Intelligent analysis method and system for titanium alloy smelting effect

Abstract

The invention relates to the technical field of titanium alloy smelting analysis, in particular to a titanium alloy smelting effect intelligent analysis method and system, and the method comprises the steps: recording a smelting event of smelting equipment in a preset time period; recording abnormal events of smelting events in each time of smelting in a single parameter and parameter pair combination mode to form an associated parameter chain; performing mode clustering on the associated parameter chain according to a combined single parameter and parameter pair form, and determining a propagation path after clustering; based on the propagation path, performing single event scoring on each batch of abnormal events, and constructing a batch-index incidence matrix; according to the change trend of each key index in the batch-index incidence matrix during multiple times of smelting, obtaining the transfer relation of abnormal events in multiple times of smelting; determining target parameters of each batch on the basis of a passing relation of abnormal events in multiple smelting, and outputting the target parameters; the accuracy of smelting analysis and the efficiency of dynamic perception are realized.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of titanium alloy smelting analysis, in particular to a titanium alloy smelting effect intelligent analysis method and system. BACKGROUND

[0002] Titanium alloy is a kind of metal material composed of titanium and other metal elements, which has high strength, low density and corrosion resistance, and is widely used in the fields of aviation, aerospace, chemical industry and the like. Titanium alloy smelting technology mainly includes vacuum self-consumption arc furnace smelting, electron beam cold bed furnace smelting, plasma cold bed furnace smelting and vacuum condensation crust furnace smelting, etc. The smelting effect of titanium alloy is affected by the adjustment of smelting parameters, and it is necessary to deduce the current titanium alloy smelting effect according to the specific process of each parameter in the smelting process to prevent low smelting efficiency and resource waste.

[0003] For example, Chinese patent publication No. CN117910886A discloses a smelting effect intelligent analysis method and system applied to titanium alloy smelting. The method comprises: obtaining element information and attribute characteristic information of titanium alloy batch smelting, and corresponding product sampling obtains a sample group of multiple smelting batches, detects each product sample in the sample group to obtain detection result data, processes the detection result data of each sample to obtain a product sample smelting effectiveness index, extracts element characteristic data of each batch according to the smelting element information and processes to obtain an interference coefficient, and then corrects and aggregates the smelting effectiveness index of each batch product sample to obtain smelting quality and efficiency evaluation data. The titanium alloy smelting quality of all batches is evaluated through the comparison result with the threshold value; thereby, the effect of titanium alloy smelting batch product samples is detected and the whole batch smelting quality effect is evaluated based on big data, and the smelting quality evaluation result of titanium alloy smelting products is obtained.

[0004] For example, Chinese patent publication No. CN120598431A discloses a titanium alloy smelting quality management method and system. The application comprises collecting real-time monitoring data in the titanium alloy smelting process; classifying and storing the real-time monitoring data to obtain historical process data; based on the equipment operation data and the preset first warning rule, equipment abnormality warning is performed, and first warning information is generated; based on the process control node data, a process capability index is obtained, based on the obtained process capability index and the preset second warning rule, process data deviation warning is performed, and second warning information is generated; based on the ingot product manufacturing and detection data and the preset third warning rule, product manufacturing and detection warning is performed, and third warning information is generated; the first warning information, the second warning information and the third warning information are displayed and processed through man-machine interaction.

[0005] The prior art analyzes the porosity, grain size, mechanical properties and other aspects during smelting to indicate the smelting effect of titanium alloy, and indicates the abnormal situation existing during smelting through the mode of running node early warning, but there are complex parameters in the smelting process, and it is difficult to identify the scene of abnormal association triggering through the mode of single-point abnormal early warning, and it is easy to ignore the real-time perception of the smelting process, the dynamic tracking of the abnormal propagation path and the prediction analysis process of the long-term trend, resulting in obvious decline in the quality consistency and equipment utilization rate of titanium alloy smelting, causing the smelting effect identification efficiency of titanium alloy to be reduced. SUMMARY

[0006] In order to solve the above technical problems, the technical scheme adopted by the present application is: a titanium alloy smelting effect intelligent analysis method, comprising: S1, acquiring smelting process parameters, mapping the smelting process parameters to the smelting equipment, and recording the smelting events of the smelting equipment in a preset time period.

[0007] S2, recording the abnormal events of each smelting in the form of single parameter and parameter pair combination of the smelting events of each smelting, forming an associated parameter chain.

