A full-process digital management method and system for wind power equipment manufacturing

Abstract

The present application relates to the technical field of wind power equipment management, in particular to a full-process digital management method and system for wind power equipment manufacturing. The method determines a basic temperature risk value according to an abnormal temperature point; determines a comprehensive temperature risk value according to the fluctuation characteristics of the normal temperature point within the temperature allowable interval, the difference degree of the temperature time sequence close to the boundary of the temperature allowable interval and the basic temperature risk value; determines a risk deviation coefficient according to the similarity between the temperature time sequence and the historical temperature sub-sequence; and determines a calibrated temperature risk value in combination with the comprehensive temperature risk value; extracts an abnormal excess height point exceeding the excess height allowable interval in the excess height time sequence; determines a comprehensive temperature-excess correlation index according to the abnormal excess height point and the temperature value at the corresponding moment in the temperature time sequence; and determines a welding severity index according to the comprehensive temperature-excess correlation index and the calibrated temperature risk value, thereby making the risk assessment result more credible and adaptive.

Description

Technical Field

[0001] This invention relates to the field of wind power equipment management technology, specifically to a digital management method and system for the entire process of wind power equipment manufacturing. Background Technology

[0002] As the core load-bearing structure of wind power equipment, the welding quality of the wind turbine tower directly affects the safe operating life of the entire wind turbine unit. The manufacturing of the tower follows a complete chain of standards, including material control, forming and processing, welding and assembly, anti-corrosion treatment, and inspection and acceptance. Among these standards, welding and assembly is the key link connecting all stages and determining the final structural strength.

[0003] Currently, in the digital management process of wind turbine tower manufacturing, routine collection and monitoring of process parameters such as welding temperature, current, and voltage have been generally achieved. However, existing technologies typically monitor and alarm thresholds independently for parameters such as welding temperature and weld formation (e.g., weld height). This method can only be used to detect obvious anomalies in a single parameter and cannot deeply reveal the intrinsic causal relationship between abnormal welding process parameters and final weld quality defects. When intermittent or complex anomalies occur, the early warning is severely delayed, missing the best management and adjustment window. Summary of the Invention

[0004] To address the technical problem of insufficient insight into the intrinsic causal relationship between abnormal welding process parameters and final weld quality defects, leading to severely delayed early warnings, this invention provides a full-process digital management method and system for wind power equipment manufacturing. The specific technical solution adopted is as follows:

[0005] This invention proposes a full-process digital management method for wind power equipment manufacturing, the method comprising:

[0006] Simultaneously acquire the welding temperature time sequence and weld reinforcement time sequence during the welding process of wind turbine towers, and obtain historical temperature time sequence under the same welding conditions;

[0007] Extract abnormal temperature points that exceed the allowable temperature range and normal temperature points that do not exceed it from the temperature time series; determine the basic temperature risk value based on the abnormal temperature points; determine the comprehensive temperature risk value based on the fluctuation characteristics of normal temperature points within the allowable temperature range, the degree of difference between the temperature time series and the boundary of the allowable temperature range, and the basic temperature risk value.

[0008] Extract historical temperature subsequences from historical temperature time series that have the same duration as the temperature time series and meet preset similarity conditions; determine the risk deviation coefficient based on the similarity between the temperature time series and the historical temperature subsequences; and determine the calibration temperature risk value by combining the comprehensive temperature risk value.

[0009] Extract abnormal residual height points that exceed the allowable residual height range from the residual height time series; determine the comprehensive temperature residual height correlation index based on the abnormal residual height points and the temperature values ​​at the corresponding times in the temperature time series; determine the welding severity index based on the comprehensive temperature residual height correlation index and the calibration temperature risk value.

[0010] Based on the comparison between the welding severity index and the preset threshold, the corresponding manufacturing process management strategy is implemented.

[0011] Furthermore, the welding temperature timing sequence is a temperature timing sequence during the welding steady-state stage; the method further includes:

[0012] The welding temperature value is collected in real time. When the welding temperature value enters the preset process temperature range for the first time, and the temperature value is within the preset process temperature range for a preset number of consecutive sampling times after the first entry, the welding process is determined to have entered the welding steady state stage.

[0013] Furthermore, the process for determining the baseline temperature risk value includes:

[0014] The total number of temperature points in the welding temperature time sequence, the number of abnormal temperature points exceeding the preset temperature allowable range, and the temperature width of the preset temperature allowable range are statistically analyzed.

[0015] Calculate the temperature difference between each abnormal temperature point and the boundary of the allowable temperature range; calculate the arithmetic mean of all temperature differences as the average exceedance.

[0016] Divide the average exceedance by the temperature width to obtain the average exceedance ratio; divide the number of abnormal temperature points by the total number of temperature points to obtain the percentage of abnormal temperature points.

[0017] Calculate the product of the percentage of abnormal temperature points and the average excess percentage as a risk factor;

[0018] Add the risk factor to the positive integer 1 to obtain the base temperature risk value.

[0019] Furthermore, the process for determining the comprehensive temperature risk value includes:

[0020] Calculate the highest and lowest temperature values ​​among all normal temperature points;

[0021] The continuous normal temperature points are divided into at least one continuous normal segment; for each continuous normal segment, the segment fluctuation intensity is calculated based on the fluctuation characteristics of the temperature value within the continuous normal segment; the arithmetic mean of the fluctuation intensity of all segments is divided by a preset fluctuation intensity reference value to obtain the volatility risk component.

[0022] The difference between the upper limit of the allowable temperature range and the highest temperature value is calculated as the upper boundary temperature value; the difference between the lowest temperature value and the lower limit of the allowable range is calculated as the lower boundary temperature value; based on the upper boundary temperature value, the lower boundary temperature value, and the preset normal temperature value, the boundary risk component is calculated.

[0023] Calculate the sum of the volatility risk component and the boundary risk component, add the summation result to the positive integer 1, and obtain the state risk coefficient;

[0024] The product of the state risk coefficient and the basic temperature risk value is used as the comprehensive temperature risk value.

[0025] Further, the step of extracting historical temperature subsequences from the historical temperature time series that have the same duration as the temperature time series and meet preset similarity conditions includes:

[0026] The temperature time series of the current welding steady-state stage is used as the reference series;

[0027] For each historical temperature time series, multiple candidate subsequences with the same length as the baseline sequence are sequentially extracted from the historical temperature time series using a sliding window method. For each candidate subsequence, the Pearson correlation coefficient between the candidate subsequence and the baseline sequence is calculated as a morphological similarity score. From all candidate subsequences, the candidate subsequence with the highest morphological similarity score is selected as the historical temperature subsequence.

[0028] Furthermore, the process for determining the risk deviation coefficient includes:

[0029] Obtain the preset reference deviation;

[0030] For each historical temperature subsequence, the Pearson correlation coefficient between the historical temperature subsequence and the temperature time series is normalized and used as a similarity value; the arithmetic mean of all similarity values ​​is calculated as the average similarity.

