when i=j, This is the normalized similarity value of the feature combination itself, representing the basic weight of the combination, such as the medium-term and long-term strain growth rates of the 7th combination. =1; when i≠j This represents the normalized average similarity between the i-th and j-th combinations. For example, the normalized value of the short-term and medium-term strain growth rates of the first combination is 0.898, while the normalized value of the seventh combination is 1. =(0.898+1) / 2=0.949. This rule is used to fill the matrix, forming a set of temporal interaction matrices. The larger the value in the matrix, the stronger the temporal correlation between the corresponding two feature combinations.

[0106] Finally, a time-series interaction matrix was used to perform time-series interaction on the set of performance feature combinations of the first critical weld anchor points. For the original data of each feature combination, such as the short-term strain growth rate of 0.0005% / h and the medium-term strain growth rate of 0.00008% / h for the first combination, a weighted fusion was performed with all elements in the corresponding row of the matrix. For example, when the first combination interacts with the seventh combination, the medium-term strain growth rate of both is 0.00008% / h, and then... The weighted calculation with a weight of 0.949 yields a mid-term strain growth rate of approximately 0.00008% / h after the interaction. For the third combination, its normalized weight is low, so the original data is weakened when interacting with other combinations, and redundant information is removed. After the interaction is completed, effective combinations with a matrix weight ≥ 0.5 are selected to form the first set of key weld anchor point performance characteristic combinations for the first interaction.

[0107] By traversing the feature similarity set for normalization, constructing a temporal interaction matrix, and using the matrix to weight the interaction of feature combinations, a coherent process was achieved, which accurately preserved the highly correlated performance characteristics of steel welds under deep peak shaving, weakened redundant information, and provided high-quality interactive data for subsequent acquisition of performance trend feature groups.

[0108] In summary, the method for evaluating the performance of steel welds under deep peak-shaving conditions provided in this application has the following technical effects:

[0109] This application obtains multiple typical operating conditions and the steel structure under test in the target application scenario. It then filters the key weld anchor points of the steel structure under test under multiple typical operating conditions by combining historical steel weld anomaly data sets. Next, it obtains the load protection method and duration, traverses multiple typical operating conditions to simulate a deep peak-shaving environment, maps the key weld anchor point set to a time-series sampling according to preset sampling indicators to obtain a set of test data sequences, determines the performance trend feature set through time-series interaction of performance trend features, extracts the failure test data set at the failure time from the test data sequence set, and combines the performance trend feature set to evaluate the steel weld performance to obtain evaluation results for multiple typical operating conditions. Finally, it performs a weighted calculation on the evaluation results of multiple typical operating conditions, thereby accurately evaluating the steel weld performance of the steel structure under test under deep peak-shaving conditions. This makes the steel weld performance evaluation results under deep peak-shaving scenarios more accurate and reliable, providing support for the safe and economical operation of the unit. It achieves the technical effect of making steel weld performance testing data more accurate and targeted, meeting the technical requirements for accurate evaluation of steel weld performance under deep peak-shaving conditions.

[0110] Example 2, as Figure 2 As shown, based on the same inventive concept as in Embodiment 1 above, this application provides a steel weld performance evaluation system under deep peak shaving conditions, the system comprising:

[0111] The critical weld anchor point acquisition module 1 is used to acquire multiple typical working conditions and the steel structure to be tested in the target application scenario, and to filter multiple critical weld anchor point sets of the steel structure to be tested under multiple typical working conditions in combination with historical steel weld anomaly data sets.

[0112] Test data sequence set acquisition module 2 is used to acquire the load protection method and the load protection duration, traverse the multiple typical working conditions to perform in-depth peak-shaving environment simulation, and perform time-series sampling of the multiple key weld anchor point sets according to preset sampling indicators to obtain multiple key weld anchor point test data sequence sets.

[0113] The performance trend feature set acquisition module 3 is used to perform time-series interaction of performance trend features on the multiple key weld anchor point test data sequence sets to determine the performance trend feature set of multiple key weld anchor points.

[0114] The steel weld performance evaluation result acquisition module 4 is used to extract the failure test data set of multiple key weld anchor points at the failure time of the multiple key weld anchor point test data sequence set, and combine the performance trend feature set of the multiple key weld anchor points to perform steel weld performance evaluation and obtain steel weld performance evaluation results under multiple typical working conditions.

[0115] The steel weld performance evaluation result determination module 5 is used to perform weighted calculations on the steel weld performance evaluation results of multiple typical working conditions to determine the steel weld performance evaluation result for the target application scenario.