[0008] S3, performing mode clustering on the associated parameter chain in the form of combined single parameter and parameter pair, and determining the propagation path after clustering.

[0009] S4, based on the propagation path, performing single event scoring on each batch of abnormal events, calculating the transmission efficiency, single event scoring and path characteristics of each batch based on the associated parameter chain, and constructing a batch-index association matrix; recording the change trend of each key index in the batch-index association matrix in multiple smeltings, and obtaining the transmission relationship of abnormal events in multiple smeltings by analyzing the correlation of the change trend of each key index.

[0010] S5, based on the transmission relationship of abnormal events in multiple smeltings, tracing the smelting events, determining the target parameters of each batch, and outputting the target parameters.

[0011] A titanium alloy smelting effect intelligent analysis system, comprising: a smelting event module for acquiring smelting process parameters, mapping the smelting process parameters to the smelting equipment, and recording the smelting events of the smelting equipment in a preset time period.

[0012] A smelting association module for recording the abnormal events of each smelting in the form of single parameter and parameter pair combination of the smelting events of each smelting, forming an associated parameter chain.

[0013] A propagation path module for performing mode clustering on the associated parameter chain in the form of combined single parameter and parameter pair, and determining the propagation path after clustering.

[0014] The transmission analysis module is used for scoring each batch of abnormal events based on a propagation path, calculating the transmission efficiency, single event score and path feature of each batch based on the associated parameter chain, and constructing a batch-index association matrix; the change trend of each key index in the batch-index association matrix is recorded during multiple smelting, and the transmission relationship of abnormal events in multiple smelting is obtained by analyzing the correlation of the change trend of each key index.

[0015] The parameter output module is used for tracing the smelting event based on the transmission relationship of abnormal events in multiple smelting, determining the target parameter of each batch, and outputting the target parameter.

[0016] The beneficial effects of the present application are as follows: first, the present application realizes the standardized recording of smelting events by obtaining smelting process parameters and mapping them to equipment, providing a data basis for subsequent analysis; then the associated parameter chain is formed in the form of single parameter and parameter pair combination, which enhances the correlation identification of abnormal events; then the associated parameter chain is converted into a propagation path through mode clustering, realizing the abstraction and classification of abnormal patterns and reducing the analysis complexity; the batch-index association matrix is constructed and the trend is analyzed, realizing the dynamic monitoring of abnormal transmission efficiency. Finally, based on the transmission relationship, the target parameter is traced, realizing the closed loop of data processing, completing the overall identification and analysis processing of the smelting process, and ensuring that abnormal events are handled in a timely and accurate manner.

[0017] Second, the present application determines the smelting process parameters identified under the current scenario by determining the smelting boundary, risk exposure time and risk mitigation time, and determines the current data according to the stable and unstable forms presented by the parameters, to enhance the form of parameter dependence and improve the accuracy of abnormal event identification and processing.

[0018] Third, the present application abstracts abnormal events into abnormal nodes and time intervals, and constructs a multi-level associated parameter chain by checking the intersection of time intervals and recursively finding chain nodes: the parameter pair abnormality is set as the root node to realize the transmission combination of abnormal events under various parameter combinations, and to avoid the problem of smelting process early warning lag. BRIEF DESCRIPTION OF DRAWINGS

[0019] The present application will be further described below in combination with the drawings and examples.

[0020] Figure 1 It is a flowchart of a titanium alloy smelting effect intelligent analysis method.

[0021] Figure 2 It is a flowchart of step S1 of a titanium alloy smelting effect intelligent analysis method.

[0022] Figure 3 It is a flowchart of step S2 of a titanium alloy smelting effect intelligent analysis method.

[0023] Figure 4 is a flowchart of step S4 of the intelligent analysis method for titanium alloy smelting effect.

[0024] Figure 5 is a system framework diagram of the intelligent analysis system for titanium alloy smelting effect. DETAILED DESCRIPTION

[0025] Embodiments of the present application are described in detail below. The embodiments described below are exemplary only, and are intended to explain the present application, and are not to be understood as limiting the present application. If a specific technique or condition is not mentioned in the embodiments, the technique or condition described in the literature in the art or according to the product manual is used.

[0026] Referring to Figure 1 An intelligent analysis method for titanium alloy smelting effect includes: S1, acquiring smelting process parameters, mapping the smelting process parameters to a smelting device, and recording smelting events of the smelting device in a preset time period.