[0031] Calculate the difference between the positive integer 1 and the average similarity, and use it as the average deviation.

[0032] Divide the average deviation by the preset reference deviation to obtain the relative deviation.

[0033] Calculate the sum of the positive integer 1 and the relative deviation, and use it as the risk deviation coefficient.

[0034] Furthermore, the process for determining the comprehensive residual temperature correlation index includes:

[0035] In the residual height time series, consecutive abnormal residual height points without intervals are grouped together, and each group is defined as an abnormal unit. If an abnormal unit contains more than or equal to two abnormal residual height points, it is marked as a continuous out-of-specification segment. If an abnormal unit contains only one abnormal residual height point, it is marked as an isolated out-of-specification point.

[0036] For continuous exceedance segments and isolated exceedance points, different association analysis rules are used to calculate the association contribution value;

[0037] The comprehensive residual temperature correlation index is obtained by weighting all related contribution values ​​with the duration of each exceeding segment and exceeding point as the weight. The weight of each continuous exceeding segment is the number of abnormal residual high points it contains, and the weight of each isolated exceeding point is 1.

[0038] Furthermore, the calculation of the correlation contribution value for consecutive exceeding segments using correlation analysis rules includes:

[0039] For each continuous excess segment, the excess height difference between the excess height value and the boundary of the allowable excess height interval at each moment within the continuous excess segment is calculated as the excess height amount at each moment within the continuous excess segment, and sorted by time to form an excess height amount sequence; the welding temperature sequence at the corresponding moment is extracted from the temperature time series, and the first comparison temperature value of the previous moment of the welding temperature sequence is extracted.

[0040] The Pearson correlation coefficient between the superscalar superscalar sequence and the welding temperature sequence was calculated and normalized to serve as the basic correlation coefficient.

[0041] Calculate the first difference between the temperature value at each moment in the welding temperature sequence and the first comparison temperature value; for each first difference, if the first difference is greater than 0, retain the first difference; if the first difference is less than or equal to 0, correct the first difference to 0.

[0042] Calculate the arithmetic mean of all corrected first differences as the average positive overtemperature; divide the average positive overtemperature by the temperature width of the preset temperature allowable range to obtain the segment temperature penalty term;

[0043] Multiply the sum of the positive integer 1 and the segment temperature penalty term by the basic correlation degree to obtain the correlation contribution value of the continuous out-of-standard segment.

[0044] Furthermore, the calculation of the correlation contribution value for isolated exceedance points using correlation analysis rules includes:

[0045] For each isolated exceedance point, calculate the residual exceedance amount of the isolated exceedance point; and obtain the temperature value at the corresponding time of the isolated exceedance point, as well as the second comparison temperature value at the time before the isolated exceedance point;

[0046] Divide the excess height by the width of the allowable excess height range to obtain the excess height ratio;

[0047] Calculate the second difference between the temperature value at the time corresponding to the isolated exceeding point and the second comparison temperature value. If the second difference is negative or zero, the second difference is counted as 0.

[0048] Divide the second difference by the temperature width to obtain the point temperature shock term;

[0049] Multiply the sum of the positive integer 1 and the point temperature shock term by the excess height ratio to obtain the correlation contribution value of isolated excess points.

[0050] A full-process digital management system for wind power equipment manufacturing, the system comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of a full-process digital management method for wind power equipment manufacturing.

[0051] The present invention has the following beneficial effects:

[0052] This invention simultaneously analyzes the welding temperature time series and the weld reinforcement time series, and establishes a quantitative correlation model between the two. Specifically, by comprehensively analyzing the temperature-reinforcement correlation index, it retrospectively analyzes whether abnormal weld reinforcement points occurring in the current welding process (involving quality issues in wind turbine towers) are highly likely to be directly caused by synchronously occurring welding temperature anomalies. This allows the system to not only detect anomalies in a timely manner but also determine whether the anomalies may have caused actual quality problems, improving the accuracy of root cause analysis. By analyzing the deviation of temperature from the allowable temperature range, a basic temperature risk value is determined to capture explicit anomalies. Then, by analyzing the fluctuation characteristics of normal temperature points and the basic temperature risk value, a comprehensive temperature risk value is determined to facilitate the early detection of potential process instability risks. Finally, by calibrating the temperature risk value, the current temperature time series is analyzed and compared with historical temperature time series (i.e., temperature time series in historical welding processes where no process risks occurred), allowing for a deeper identification of hidden anomalies deviating from normal process experience. This dynamic analysis of the probability of risk occurrence from multiple perspectives makes the risk assessment results more credible and adaptable. Attached Figure Description

[0053] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the 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.

[0054] Figure 1A flowchart of a full-process digital management method for wind power equipment manufacturing, provided as an embodiment of the present invention;

[0055] Figure 2 An example diagram illustrating the comprehensive temperature risk determination process provided in one embodiment of the present invention;

[0056] Figure 3 This is an example diagram illustrating the process of determining the comprehensive residual temperature correlation index provided in one embodiment of the present invention. Detailed Implementation

[0057] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a full-process digital management method and system for wind power equipment manufacturing proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0058] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0059] The following description, in conjunction with the accompanying drawings, details a specific solution for a fully digital management method and system for wind power equipment manufacturing provided by this invention.

[0060] Please see Figure 1 The diagram illustrates a flowchart of a full-process digital management method for wind power equipment manufacturing, provided by an embodiment of the present invention. The method includes:

[0061] S101: Synchronously acquire the welding temperature time sequence and weld reinforcement time sequence during the welding process of wind turbine towers, and obtain historical temperature time sequence under the same welding conditions.

[0062] It is important to understand that welding temperature is the most direct and critical process parameter reflecting the welding heat input, and weld reinforcement is the core geometric indicator for evaluating weld formation quality. Moreover, during the welding process, welding temperature directly determines the melting, flow, and solidification behavior of the base material and welding wire, and is the fundamental driving force for weld formation (including reinforcement). There is a physical causal relationship between the two. Therefore, this invention synchronously collects the welding temperature time series and the weld reinforcement time series to facilitate subsequent in-depth analysis of the temporal correlation between "temperature anomaly at a certain moment" and "weld formation defects caused by temperature anomaly".

[0063] To achieve continuous, distributed monitoring of the temperature along the entire length of the weld, a distributed fiber optic temperature sensor is preferred. The specific deployment method of the distributed fiber optic temperature sensor is determined based on the actual engineering scenario, and will not be detailed in this embodiment. For example, in a specific implementation, the sensing fiber is laid tightly along a predetermined path along the weld of the tower to be welded. Based on the principle of optical time-domain reflectometry, the distributed fiber optic temperature sensor can calculate the temperature at every point along the fiber length (i.e., the weld direction) in real time, forming a spatial temperature distribution curve. By scanning at a fixed frequency (e.g., 2Hz), the welding temperature time sequence changing over time can be obtained.