[0116] Furthermore, the experimental data sequence set acquisition module 2 is used to perform the following steps:

[0117] The preset sampling indicators include time-strain curve, fracture elongation, fracture morphology, and stress-strain cycle curve.

[0118] Furthermore, the key weld anchor point acquisition module 1 is used to perform the following steps:

[0119] The historical steel weld anomaly data set is retrieved using the multiple typical working conditions and the steel structure under test to obtain multiple typical working condition historical steel weld anomaly data sets; a first typical working condition historical steel weld anomaly data set is extracted from the multiple typical working condition historical steel weld anomaly data sets, and the first typical working condition historical steel weld anomaly data set is retrieved by weld position to obtain a first typical working condition historical steel weld position set; the first typical working condition historical steel weld position set is filtered to determine a first critical weld anchor point set; the first critical weld anchor point set is added to multiple critical weld anchor point sets.

[0120] Furthermore, the key weld anchor point acquisition module 1 is used to perform the following steps:

[0121] Based on the first typical working condition historical steel weld location set, weld anomalies are identified in the steel structure under test to obtain identified steel structures under test; the welds corresponding to the maximum number of weld anomaly identifications are extracted from the identified steel structures under test as the first key weld anchor points; according to the preset first filtering bandwidth and the preset second filtering bandwidth, the first key weld anchor points are further filtered in the identified steel structures under test to obtain the first key weld anchor point set.

[0122] Furthermore, the key weld anchor point acquisition module 1 is used to perform the following steps:

[0123] According to the first and second filtering bandwidths, the first critical weld anchor points are derived to obtain first derived critical weld anchor points and second derived critical weld anchor points. It is determined whether the clustered neighborhood density of the first derived critical weld anchor points is greater than or equal to the clustered neighborhood density of the first critical weld anchor points. If so, the first derived critical weld anchor points are further derived and filtered according to the first filtering bandwidth to obtain a first set of derived filtered critical weld anchor points. It is then determined whether the clustered neighborhood density of the second derived critical weld anchor points is greater than or equal to the clustered neighborhood density of the first critical weld anchor points. If so, the second derived critical weld anchor points are further derived and filtered according to the second filtering bandwidth to obtain a second set of derived filtered critical weld anchor points. The union of the first set of derived filtered critical weld anchor points and the second set of derived filtered critical weld anchor points is obtained to obtain a first set of critical weld anchor points.

[0124] Furthermore, the key weld anchor point acquisition module 1 is used to perform the following steps:

[0125] When the clustered neighborhood density of the first derived critical weld anchor point is less than the clustered neighborhood density of the first critical weld anchor point, the difference between the clustered neighborhood density of the first derived critical weld anchor point and the clustered neighborhood density of the first critical weld anchor point is calculated. If the difference in clustered neighborhood density is less than or equal to a preset density difference threshold, the first derived critical weld anchor point is further derived and filtered according to the first filtering bandwidth to obtain the first derived filtered critical weld anchor point set.

[0126] Furthermore, the performance trend feature set acquisition module 3 is used to perform the following steps:

[0127] A multi-scale performance feature recognizer is invoked to analyze the test data sequence set of the multiple key weld anchor points to obtain a set of performance feature groups for multiple key weld anchor points; the performance trend feature time series interaction of the set of performance feature groups for multiple key weld anchor points is performed to determine the set of performance trend features for multiple key weld anchor points.

[0128] Furthermore, the performance trend feature set acquisition module 3 is used to perform the following steps:

[0129] A first key weld anchor point performance feature group is randomly extracted from multiple sets of key weld anchor point performance feature groups; the first key weld anchor point performance feature group is combined in pairs to obtain a first key weld anchor point performance feature combination set; feature similarity is identified for each combination in the first key weld anchor point performance feature combination set to obtain a first feature similarity set; based on the first feature similarity set, the first key weld anchor point performance feature combination set is subjected to temporal interaction to obtain a first interactive key weld anchor point performance feature combination set; based on the first key weld anchor point performance feature group, the mean of the first interactive key weld anchor point performance feature combination set is calculated to obtain a first interactive key weld anchor point performance trend feature group; the mean of the first interactive key weld anchor point performance trend feature group is calculated to obtain a first key weld anchor point performance trend feature, and the first key weld anchor point performance trend feature is added to multiple sets of key weld anchor point performance trend features.