[0027] S2, recording abnormal events during each smelting in a single parameter and parameter pair combination manner, forming a correlation parameter chain.

[0028] S3, performing mode clustering on the correlation parameter chain in the form of the combined single parameter and parameter pair, and determining a propagation path after clustering.

[0029] S4, based on the propagation path, performing single event scoring on each batch of abnormal events, calculating the transfer efficiency, single event scoring and path characteristics of each batch based on the correlation parameter chain, and constructing a batch-index correlation matrix; recording the change trend of each key index in the batch-index correlation matrix in multiple smeltings, obtaining the transfer relationship of abnormal events in multiple smeltings by analyzing the correlation of the change trend of each key index.

[0030] S5, based on the transfer relationship of abnormal events in multiple smeltings, tracing the smelting events, determining the target parameters of each batch, and outputting the target parameters.

[0031] The smelting process parameters are acquired in step S1, including smelting power, current, voltage, vacuum degree, leakage rate, cooling water flow / temperature, crucible rotation speed and smelting time of each stage and other parameters, which are combined in the form of time-stamped smelting process parameters; the above-mentioned preset time period is the length of time for completing a batch of titanium alloy production as the preset time period. The smelting event refers to a specific behavior occurring in the whole smelting process which can be monitored and identified and has a direct or indirect impact on the stability of the production process or the quality of the final product, including smelting events corresponding to raw materials such as raw material ratio deviation events, raw material granularity unevenness events and raw material impurity exceeding events, smelting events related to equipment such as VAR electrode short circuit, EB furnace electron gun beam instability, smelting cavity cleanliness not meeting standards and vacuum system leakage, smelting events corresponding to process parameter abnormalities such as molten pool over-temperature, smelting atmosphere adaptation degree abnormality, smelting speed fluctuation and over-temperature and over-pressure amplitude frequency abnormality, and quality-related smelting events such as ingot inclusion porosity, grain size abnormality and performance not meeting standards.

[0032] In conventional titanium alloy smelting, 2-3 times of smelting are usually required in titanium alloy arc smelting to improve the effect, and at this time, the difference in the effect of each batch is also a key, the purpose of acquiring the smelting process parameters is to judge the smelting abnormal part existing in the process before and after the titanium alloy production, to emphasize the consistency and transmission process of the smelting abnormality, to explain the conditions of the smelting abnormality, and to improve the efficiency of the subsequent titanium alloy smelting production.

[0033] As shown in Figure 2 The implementation mode of step S1 includes: S11, based on the mapping smelting equipment, determining the smelting boundary of the current smelting process parameter, and the risk exposure time of the smelting process parameter reaching the smelting boundary and the risk mitigation time of reducing from the smelting boundary to the normal value on each smelting equipment. When calculating the smelting boundary, the arithmetic mean of each parameter (such as vacuum degree, current) is calculated based on 100 batches of normal data of the current equipment smelting; then the standard deviation is calculated based on the above-mentioned 100 batches of data, and when the currently identified titanium alloy type changes, the normal data of the corresponding model is re-collected.

[0034] S12, the average value and the standard deviation of the smelting process parameter in the risk exposure time and the risk mitigation time are counted, the stability of the smelting process parameter in each smelting is determined, and the smelting event is extracted with the smelting boundary when the stability of the smelting process parameter meets the requirements.

[0035] S13, when the stability of the smelting process parameter does not meet the requirements, the smelting process parameter is divided into multiple stages, the smelting process parameter is associated according to each stage, the correlation of the smelting process parameter in different stages is determined, the smelting process parameter with the maximum correlation is output as the extracted smelting event.

[0036] The system monitors the values of the smelting process parameters based on the safety boundary of the preloaded parameters of the current device. When a parameter starts to deviate from the normal value and moves towards the boundary, the system starts timing and calculates the time taken by the parameter to reach the boundary value from the normal range value, which is recorded as the risk exposure time. At this time, the safety boundary represents the confidence interval of each parameter under normal operation of the device. The data within this confidence interval is taken as the normal value, that is, the current smelting boundary is set in the form of mean ± three times the standard deviation. Similarly, when the parameter returns to the normal range from the smelting boundary, it is recorded as the risk mitigation time. These two times represent the risk increase and decrease processes during the current production, and the corresponding times are recorded.