[0064] To achieve non-contact, online measurement of weld reinforcement height, a laser displacement sensor or a line laser profile sensor is preferred. The laser displacement sensor or line laser profile sensor is usually mounted on the welding head or a separate, servo-driven measuring arm. This ensures that the sensor's laser scanning line is perpendicular to the weld and covers the solidified weld area behind the molten pool. The sensor then measures the weld cross-sectional profile in real time and extracts the reinforcement height value of the weld centerline (i.e., the height difference between the highest point of the weld surface and the base material surface) through a built-in algorithm. The laser displacement sensor or line laser profile sensor operates at the same or a matching sampling frequency (e.g., 2Hz) as the temperature sensor to output a weld reinforcement height time sequence.

[0065] It should be noted that a unified clock source or a high-precision synchronization timestamp can be used to ensure that the welding temperature timing sequence and the weld height timing sequence are strictly aligned at each sampling moment, generating a one-to-one corresponding data pair.

[0066] The same welding conditions refer to completed welding processes that are the same or highly similar to the tower to be analyzed in terms of material grade, plate thickness, bevel type, welding process, and main welding parameters.

[0067] It should be noted that the historical temperature time series is a complete temperature time series of welding processes that were collected and stored using the same temperature sensors in past production and subsequently verified as qualified by non-destructive testing.

[0068] S102: Extract abnormal temperature points that exceed the allowable temperature range and normal temperature points that do not exceed the allowable temperature range from the temperature time series; determine the basic temperature risk value based on the abnormal temperature points; determine the comprehensive temperature risk value based on the fluctuation characteristics of normal temperature points within the allowable temperature range, the degree of difference between the temperature time series and the boundary of the allowable temperature range, and the basic temperature risk value.

[0069] It is important to understand that simply counting abnormal temperature points is far from sufficient to reflect the true stability of the welding process. It is also necessary to analyze the severity of the risks posed by abnormal temperature points and to delve into the potential risk trends of normal temperature points in order to avoid the risk signals being diluted by averaging all temperature points together.

[0070] A temperature time series is a set of temperature data continuously measured and recorded by temperature sensors during the welding process, arranged in chronological order. It can be represented by an array, where each element represents the welding temperature value measured at a specific sampling moment. Each data element in the temperature time series can be called a temperature point, representing the temperature measurement value at a certain sampling moment.

[0071] It should be noted that the specific value of the allowable temperature range can be determined based on the material, thickness, welding method, and welding material type of the workpiece being welded. The allowable temperature range is not fixed and is not specifically limited in this embodiment. For example, if the ideal molten pool temperature for welding process evaluation is set at around 1150°C, considering the fluctuation of the welding power source, the slight changes in heat conduction, and sensor measurement noise, a reasonable fluctuation range is usually allowed in actual process scenarios. For example, for welding medium and thick plates, a typical allowable fluctuation range may be set to fluctuate by 50°C. Then the allowable temperature range can be set as: [1150-50, 1150+50] = [1100°C, 1200°C], where 1100°C is the lower limit of the allowable temperature range and 1200°C is the upper limit of the allowable temperature range. Both the lower and upper limits are the boundaries of the allowable temperature range.

[0072] It should be noted that the welding temperature time sequence is the temperature time sequence during the steady-state stage of welding.

[0073] It is important to understand that there is an unstable transition period of "arc ignition-heating" at the beginning of welding. During this period, the temperature changes drastically, and the data is not representative. Moreover, all subsequent risk assessment algorithms (such as fluctuation analysis and boundary proximity analysis) are based on the assumption that the data is stable and representative of the process. In the unstable heating phase, the drastic inherent fluctuations will be misjudged as high risk by the algorithm, resulting in false alarms. Therefore, the analysis of temperature time series is strictly limited to the steady-state stage of intelligent judgment.

[0074] Accordingly, the weld reinforcement height timing sequence referred to in this invention, in an ideal and preferred embodiment, also specifically refers to the reinforcement height data sequence that is strictly synchronized with the welding steady-state stage in time. Since the temperature and reinforcement height sensors are triggered synchronously by a unified clock during the data acquisition process, once the system determines that it has entered the welding steady-state stage based on the temperature value, it will automatically use the starting point of the time as a marker and ensure that all temperature and reinforcement height data collected thereafter are included in the corresponding timing sequence.

[0075] In this embodiment, welding temperature values ​​are collected in real time. When the welding temperature value is detected to enter the preset process temperature range for the first time, and the temperature value is within the preset process temperature range for a preset number of consecutive sampling times after the first entry, it is determined that the welding process has entered the welding steady state stage.

[0076] It should be noted that the specific value of the preset process temperature range is set around the process target temperature. This embodiment does not make specific limitations. For example, assuming that the process target temperature for submerged arc welding of a certain Q355D steel is 1200°C, to ensure that the process stably enters the ideal state, a relatively strict criterion is set, such as a fluctuation of 25°C, then the preset process temperature range is [1175°C, 1225°C].

[0077] It should be noted that the specific value of the preset quantity can be set according to different welding methods, material thicknesses, and long-term accumulated experience in process stability. This embodiment does not impose specific limitations. For example, for submerged arc welding with relatively stable arcs, the preset quantity can be set to 5-8; for certain gas shielded welding where the arc is easily disturbed, the preset quantity can be appropriately increased to 8-12.

[0078] It should be noted that after determining that the welding process has entered the steady-state stage, only the temperature data of this stage will be collected and analyzed to form a welding temperature time series.

[0079] It is important to understand that in welding quality control, parameter exceeding the standard is the most irrefutable evidence of anomalies. While a single or a few minor exceedances may be caused by accidental interference (such as momentary arc instability), frequent or significant exceedances strongly indicate systemic loss of control, such as improper process parameter settings, equipment failure, or operational inaccuracies. Therefore, by analyzing the proportion of abnormal temperature points and the severity of data exceeding the standard, this vague judgment of "abnormality" can be transformed into a measurable and comparable value, namely the basic temperature risk value.

[0080] In this embodiment, the total number of temperature points in the welding temperature time sequence, the number of abnormal temperature points exceeding the preset allowable temperature range, and the temperature width of the preset allowable temperature range are counted. The temperature difference between the temperature value of each abnormal temperature point and the boundary of the allowable temperature range is calculated. The arithmetic mean of all temperature differences is calculated as the average exceedance. The average exceedance is divided by the temperature width to obtain the average exceedance ratio. The number of abnormal temperature points is divided by the total number of temperature points to obtain the percentage of abnormal temperature points. The product of the percentage of abnormal temperature points and the average exceedance ratio is calculated as the risk factor. The risk factor is added to the positive integer 1 to obtain the basic temperature risk value.

[0081] It can be understood that the temperature width of the preset allowable temperature range = the upper limit of the allowable temperature range - the lower limit of the allowable temperature range; if the abnormal temperature point is higher than the upper limit of the allowable temperature range, then the temperature difference = the temperature value of the abnormal temperature point - the upper limit of the allowable temperature range; if the abnormal temperature point is lower than the lower limit of the allowable temperature range, then the temperature difference = the lower limit of the allowable temperature range - the temperature value of the abnormal temperature point.