[0130] Furthermore, the performance trend feature set acquisition module 3 is used to perform the following steps:

[0131] The first feature similarity set is traversed and normalized to construct a time-series interaction matrix set. The time-series interaction matrix set is then used to perform time-series interaction on the first key weld anchor point performance feature combination set to obtain the first interactive key weld anchor point performance feature combination set.

[0132] The steel weld performance evaluation system under deep peak shaving conditions provided in this embodiment of the invention can execute the steel weld performance evaluation method under deep peak shaving conditions provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method.

[0133] Although this application makes various references to certain modules in the system according to the embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy distinction between each other and are not used to limit the scope of protection of this invention.

[0134] The specific embodiments described above do not constitute a limitation on the scope of protection of this application. Those skilled in the art should understand that various modifications, combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application. In some cases, the actions or steps described in this application can be performed in a different order than that shown in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A method for evaluating the performance of a steel weld under deep peaking conditions, characterized in that, The method comprises: obtaining a plurality of typical working conditions of a target application scenario and a to-be-tested steel structure, and screening a plurality of key weld anchor point sets of the to-be-tested steel structure under the plurality of typical working conditions in combination with a historical steel weld abnormal data set; obtaining a load preservation mode and a preservation time length, traversing the plurality of typical working conditions to perform deep peak-shaving environment simulation, performing mapping time sequence sampling on the plurality of key weld anchor point set according to a preset sampling index, and obtaining a plurality of key weld anchor point test data sequence sets; performing performance trend feature time sequence interaction on the plurality of key weld anchor point test data sequence sets to determine a plurality of key weld anchor point performance trend feature sets; extracting a plurality of key weld anchor point failure test data sets of the plurality of key weld anchor point test data sequence sets at failure moments, respectively, combining the plurality of key weld anchor point performance trend feature sets to perform steel weld performance evaluation, and obtaining a plurality of typical working condition steel weld performance evaluation results; performing weighted calculation on the plurality of typical working condition steel weld performance evaluation results to determine a steel weld performance evaluation result of the target application scenario.

2. The method of evaluating steel weld performance under deep cycling surge conditions of claim 1, wherein, The preset sampling index comprises a time-strain curve, a fracture elongation, a fracture morphology, and a stress-strain cycle curve.

3. The method of evaluating steel weld performance under deep cycling surge conditions of claim 1, wherein, Obtaining a plurality of typical working conditions of a target application scenario and a to-be-tested steel structure, and screening a plurality of key weld anchor point sets of the to-be-tested steel structure under the plurality of typical working conditions in combination with a historical steel weld abnormal data set comprises: searching the historical steel weld abnormal data set by using the plurality of typical working conditions and the to-be-tested steel structure to obtain a plurality of typical working condition historical steel weld abnormal data sets; extracting a first typical working condition historical steel weld abnormal data set from the plurality of typical working condition historical steel weld abnormal data sets, and searching the first typical working condition historical steel weld abnormal data set by using a weld position to obtain a first typical working condition historical steel weld position set; screening the first typical working condition historical steel weld position set to determine a first key weld anchor point set; adding the first key weld anchor point set into the plurality of key weld anchor point sets.

4. The method of evaluating steel weld performance under deep cycling surge conditions of claim 3, wherein, Screening the first typical working condition historical steel weld position set to determine a first key weld anchor point set comprises: performing weld abnormality identification on the to-be-tested steel structure based on the first typical working condition historical steel weld position set to obtain an identified to-be-tested steel structure; extracting a weld corresponding to a maximum value of weld abnormality identification quantity from the identified to-be-tested steel structure as a first key weld anchor point; deriving and screening the first key weld anchor point in the identified to-be-tested steel structure according to a preset first screening bandwidth and a preset second screening bandwidth to obtain the first key weld anchor point set.

5. The method of evaluating steel weld performance under deep cycling surge conditions of claim 4, wherein, Deriving and screening the first key weld anchor point in the identified to-be-tested steel structure according to a preset first screening bandwidth and a preset second screening bandwidth to obtain the first key weld anchor point set comprises: deriving the first key weld anchor point and the second key weld anchor point respectively according to the first screening bandwidth and the second screening bandwidth to obtain a first derived key weld anchor point and a second derived key weld anchor point; determining whether the cluster neighborhood density of the first derived key weld anchor point is greater than or equal to the cluster neighborhood density of the first key weld anchor point, and if so, continuing to derive screen the first derived key weld anchor point according to the first screening bandwidth to obtain a first derived screened key weld anchor point set; determining whether the cluster neighborhood density of the second derived key weld anchor point is greater than or equal to the cluster neighborhood density of the first key weld anchor point, and if so, continuing to derive screen the second derived key weld anchor point according to the second screening bandwidth to obtain a second derived screened key weld anchor point set; performing a union operation on the first derived screened key weld anchor point set and the second derived screened key weld anchor point set to obtain the first key weld anchor point set.