[0037] At this time, the recorded time can be recorded in the form of continuous 7-point rise, that is, when the current value is gradually rising and approaching the upper limit value of the confidence interval, the system starts timing at the 7th time point or the 6th time point, records the process and time of the current smelting process parameter exceeding the upper limit value of the confidence interval, to obtain the risk exposure time of deviation from the upper limit value. Similarly, the processing method of approaching the lower limit value and the upper limit value of the confidence interval is consistent, and the continuous 7-point decline is used to obtain the risk exposure time of deviation from the lower limit value.

[0038] The risk mitigation time is timed at the time point when the smelting process parameter exceeds the upper and lower limit values of the confidence interval and reverses, and the time of decreasing to the confidence interval is recorded.

[0039] The system does not only process the event of once approaching the smelting boundary. At this time, the average value and standard deviation of these times are calculated to evaluate the stability in the form of standard deviation and average value. If the stability meets the requirements, the event significantly deviating from the current data will be marked as an abnormal event.

[0040] If the stability does not meet the requirements, these values will be relatively scattered, resulting in difficulty in predicting and comparing the smelting evaluation process. At this time, the root cause of instability is found. The time of the parameter reaching the boundary is associated with different stages of smelting, such as preheating stage, melting stage, refining stage, pouring stage, etc. Focus on which time period is more unstable and find a time period with obvious problems.

[0041] Therefore, the implementation manner of step S12 further includes: S121, based on the smelting demand information of the titanium alloy, a preset smelting stability threshold is used to analyze the current smelting process parameter; according to the processing flow of the titanium alloy smelting, the preset smelting stability threshold is compared with the average value and the standard deviation of the current smelting process parameter.

[0042] S122, if any of the average value or the standard deviation of the current smelting process parameter exceeds the upper and lower limit range of the preset smelting stability threshold value, it is considered that the stability of the smelting process parameter does not meet the requirements.

[0043] S123, only if the average value and the standard deviation of the current smelting process parameter are both within the upper and lower limit range of the preset smelting stability threshold value, it is considered that the stability of the smelting process parameter meets the requirements.

[0044] It should be noted that the current preset smelting stability threshold value is based on the average value and the standard deviation of the smelting process parameter which is continuously distributed in the historical data. The average value and the standard deviation obtained by multiple batches or multiple time periods are sorted from small to large. The average value and the standard deviation at the 75% data after sorting are taken as the upper limit value of the preset smelting stability threshold value. The smallest average value and standard deviation during sorting are taken as the lower limit value of the preset smelting stability threshold value. The average value and the standard deviation of the smelting process parameter according to the risk exposure time and the risk mitigation time are compared with the preset smelting stability threshold value. When it exceeds the upper and lower limit value, it is considered as unstable data, indicating that the data fluctuation is large and the average value is large in the corresponding time period. At this time, there is a relative limitation in identifying smelting events only by smelting boundary. The abnormal situation existing in the smelting event should be focused on the time period of the abnormal event to identify the abnormal situation existing in the smelting event.

[0045] As for the processing mode of step S13, it is inclined to process smelting process parameters in different stages. The smelting process is divided into multiple stages. The Pearson correlation coefficient is calculated for the parameters in each stage to find the parameter pair with the largest correlation. The parameter pair with the largest global correlation is found by comparing the correlation results of all stages. This parameter pair represents the key smelting event in the current smelting process, representing the positioning of abnormal events from the fuzzy time period to the specific smelting stage, and directly outputting the specific stage time of the abnormal event according to the correlation coefficient of the parameter pair in each stage. For example, the smelting power and the vacuum degree are highly correlated, such as a correlation coefficient greater than 0.8, indicating a vacuum system failure, resulting in gas inclusion. According to the obtained correlation coefficient, it is defined that there may be smelting events in each stage, and these parameter pairs are taken as the output smelting events, which can locate the abnormal situations existing in multiple stages. At this time, the calculated data is first normalized and standardized to eliminate the dimension between multiple groups of data to complete the calculation of the corresponding data. It should be noted that the calculated smelting process parameters will be calculated according to different parameters in multiple divided stages, such as smelting power, current, voltage, vacuum degree, leakage rate, cooling water flow / temperature, crucible rotation speed and smelting time in each stage.