[0082] The average exceedance quantifies the average severity of temperature exceedance events.

[0083] The average exceedance ratio is an indicator obtained by normalizing the dimensions based on the average exceedance range. In other words, it reflects the average severity of the temperature exceedance range as a percentage of the entire allowable temperature range by dividing the average exceedance range by the temperature width.

[0084] The percentage of abnormal temperature points quantifies the proportion of time during the analysis period when the temperature was out of control.

[0085] It is important to understand that the number of abnormal points cannot exceed the total number of temperature points. If the proportion of a certain abnormal temperature point is larger, it means that the time when the temperature is out of control is higher. If the average excess proportion is larger, it reflects that all abnormal temperature points are more likely to have serious exceedances and the degree of exceedance is greater. In this case, the risk factor will also be greater. The greater the risk factor, the higher the risk, and the welding process will continue to be seriously out of control.

[0086] It should be noted that the reason for adding the risk factor to the positive integer 1 is as follows: if no temperature point exceeds the standard in the entire temperature time series, then the risk factor is 0. If 1 is not added, the final basic risk value will be defined as 0. In the subsequent calculation of the comprehensive temperature risk value, as long as the historical performance is perfect (risk factor is 0), no matter how great the potential risk is at the current normal temperature point, the comprehensive risk will be forcibly reduced to zero. This is obviously a dangerous logical loophole, which will cause it to completely fail to work for abnormal situations that were good in the past but have suddenly changed.

[0087] It is important to understand that even if all temperature points are within acceptable limits, the basic risk value is closer to 1. However, if the normal temperature points fluctuate drastically within the allowable range, or operate close to the boundary of the allowable temperature range, this indicates that the welding process control is extremely unstable or operating at the safety limit. This state is like "walking a tightrope"—although not exceeding the limit, the potential risk is extremely high, and failure may occur at any time due to minor disturbances. Therefore, by analyzing the "fluctuation characteristics" and "closeness to the boundary" of the normal temperature points, the state risk coefficient is determined. Then, this potential instability and vulnerability is quantified into a risk increment by combining the state risk coefficient.

[0088] The process of determining comprehensive temperature risk is as follows: Figure 2 As shown, it includes:

[0089] S102-1: Calculate the highest and lowest temperature values ​​among all normal temperature points.

[0090] S102-2: Divide the continuous normal temperature points into at least one continuous normal segment; for each continuous normal segment, calculate the segment fluctuation intensity based on the fluctuation characteristics of the temperature value within the continuous normal segment; divide the arithmetic mean of the fluctuation intensity of all segments by the preset fluctuation intensity reference value to obtain the volatility risk component.

[0091] It should be noted that the sampling time sequence traverses the welding temperature time series. When a data point within the preset temperature allowable range (i.e., a normal temperature point) is encountered, it is marked and included in the currently constructed segment. This process continues, and as long as subsequent consecutive points are also normal temperature points, they are grouped into the same segment. Once an abnormal temperature point is encountered, the current normal segment ends, the segment is saved, and it waits for the next normal temperature point to appear before starting the construction of a new segment. After the traversal is completed, all saved sequences consisting of consecutive normal points constitute the divided consecutive normal segments. If there are no abnormal points during the entire analysis period, all normal points constitute a complete consecutive normal segment.

[0092] To accurately determine the intensity of a segment fluctuation, as an example, the absolute difference in temperature values ​​between adjacent moments within a continuous normal segment is calculated as the instantaneous temperature change; the arithmetic mean of all instantaneous temperature changes within a continuous normal segment is calculated as the segment fluctuation intensity.

[0093] The intensity of the segment fluctuation quantifies the degree of fluctuation or shaking of the welding temperature value around its own trend during the duration of a continuous normal segment. The greater the intensity of the segment fluctuation of a continuous normal segment, the more unstable the temperature is within that continuous normal segment.

[0094] For example, suppose there are three temperature values ​​at three sampling times within a continuous normal segment, which are W1, W2 and W3 in chronological order. Then W1 and W2 are temperature values ​​at adjacent times, and W2 and W3 are temperature values ​​at adjacent times. The instantaneous temperature changes are: |W1-W2| and |W2-W3|.

[0095] It should be noted that the specific value of the preset fluctuation intensity reference value can be calculated based on the temperature time sequence of historical qualified welding processes. For example, assuming that for a submerged arc welding process of a tower with a specific material and thickness, through long-term data accumulation, the average value of the typical temperature fluctuation within the allowable range during the steady-state stage of welding (calculated using the aforementioned segment fluctuation intensity method) is about 3.5°C, then 3.5°C can be set as the preset fluctuation intensity reference value.

[0096] S102-3: Calculate the difference between the upper limit of the allowable temperature range and the highest temperature value as the upper boundary temperature value; calculate the difference between the lowest temperature value and the lower limit of the allowable range as the lower boundary temperature value; calculate the boundary risk component based on the upper boundary temperature value, the lower boundary temperature value, and the preset normal temperature value.

[0097] To accurately obtain the boundary risk components, as an example, the arithmetic mean of the upper and lower boundary temperature values ​​is calculated as the temperature average; the preset normal temperature value is divided by the sum of the temperature average and the smallest positive number to obtain the boundary risk components.

[0098] It should be noted that the specific value of the preset normal temperature can be derived from the statistical data of historical qualified welding processes. This embodiment does not impose a specific limitation. For example, assuming that for a specific process, by analyzing historical qualified data, it is found that the maximum value of the compliant temperature point is usually about 40°C away from the upper limit of the temperature allowable range, and the minimum value is about 30°C away from the lower limit, then the typical average value of the distance between the upper and lower boundaries, i.e. (40+30)÷2=35°C, can be used to set the preset normal temperature value to 35°C.

[0099] It is important to understand that the average temperature represents the average safe distance between the normal temperature and the boundary of the allowable range. The larger the average temperature, the more "away" the normal data is from the boundary, and the more sufficient the safety margin. Conversely, the smaller the average temperature, the more "crowded" the compliant data is near the boundary, the tighter the safety margin, and the greater the possibility of potential boundary risks.

[0100] It should be noted that if the maximum value of all normal temperature data is exactly at the upper limit and the minimum value is exactly at the lower limit, the average temperature will be zero. All normal temperature data will be "compressed" to the two boundary lines of the allowable range, and the safety margin will be completely exhausted. In this case, a very small positive number (such as 0.1°C) should be preset as the calculation protection value.

[0101] S102-4: Calculate the sum of the volatility risk component and the boundary risk component, add the summation result to the positive integer 1, and obtain the state risk coefficient.