6. The method of evaluating steel weld performance under deep cycling surge conditions of claim 5, wherein, When the cluster neighborhood density of the first derived key weld anchor point is less than the cluster neighborhood density of the first key weld anchor point, calculating the difference between the cluster neighborhood density of the first derived key weld anchor point and the cluster neighborhood density of the first key weld anchor point, and if the difference is less than or equal to a preset density difference threshold, continuing to derive screen the first derived key weld anchor point according to the first screening bandwidth to obtain a first derived screened key weld anchor point set.

7. The method of evaluating steel weld performance under deep cycling surge conditions of claim 1, wherein, performing performance trend feature time sequence interaction on the plurality of key weld anchor point test data sequence sets to determine a plurality of key weld anchor point performance trend feature sets, including: calling a multi-scale performance feature identifier to analyze the plurality of key weld anchor point test data sequence sets to obtain a plurality of key weld anchor point performance feature group sets; performing performance trend feature time sequence interaction on the plurality of key weld anchor point performance feature group sets to determine the plurality of key weld anchor point performance trend feature sets.

8. The method of evaluating steel weld performance under deep cycling surge conditions of claim 7, wherein, performing performance trend feature time sequence interaction on the plurality of key weld anchor point performance feature group sets to determine the plurality of key weld anchor point performance trend feature sets, including: randomly extracting a first key weld anchor point performance feature group from the plurality of key weld anchor point performance feature group sets; performing pairwise enumeration combination on the first key weld anchor point performance feature group to obtain a first key weld anchor point performance feature combination set; performing feature similarity recognition on each combination in the first key weld anchor point performance feature combination set to obtain a first feature similarity set; based on the first feature similarity set, performing time sequence interaction on the first key weld anchor point performance feature combination set to obtain a first interactive key weld anchor point performance feature combination set; based on the first key weld anchor point performance feature group, performing retrieval mean value calculation on the first interactive key weld anchor point performance feature combination set to obtain a first interactive key weld anchor point performance trend feature group; calculating the mean value of the first interactive key weld anchor point performance trend feature group to obtain a first key weld anchor point performance trend feature, and adding the first key weld anchor point performance trend feature to the plurality of key weld anchor point performance trend feature sets.

9. The method of evaluating steel weld performance under deep cycling surge conditions of claim 8, wherein, The first feature similarity set is normalized, a time sequence interaction matrix set is constructed, and the first key weld anchor performance feature combination set is time sequence interacted by using the time sequence interaction matrix set, to obtain a first interaction key weld anchor performance feature combination set.

10. A system for evaluating the performance of steel welds under deep cycling conditions, characterized in that, The system for implementing the steel weld performance evaluation method under deep peak regulation conditions according to any one of claims 1-9 comprises: A key weld anchor set acquisition module is configured to acquire a plurality of typical working conditions of a target application scenario and a steel structure to be tested, and screen a plurality of key weld anchor sets of the steel structure to be tested under the plurality of typical working conditions in combination with a historical steel weld abnormal data set. An experimental data sequence set acquisition module is configured to acquire a load preservation mode and a preservation duration, simulate a deep peak regulation environment by traversing the plurality of typical working conditions, and perform mapping time sequence sampling on the plurality of key weld anchor sets according to a preset sampling index, to obtain a plurality of key weld anchor experimental data sequence sets. A performance trend feature set acquisition module is configured to perform performance trend feature time sequence interaction on the plurality of key weld anchor experimental data sequence sets, to determine a plurality of key weld anchor performance trend feature sets. A steel weld performance evaluation result acquisition module is configured to extract a plurality of key weld anchor failure experimental data sets at a failure time of the plurality of key weld anchor experimental data sequence sets, perform steel weld performance evaluation in combination with the plurality of key weld anchor performance trend feature sets, and obtain a plurality of typical working condition steel weld performance evaluation results. A steel weld performance evaluation result determination module is configured to perform weighted calculation on the plurality of typical working condition steel weld performance evaluation results, to determine a steel weld performance evaluation result of the target application scenario.

Images