[0046] The implementation of step S13 further includes: forming parameter pairs by pairing the smelting process parameters in each stage, calculating the Pearson correlation coefficient of all parameter pairs to obtain the correlation of parameter pairs in each stage.

[0047] For each stage, find the parameter pair with the largest correlation, and consider it as the correlation result under the corresponding stage.

[0048] Compare the correlation results of all stages to determine the globally maximum correlation parameter pair and output it as a smelting event. This output smelting event will record the corresponding parameter pair and the timestamp of the parameter pair to explain the most significant correlation parameter pair under the current smelting process analysis, which represents the most serious process imbalance and explains the main anomaly in the extreme discrete scenario of data, serving as the main part of subsequent processing.

[0049] In an embodiment of the present application, when the smelting event needs to be processed in step S2, the abnormal events existing at that time are associated in a chain to form a parameter chain displayed in a chain. At this time, the abnormal events represent the smelting events extracted in step S13 and the smelting events exceeding the smelting boundary in step S12, which belong to the abnormal events existing in the current smelting process.

[0050] As shown in Figure 3 The implementation of step S2 includes: S21, based on the abnormal events at each smelting, extracting the smelting process parameters associated with the abnormal events in the form of single parameters and parameter pairs, and considering each monitored single parameter and parameter pair as an abnormal node.

[0051] S22, for each abnormal node, the time period when the abnormal event occurs is counted, and the continuously occurring abnormal nodes are merged into an abnormal time interval, and the abnormal time interval contains at least one abnormal node; when merging the abnormal time interval, the time interval between the time periods covered by the two abnormal events needs to be determined, for example, the current monitoring abnormal node is displayed in the form of fluctuation curve, the time value of the first trough when the fluctuation occurs is taken as the basis for setting at this time, so that the continuously occurring abnormal events are recognized, and the time value corresponding to the trough will represent the time interval of the current continuous abnormal event, and the interval of the merged abnormal time interval is only set to 1.2 or other multiples of the time value corresponding to the trough, which only needs to be slightly larger than the trough to recognize the relatively comprehensive abnormal time interval; when identifying the trough, the inflection point of the previous 5 sampling points is selected, the latter 5 sampling points are raised, and the inflection point value is lower than the adjacent 10 sampling points, so that the corresponding trough is obtained; the time value corresponding to the first trough is T, and the merging interval is 1.2T, from the first time the parameter exceeds the boundary; if the interval of the two abnormal intervals is less than or equal to 1.2T, they are merged into one interval. As for the case of not merging, the average abnormal time of the current smelting process parameter when the abnormal event occurs is viewed through the historical data, and a time value of 5-10 times is set to explain the isolated abnormal event under the current abnormal time interval.

[0052] S23, for each abnormal time interval, check whether the abnormal time interval composed of all other abnormal nodes intersects with the current abnormal time interval, that is, there is time interval overlap or time point inclusion, which means that there is a relatively overlapping part on the two time intervals; if there is intersection, the abnormal node in the current abnormal time interval is regarded as the root node, and the other abnormal nodes are regarded as the secondary chain nodes of the abnormal node in the current abnormal time interval. It should be noted that the secondary chain nodes are gradually set according to the order of time. For example, the abnormal time interval of the vacuum degree is [t1, t2], and the abnormal time interval of the current is [t1.5, t2.5], at this time, the abnormal time interval of the current is contained in the abnormal time interval of the vacuum degree, that is, the chain structure of vacuum degree→current, and the current is the secondary chain node thereof.

[0053] S24, if there is no intersection, the current abnormal time interval is pushed forward according to the time stamp to form the output associated parameter chain.

[0054] S25, all secondary chain nodes are counted, and the tertiary chain nodes under the secondary chain nodes are found according to the way of abnormal time interval intersection, and the tertiary chain nodes are recursively found until there is no intersection in the abnormal time interval, and the formed chain nodes and corresponding abnormal nodes are taken as the output associated parameter chain.

[0055] The process described above is to associate the parameters with abnormality in the smelting process according to the time period of occurrence, and output the corresponding associated parameters after association. It should be noted that if there are multiple levels of chain association, the maximum recursion depth needs to be set, such as setting 5 levels as the maximum depth, to prevent the structure from being too complex due to gradual cycling of some parameters.

[0056] Preferably, when comparing abnormal time intervals, if the corresponding abnormal node is a single-parameter abnormal node and a parameter pair abnormal node, the relationship abnormal first and single-parameter abnormal first also need to be identified.