[0102] It is important to understand that the volatility risk component measures the degree of deviation of the normal temperature value from a recognized and stable benchmark level (i.e., the preset volatility intensity reference value) within the allowable temperature range. The larger a certain volatility risk component is, the more drastic the fluctuation of the normal temperature value is compared with the typical level, and the greater the potential risk. Furthermore, if the boundary risk component is larger, it reflects that the safety margin is closer to the boundary and is in the highest risk state. In this case, the state risk coefficient should be larger, which will have a higher amplification effect on the basic risk value.

[0103] It should be noted that the reason for adding the summation result to the positive integer 1 is that the positive integer 1 here is the absolute lower limit of the state risk coefficient, which represents the ideal state when both the volatility risk component and the boundary risk component are zero, that is, no volatility or boundary proximity risk can be ignored.

[0104] S102-5: Calculate the product of the state risk coefficient and the basic temperature risk value as the comprehensive temperature risk value.

[0105] S103: Extract historical temperature subsequences from the historical temperature time series that have the same duration as the temperature time series and meet the preset similarity conditions; determine the risk deviation coefficient based on the similarity between the temperature time series and the historical temperature subsequences; and determine the calibration temperature risk value by combining the comprehensive temperature risk value.

[0106] It's important to understand that since historical temperature time series originate from past independent welding processes, their absolute starting points are completely different from the current temperature time series. Therefore, to accurately extract historical temperature subsequences, a "logical alignment" is required to ensure that the data segments are of equal size. This "logical alignment" is not a simple alignment based on calendar time. Instead, it first processes the historical temperature time series in the same way as the current temperature time series, identifying the steady-state welding stage. All subsequent "logical alignments" are performed within the steady-state stage data of each historical temperature time series. Then, through the core algorithm of consistent duration + sliding window + similarity analysis, the algorithm essentially automatically finds the optimal logical alignment point by sliding across the historical temperature time series to find the starting point from which the extracted subsequence is most similar in shape to the current temperature time series.

[0107] In this embodiment, the temperature time series of the current welding steady-state stage is used as the reference sequence. For each historical temperature time series, multiple candidate subsequences with the same length as the reference sequence are sequentially extracted from the historical temperature time series using a sliding window. For each candidate subsequence, the Pearson correlation coefficient between the candidate subsequence and the reference sequence is calculated as a morphological similarity score. From all candidate subsequences, the candidate subsequence with the highest morphological similarity score is selected as the historical temperature subsequence.

[0108] It is understood that the specific calculation method of the Pearson correlation coefficient is a technical means well known to those skilled in the art, and will not be described in detail in this embodiment.

[0109] It is important to understand that even if the current temperature data does not significantly exceed the standard or fluctuate, if the temperature change pattern deviates significantly from the stable pattern shared by many successful welding cases in history, this is itself a strong abnormal signal. Therefore, the degree of deviation of this pattern can be quantified by the risk deviation coefficient. The greater the degree of deviation, the higher the risk deviation coefficient. In subsequent calculations, the risk deviation coefficient will act as a risk amplifier and, combined with the comprehensive temperature risk value, obtain the calibration temperature risk value. This means that even if the surface risk of a current welding case is not high, but the degree of deviation from the process pattern of historical successful welding cases is high, the final risk assessment result will be significantly improved, thereby forcing the system to be highly vigilant.

[0110] In this embodiment, a preset reference deviation is obtained; for each historical temperature subsequence, the Pearson correlation coefficient between the historical temperature subsequence and the temperature time series is normalized and used as a similarity value; the arithmetic mean of all similarity values ​​is calculated as the average similarity; the difference between the positive integer 1 and the average similarity is calculated as the average deviation; the average deviation is divided by the preset reference deviation to obtain the relative deviation; the sum of the positive integer 1 and the relative deviation is calculated as the risk deviation coefficient.

[0111] It should be noted that the specific value of the preset reference deviation is derived from the internal cross-comparison and statistical analysis of historical qualified welding process data. This embodiment does not impose a specific limitation. For example, suppose we analyze 50 historical qualified temperature sequences for a specific process. Using the same method as this invention, we calculate the average deviation between each pair of sequences (i.e., for each pair of sequences, we calculate 1 - average similarity). We then statistically analyze the average deviation calculated for all paired sequences and find that the arithmetic mean of all average deviations is approximately 0.15. Therefore, 0.15 can be set as the preset reference deviation.

[0112] It should be noted that the Pearson correlation coefficient has a range of [-1, 1]. To convert the Pearson correlation coefficient into a more intuitive similarity value with a range of [0, 1], linear normalization is required. The specific method is as follows: add 1 to the Pearson correlation coefficient to shift the range to [0, 2], and then divide by 2. That is, similarity value = (Pearson correlation coefficient + 1) ÷ 2. After this transformation, a Pearson correlation coefficient of 1 (perfectly positive correlation) corresponds to a similarity value of 1 (perfectly similar); a Pearson correlation coefficient of -1 (perfectly negative correlation) corresponds to a similarity value of 0 (completely dissimilar); and a Pearson correlation coefficient of 0 (no linear correlation) corresponds to a similarity value of 0.5.

[0113] Average similarity reflects the average level of similarity between the temperature variation pattern of the current welding process and the temperature patterns of a set of historical successful welding processes (i.e., historical temperature subsequences).

[0114] It is important to understand that if the average deviation between a certain temperature time series and a certain historical temperature subsequence is greater, it means that the temperature pattern of the current welding process deviates more severely from the historical normal pattern; if a certain relative deviation is greater, it means that the temperature pattern of the current welding process deviates more than the historical normal range, that is, the possibility of risk in the current welding process is greater, and the risk deviation coefficient is also greater.

[0115] It should be noted that the positive integer 1 plays a role in calculating the risk deviation coefficient. Since the relative deviation is always non-negative, the risk deviation coefficient is always greater than or equal to 1. This ensures that when the risk deviation coefficient is multiplied by the comprehensive temperature risk value, it will always be a risk maintenance or amplification factor, and will never become a risk reduction factor (i.e., the risk deviation coefficient will never be less than 1). This ensures that any deviation from the historical successful pattern will not be considered to reduce the overall risk, but will only maintain or increase the risk. Moreover, when the relative deviation is 0 (i.e., the current pattern is completely consistent with the historical pattern), the risk deviation coefficient = 1, which means that the existing temperature risk will not be amplified.

[0116] In this embodiment, the product of the risk deviation coefficient and the comprehensive temperature risk value is calculated as the calibration temperature risk value.

[0117] S104: Extract abnormal residual height points that exceed the allowable residual height range in the residual height time series; determine the comprehensive temperature residual height correlation index based on the abnormal residual height points and the temperature values ​​at the corresponding times in the temperature time series; determine the welding severity index based on the comprehensive temperature residual height correlation index and the calibration temperature risk value.

[0118] The weld reinforcement time series is a set of weld reinforcement data arranged in chronological order, continuously measured and recorded by temperature sensors during the welding process. It can be represented by an array, where each element represents the weld reinforcement value measured at a specific sampling time. Each data element in the weld reinforcement time series can be called a weld reinforcement point, representing the weld reinforcement measurement value at a certain sampling time.

[0119] It should be noted that the specific value of the allowable range of weld reinforcement can be determined comprehensively based on national / industry welding technical standards (such as GB, ISO, AWS relevant specifications), process requirements of product design drawings, and welding process qualification test results for specific materials and joint types. The allowable range of weld reinforcement is not fixed and this embodiment does not impose specific limitations. For example, for the welding process of wind turbine tower steel plates of a certain thickness, the qualified range of weld reinforcement is usually set to 0mm to 3mm after process qualification. Welds below 0mm (i.e., depression) or above 3mm are considered abnormal and need to be monitored and evaluated. Then the allowable range of weld reinforcement can be set as: [0mm, 3mm].

[0120] It is important to understand that although abnormal welding temperature (process cause) and excessive weld reinforcement (quality effect) are physically strongly correlated, they are not simply synchronous in terms of data. Therefore, in order to quantify the extent to which "observed weld defects can be attributed to the simultaneous welding temperature abnormality," we can differentiate between persistent and transient abnormalities. These two isolated alarms, temperature abnormality and excessive weld reinforcement, are woven into a quantifiable "causal correlation strength" index, namely the comprehensive temperature-reinforcement correlation index. The higher the value of the comprehensive temperature-reinforcement correlation index, the more likely that temperature runaway is the main culprit for the current quality problem; the lower the comprehensive temperature-reinforcement correlation index, the more it suggests that other causes (such as wire feeding, gas, etc.) need to be investigated.

[0121] The process of determining the comprehensive residual temperature correlation index is as follows: Figure 3 As shown, it includes:

[0122] S104-1: Group consecutive abnormal residual height points in the residual height time series without any intervals into a group, and define each group as an abnormal unit; if an abnormal unit contains more than or equal to two abnormal residual height points, it is marked as a continuous out-of-standard segment; if an abnormal unit contains only one abnormal residual height point, it is marked as an isolated out-of-standard point.

[0123] It is important to understand that continuous exceedances usually indicate stable, systemic process deviations, which are more harmful; while isolated exceedances are more likely to be occasional phenomena caused by momentary disturbances (such as splashing or slight vibrations), which are relatively easy to recover from. Therefore, distinguishing and analyzing them can enable the system to more accurately determine the root cause and severity of the problem, avoid overreacting to occasional fluctuations, and prevent underestimating systemic problems.

[0124] S104-2: For continuous exceeding segments and isolated exceeding points, different association analysis rules are used to calculate the association contribution value.

[0125] To accurately calculate the correlation contribution value for consecutive exceedance segments using correlation analysis rules, as an example, for each consecutive exceedance segment, the difference between the residual height value at each time point within the consecutive exceedance segment and the boundary of the allowable residual height interval is calculated. This difference is taken as the residual height exceedance amount at each time point within the consecutive exceedance segment, and sorted by time to form a residual height exceedance amount sequence. The welding temperature sequence at the corresponding time point is extracted from the temperature time series, and the first comparison temperature value of the previous time point in the welding temperature sequence is extracted. The Pearson correlation coefficient between the residual height exceedance amount sequence and the welding temperature sequence is calculated and normalized. The following steps are taken: First, the temperature value at each moment in the welding temperature sequence is calculated as the basic correlation degree. Then, the first difference between the temperature value and the first comparison temperature value is calculated. For each first difference, if the first difference is greater than 0, it is retained; if the first difference is less than or equal to 0, it is corrected to 0. The arithmetic mean of all corrected first differences is calculated as the average positive overheat amount. The average positive overheat amount is divided by the temperature width of the preset allowable temperature range to obtain the segment temperature penalty term. The sum of the positive integer 1 and the segment temperature penalty term is multiplied by the basic correlation degree to obtain the correlation contribution value of the continuous overheating segment.

[0126] It should be noted that the specific method for calculating the residual height difference is similar to the specific method for calculating the temperature difference, only the input parameters are different, which will not be repeated in this embodiment.

[0127] The corresponding time period specifically refers to the time segment that is completely aligned with the continuous exceedance of the temperature limit. For example, if the temperature limit is continuously exceeded within the 5-minute segment from 10:00 to 10:05, the system will accurately extract the temperature data for the same time segment (10:00 to 10:05) from the temperature time series.

[0128] The first comparison temperature value refers to the temperature value of the normal welding state immediately preceding the start of the continuous exceeding the standard section.

[0129] It should be noted that the Pearson correlation coefficient between the superscalar sequence and the temperature sequence is normalized to obtain the basic correlation degree. This is the same normalization method used in step S103 to normalize the Pearson correlation coefficient as a similarity value, and will not be repeated in this embodiment.

[0130] It is important to understand that abnormal increases in welding temperature are more likely to cause changes in the weld pool state and overheating of the metal, leading to defects such as excessive weld reinforcement. On the other hand, a decrease or stagnation in temperature has a weaker correlation with excessive weld reinforcement. Therefore, by correcting non-positive values ​​to 0, the calculated average positive overheat can more purely reflect temperature fluctuations that have a positive impact on abnormal weld reinforcement, allowing the analysis to focus more on key risks.

[0131] It is important to understand that the basic correlation degree takes a value between [0, 1]. The larger the basic correlation degree value, the stronger the synchronicity between the changes in the excess height overscalar sequence and the welding temperature sequence, that is, the more significant the risk correlation. The average positive overtemperature amount divided by the temperature width yields a dimensionless segment temperature penalty term, which is used to quantify the degree of abnormal temperature rise. The larger the segment temperature penalty term, the deeper the degree of abnormal temperature rise. The formula constructs an amplification coefficient of not less than 1 by "1 + segment temperature penalty term". Its core logic is: even if there is only a weak correlation between temperature and excess height (low basic correlation degree), as long as the temperature deviates significantly in the positive direction (large segment temperature penalty term), it will be identified as a potential risk and its correlation contribution value will be amplified. Conversely, if the two are strongly correlated and accompanied by a high temperature rise, the correlation contribution value will be significantly amplified. This ensures that the correlation contribution value increases with the increase of correlation strength and temperature anomaly degree.

[0132] To accurately calculate the correlation contribution value for isolated exceedance points using correlation analysis rules, as an example, for each isolated exceedance point, the excess height of the isolated exceedance point is calculated; the excess height is divided by the width of the allowable excess height interval to obtain the excess height ratio; the temperature value at the corresponding time of the isolated exceedance point and the second comparison temperature value at the previous time of the isolated exceedance point are obtained; the second difference between the temperature value at the corresponding time of the isolated exceedance point and the second comparison temperature value is calculated; if the second difference is negative or zero, the second difference is counted as 0; the second difference is divided by the temperature width to obtain the point temperature impact term; the sum of the positive integer 1 and the point temperature impact term is multiplied by the excess height ratio to obtain the correlation contribution value of the isolated exceedance point.