[0057] The relationship abnormal first indicates that if the parameter pair relationship abnormality occurs earlier than the single abnormality of its constituent parameters, the system determines it as an early fault symptom, and the combined parameter abnormality needs to be taken as the root node, and the constituent single parameter is taken as its chain node.

[0058] The single-parameter abnormal first indicates that if the single-parameter abnormality occurs first, the parameters affected by the association are traced back to locate the initial fault point; that is, the single-parameter abnormality appears first and then the common abnormality of the two parameters appears. Here, only the original time sequence needs to be described.

[0059] The implementation of the two-level chain node in step S23 also includes: judging whether the abnormal time interval of the current parameter pair is earlier than the single abnormal time interval of its constituent parameters, if it is earlier, taking the abnormal time interval of the current parameter pair as the root node, and taking its constituent parameters as two-level chain nodes.

[0060] If it is not earlier, the formation of the two-level chain node is not processed, and the associated parameter chain is still output in the form of time before and after according to the original processing order.

[0061] The limiting operation of step S23 here is to process the comparison of parameter pairs and single parameters. In step S1, parameters with abnormal fluctuations that are too large or significantly exceed the normal abnormal fluctuations are combined and output in the form of parameter pairs with the largest correlation. Here, it needs to be verified whether the subsequent single parameter is continuously abnormal due to the parameter pair combination abnormality, and the main parameter causing the subsequent abnormality is explained, so as to correct the associated parameter chain composed according to the time sequence in the form of relative causal chain.

[0062] In an embodiment of the present application, when performing mode clustering, the generated associated parameter chain ignores the time stamp and specific data, only retains the structure topology and node type of the chain, and clusters the formed chain. Each cluster represents a repeatedly occurring abnormal mode, such as a cooling system failure mode, a vacuum leakage mode, and the like.

[0063] The chain structure identified at this time, such as the chain structure of cooling water flow→flow-temperature relationship→molten pool temperature, will be set with an encoding, and will be clustered with the conventional abnormal chain structure in a manner similar to the figure or machine learning, to identify the abnormal pattern repeatedly appearing in the current output associated parameter chain, and the corresponding abnormal pattern will be output as a clustered propagation path in the form of a path to explain the root cause, typical propagation path and impact of each abnormal pattern. These specific clustered contents are used as paths, and then the single event analysis and correlation cumulative analysis in the chain structure can be performed for this path to explain the further correlation between abnormal events during smelting. The final output propagation path will explain each node according to the originally formed associated parameter chain, forming multiple groups of chain structures related to abnormal patterns such as cooling system failure mode, vacuum system leakage mode, etc. These chain structures are regarded as the output propagation path.

[0064] The implementation of step S3 includes setting a node type for each chain node in the associated parameter chain and recording the structure topology of the current associated parameter chain.

[0065] The associated parameter chain corresponding node type and structure topology are used for abnormal pattern clustering. Each chain structure in the associated parameter chain is regarded as input data and is converted into a vector in the form of an encoding. The vector contains node type and structure topology information. The cosine similarity is used to calculate the similarity between any two chain structures in the current associated parameter chain. The hierarchical clustering method is used. First, each chain structure is regarded as an initial cluster. Then, the two clusters with the highest similarity are merged to form a new cluster. The similarity between the new cluster and other clusters is calculated, and the merging is repeated until the preset number of clusters or the similarity between clusters is lower than the threshold value. The multiple groups of chain structures that complete pattern clustering are used as the output propagation path.

[0066] At this time, the corresponding chain structure is converted into a vector form to calculate the similarity of the node type and structure topology of different chains. The similarity is used to gradually merge the clusters. At this time, the number of each cluster can be set to 5 to identify the typical abnormal pattern output by the associated parameter chain in the current scenario. As for the part with a similarity lower than the threshold value, the threshold value can be selected as 0.65, which is a commonly used threshold value, balancing the accuracy and integrity of clustering, and can form a relatively aggregated abnormal pattern.

[0067] The current chain structure is regarded as input data, and the clustering results of the abnormal patterns labeled in the historical data are compared with the current chain structure to determine the abnormal pattern corresponding to the current propagation path.