[0133] The time corresponding to an isolated out-of-range point specifically refers to the precise moment when that isolated out-of-range point itself occurs. For example, if the excess height only exceeds the limit at the sampling point of 10:03, then the system will extract the welding temperature value recorded at 10:03 from the temperature time series.

[0134] The second comparison temperature value refers to the temperature value at the normal moment immediately before the isolated exceedance point occurs. Using the previous example, it is the temperature value at 10:02.

[0135] The second difference refers to the value obtained by subtracting the temperature value at the moment the isolated excess point occurs from the temperature value at the moment it occurs (i.e., the second comparison temperature value), which measures whether the welding temperature has undergone an instantaneous positive jump before the excess height anomaly point appears.

[0136] It should be noted that the width of the allowable excess height range = the upper limit of the allowable excess height range - the lower limit of the allowable excess height range, and as can be seen from the example of the value of the allowable excess height range, the width of the allowable excess height range is greater than zero.

[0137] It is important to understand that, for isolated anomalies, engineering inferences suggest that they are more likely to be caused by transient disturbances. If the temperature did not rise or even fall before the anomaly occurred, then this residual height anomaly is unlikely to be caused by a transient increase in welding heat input. In the case that the second difference is negative or zero, the second difference value being zero can exclude such occasional anomalies with "no temperature shock background". This allows the point temperature shock term calculated subsequently to only capture and quantify the anomaly risk that may be caused by a transient temperature rise.

[0138] It's important to understand that the excess height directly quantifies the degree to which anomalies in the excess height exceed the acceptable range. The larger the excess height, the more severely the anomaly exceeds the acceptable range. The second difference divided by the temperature width yields a dimensionless point temperature shock term, used to quantify the relative intensity of the instantaneous temperature rise at the previous moment. The larger the second difference, the more positively the welding temperature rises. The formula constructs an amplification factor of no less than 1 through "1 + point temperature shock term." The core logic is: even if the excess height itself is not large, if it is accompanied by a significant instantaneous temperature rise before it occurs (large point temperature shock term), it suggests that the exceedance is more likely caused by harmful heat input shock, and its risk contribution should be amplified. Conversely, if there is no temperature rise or fall (point temperature shock term is 0), the associated contribution value is determined only by the excess height itself. This ensures that the risk contribution value of isolated points increases simultaneously with the severity of the excess height and the intensity of the preceding temperature shock.

[0139] S104-3: Using the duration of each exceeding segment and exceeding point as weight, a weighted average of all associated contribution values ​​is performed to obtain the comprehensive temperature surplus correlation index. The weight of each continuous exceeding segment is the number of abnormal surplus high points it contains, and the weight of each isolated exceeding point is 1.

[0140] It should be noted that, under a fixed sampling frequency, the number of outliers is directly equivalent to the duration of the outlier. The longer the duration of continuous outliers, the more persistent the welding process deviates from the normal state, which usually means a more stable and more serious systemic process problem. Therefore, it should be given a higher weight. So, for a continuous outlier segment, its weight is set as the number of residual height outliers contained in the continuous outlier segment.

[0141] It should be noted that since each isolated out-of-specification point represents an independent and transient abnormal event, its basic importance unit can be regarded as "1", and its weight is fixed at 1 to ensure that its contribution value is not amplified. This contrasts with continuous out-of-specification segments, highlighting the design logic that persistent anomalies are more critical than occasional anomalies. Therefore, the weight of each isolated out-of-specification point is 1.

[0142] In this embodiment, the product of the comprehensive residual temperature correlation index and the calibration temperature risk value is normalized to obtain the welding severity index.

[0143] It is important to understand that the calibration temperature risk value quantifies the historical deviation and fluctuation risk of the temperature time series itself, while the comprehensive temperature residual correlation index quantifies the correlation strength between temperature and residual height anomalies. The comprehensive temperature residual correlation index and the calibration temperature risk value are multiplied and normalized to obtain the welding severity index. The larger the welding severity index value, the higher the temperature itself is (high calibration temperature risk value) and the stronger the correlation with quality defects (high comprehensive temperature residual correlation index), which represents a dual high-risk state.

[0144] It should be noted that, in order to normalize the product of the comprehensive residual temperature correlation index and the calibration temperature risk value, as a preferred implementation method, it is divided by a preset reference value, i.e., Welding severity index = (Calibration temperature risk value × Comprehensive residual temperature correlation index) ÷ Reference value. This method ensures that when the product reaches or exceeds the historical typical abnormal level, the welding severity index will approach or exceed 1, thus intuitively and quantitatively reflecting the degree of deviation of the current welding status.

[0145] It should be noted that the specific value of the reference benchmark can be determined based on the maximum product value obtained from the statistical analysis of historical qualified batch data. This embodiment does not impose a specific limitation. For example, based on the welding data of 100 historical qualified batches, the product of the calibration temperature risk value and the comprehensive temperature residual correlation index of each batch can be calculated, and the maximum value or the 95th percentile (e.g., 8.5) can be taken as the reference benchmark value.

[0146] S105: Based on the comparison results between the welding severity index and the preset threshold, execute the corresponding manufacturing process management strategy.

[0147] It should be noted that the welding severity index can be compared with preset thresholds in a tiered manner, triggering corresponding process management actions. For example, the system can set a yellow warning threshold (e.g., 0.7) and a red intervention threshold (e.g., 0.9). When the welding severity index is below the warning threshold, the process proceeds normally; when the welding severity index reaches the warning threshold, the system can automatically record the event and alert the operator; when the welding severity index reaches or exceeds the intervention threshold, the system can automatically implement stricter strategies, such as pausing welding, issuing alarms, locking the workstation, or forcibly triggering process parameter review and adjustment procedures.

[0148] A full-process digital management system for wind power equipment manufacturing includes a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of a full-process digital management method for wind power equipment manufacturing.

[0149] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0150] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A full-process digital management method for wind power equipment manufacturing, characterized in that, The method includes: Simultaneously acquire the welding temperature time sequence and weld reinforcement time sequence during the welding process of wind turbine towers, and obtain historical temperature time sequence under the same welding conditions; Extract abnormal temperature points that exceed the allowable temperature range and normal temperature points that do not exceed it from the temperature time series; determine the basic temperature risk value based on the abnormal temperature points; determine the comprehensive temperature risk value based on the fluctuation characteristics of normal temperature points within the allowable temperature range, the degree of difference between the temperature time series and the boundary of the allowable temperature range, and the basic temperature risk value. Extract historical temperature subsequences from historical temperature time series that have the same duration as the temperature time series and meet preset similarity conditions; determine the risk deviation coefficient based on the similarity between the temperature time series and the historical temperature subsequences; and determine the calibration temperature risk value by combining the comprehensive temperature risk value. Extract abnormal residual height points that exceed the allowable residual height range from the residual height time series; determine the comprehensive temperature residual height correlation index based on the abnormal residual height points and the temperature values ​​at the corresponding times in the temperature time series; determine the welding severity index based on the comprehensive temperature residual height correlation index and the calibration temperature risk value. Based on the comparison between the welding severity index and the preset threshold, the corresponding manufacturing process management strategy is implemented.

2. The method for full-process digital management of wind power equipment manufacturing according to claim 1, characterized in that, The welding temperature timing sequence is a temperature timing sequence during the steady-state stage of welding; the method further includes: The welding temperature value is collected in real time. When the welding temperature value enters the preset process temperature range for the first time, and the temperature value is within the preset process temperature range for a preset number of consecutive sampling times after the first entry, the welding process is determined to have entered the welding steady state stage.

3. The method for full-process digital management of wind power equipment manufacturing according to claim 2, characterized in that, The process for determining the baseline temperature risk value includes: The total number of temperature points in the welding temperature time sequence, the number of abnormal temperature points exceeding the preset temperature allowable range, and the temperature width of the preset temperature allowable range are statistically analyzed. Calculate the temperature difference between each abnormal temperature point and the boundary of the allowable temperature range; calculate the arithmetic mean of all temperature differences as the average exceedance. Divide the average exceedance by the temperature width to obtain the average exceedance ratio; divide the number of abnormal temperature points by the total number of temperature points to obtain the percentage of abnormal temperature points. Calculate the product of the percentage of abnormal temperature points and the average excess percentage as a risk factor; Add the risk factor to the positive integer 1 to obtain the base temperature risk value.

4. The method for full-process digital management of wind power equipment manufacturing according to claim 2, characterized in that, The process for determining the comprehensive temperature risk value includes: Calculate the highest and lowest temperature values ​​among all normal temperature points; The continuous normal temperature points are divided into at least one continuous normal segment; for each continuous normal segment, the segment fluctuation intensity is calculated based on the fluctuation characteristics of the temperature value within the continuous normal segment; the arithmetic mean of the fluctuation intensity of all segments is divided by a preset fluctuation intensity reference value to obtain the volatility risk component. The difference between the upper limit of the allowable temperature range and the highest temperature value is calculated as the upper boundary temperature value; the difference between the lowest temperature value and the lower limit of the allowable range is calculated as the lower boundary temperature value; based on the upper boundary temperature value, the lower boundary temperature value, and the preset normal temperature value, the boundary risk component is calculated. Calculate the sum of the volatility risk component and the boundary risk component, add the summation result to the positive integer 1, and obtain the state risk coefficient; The product of the state risk coefficient and the basic temperature risk value is used as the comprehensive temperature risk value.

5. The method for full-process digital management of wind power equipment manufacturing according to claim 2, characterized in that, The step of extracting historical temperature subsequences from historical temperature time series that have the same duration as the temperature time series and meet preset similarity conditions includes: The temperature time series of the current welding steady-state stage is used as the reference series; For each historical temperature time series, multiple candidate subsequences with the same length as the baseline sequence are sequentially extracted from the historical temperature time series using a sliding window method. For each candidate subsequence, the Pearson correlation coefficient between the candidate subsequence and the baseline sequence is calculated as a morphological similarity score. From all candidate subsequences, the candidate subsequence with the highest morphological similarity score is selected as the historical temperature subsequence.

6. The method for full-process digital management of wind power equipment manufacturing according to claim 5, characterized in that, The process for determining the risk deviation coefficient includes: Obtain the preset reference deviation; For each historical temperature subsequence, the Pearson correlation coefficient between the historical temperature subsequence and the temperature time series is normalized and used as a similarity value; the arithmetic mean of all similarity values ​​is calculated as the average similarity. Calculate the difference between the positive integer 1 and the average similarity, and use it as the average deviation. Divide the average deviation by the preset reference deviation to obtain the relative deviation. Calculate the sum of the positive integer 1 and the relative deviation, and use it as the risk deviation coefficient.

7. The method for full-process digital management of wind power equipment manufacturing according to claim 2, characterized in that, The process for determining the comprehensive residual temperature correlation index includes: In the residual height time series, consecutive abnormal residual height points without intervals are grouped together, and each group is defined as an abnormal unit. If an abnormal unit contains more than or equal to two abnormal residual height points, it is marked as a continuous out-of-specification segment. If an abnormal unit contains only one abnormal residual height point, it is marked as an isolated out-of-specification point. For continuous exceedance segments and isolated exceedance points, different association analysis rules are used to calculate the association contribution value; The comprehensive residual temperature correlation index is obtained by weighting all related contribution values ​​with the duration of each exceeding segment and exceeding point as the weight. The weight of each continuous exceeding segment is the number of abnormal residual high points it contains, and the weight of each isolated exceeding point is 1.

8. The method for full-process digital management of wind power equipment manufacturing according to claim 7, characterized in that, The calculation of the correlation contribution value for consecutive exceedance segments using correlation analysis rules includes: For each continuous excess segment, the excess height difference between the excess height value and the boundary of the allowable excess height interval at each moment within the continuous excess segment is calculated as the excess height amount at each moment within the continuous excess segment, and sorted by time to form an excess height amount sequence; the welding temperature sequence at the corresponding moment is extracted from the temperature time series, and the first comparison temperature value of the previous moment of the welding temperature sequence is extracted. The Pearson correlation coefficient between the superscalar superscalar sequence and the welding temperature sequence was calculated and normalized to serve as the basic correlation coefficient. Calculate the first difference between the temperature value at each moment in the welding temperature sequence and the first comparison temperature value; for each first difference, if the first difference is greater than 0, retain the first difference; if the first difference is less than or equal to 0, correct the first difference to 0. Calculate the arithmetic mean of all corrected first differences as the average positive overtemperature; divide the average positive overtemperature by the temperature width of the preset temperature allowable range to obtain the segment temperature penalty term; Multiply the sum of the positive integer 1 and the segment temperature penalty term by the basic correlation degree to obtain the correlation contribution value of the continuous out-of-standard segment.

9. A method for full-process digital management of wind power equipment manufacturing according to claim 8, characterized in that, The calculation of the correlation contribution value for isolated out-of-standard points using correlation analysis rules includes: For each isolated exceedance point, calculate the residual exceedance amount of the isolated exceedance point; and obtain the temperature value at the corresponding time of the isolated exceedance point, as well as the second comparison temperature value at the time before the isolated exceedance point; Divide the excess height by the width of the allowable excess height range to obtain the excess height ratio; Calculate the second difference between the temperature value at the time corresponding to the isolated exceeding point and the second comparison temperature value. If the second difference is negative or zero, the second difference is counted as 0. Divide the second difference by the temperature width to obtain the point temperature shock term; Multiply the sum of the positive integer 1 and the point temperature shock term by the excess height ratio to obtain the correlation contribution value of isolated excess points.

10. A fully digital management system for the entire process of wind power equipment manufacturing, characterized in that, The system includes a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of the method as described in any one of claims 1 to 9.

Images