[0068] In an embodiment of the present application, step S4 calculates a score for each propagation path formed under each batch, for a propagation path generated in one smelting, to quantify the impact of the anomaly. It should be noted that when the current key indicators are calculated, the change trend of multiple propagation paths under the same abnormal mode cluster in multiple batches is analyzed based on these data.

[0069] As shown in Figure 4 S41, statistics are performed on the propagation paths under each batch, weights are set for each node in the propagation path based on the abnormal mode to which the propagation path belongs, and the single-event score corresponding to the propagation path is set in a weighted summation manner according to the abnormal degree of each node on the propagation path.

[0070] At this time, the total number of nodes in the propagation path is counted, the longer the path, the wider the range of fault conduction, and the higher the severity; and weights are set for different types of nodes according to prior knowledge or historical data; for example, based on historical fault statistics, the weight of each node on the propagation path is set according to the ratio of the frequency of the corresponding type of fault to the total frequency of faults; the abnormal degree value of each node on the propagation path is calculated by subtracting the upper limit value of the corresponding parameter in normal control from the maximum value during the abnormal period of the value represented by the node and dividing the upper limit value, and then the single-event score of the entire propagation path is calculated by using the weights to perform a weighted summation.

[0071] S42, the propagation paths of multiple batches are integrated, the average length of the propagation paths under the same abnormal mode, the maximum abnormal degree value of each propagation path during the smelting, and the abnormal mode label corresponding to the propagation path are regarded as path characteristics; and the transmission efficiency is set by using the ratio of the abnormal degree difference between the root node and the terminal node of the propagation path to the time length of the propagation path, to complete the construction of the batch-index association matrix.

[0072] Then, the single-event scores of multiple smeltings are integrated, the characteristic values of each propagation path are extracted, and the long-term change trend of the propagation path during multiple integrations is analyzed.

[0073] It should be noted that the transmission efficiency is obtained by subtracting the abnormal degree value of the last node from the abnormal degree value of the root node of the propagation path and dividing the time length of the corresponding propagation path, and the transmission efficiency measures the efficiency or ability of the abnormality from the root cause to the final effect, and the greater the value, the stronger the ability of the system to amplify the fault, which is an unhealthy performance. If the transmission efficiency tends to increase, it is a major risk signal that needs to be paid attention to and handled in a timely manner. This value is normalized before calculation, and the calculated transmission efficiency is generally small, to emphasize the deterioration of the overall system in the presence of serious abnormalities.

[0074] It should be noted that the abnormal mode label does not perform subsequent trend analysis, and is only used to label the parameters in the current formed matrix; as for the average length of the propagation path, it reflects the time and related parameters involved when each anomaly occurs, and can measure the horizontal risk impact caused by the abnormal event, and as for the maximum abnormal degree value, it measures the maximum abnormality of a single node, which belongs to the analysis of a single node to judge the relative change of the whole.

[0075] S43, using the transmission efficiency, path characteristics and single event score as the key indicator value of the batch-index correlation matrix, and calculating the change trend along the time sequence to obtain the trend line slope corresponding to each key indicator.

[0076] S44, calculate the Pearson correlation coefficient of the trend line slope corresponding to multiple key indicators, and the calculated value is regarded as the correlation of the change trend of each key indicator, and when the correlation value of the change trend of each key indicator and the trend line slope corresponding to the corresponding key indicator is consistent, the corresponding abnormal event is taken as the output transmission relationship.

[0077] The row of the batch-index correlation matrix represents the batch ID of each smelting, and the list represents the indicators to be monitored, such as the single event score of the current smelting, the average length of the propagation path under the same abnormal mode during the current smelting, the maximum abnormal degree value of each propagation path during the current smelting and the transmission efficiency of the corresponding abnormality, etc. These indicators are used as key indicators for subsequent analysis.

[0078] For the key indicator value of each column of the batch-index correlation matrix, calculate its change trend along the time sequence, use linear regression to calculate its trend line, and according to the slope of the corresponding indicator, explain the change trend at the corresponding abnormal time during the current titanium alloy smelting, and finally perform correlation analysis on the change trend, analyze the key indicators one by one through the form of calculating the Pearson correlation coefficient, and obtain the correlation value.

[0079] As for the consistency, it means the consistency of the trend corresponding to the current key indicator, the path consistency of the propagation path and the stability of the transmission efficiency. The data that meets the analysis of the three aspects will be output as the transmission relationship at this time.

[0080] The path consistency indicates multiple occurrences of the same abnormal mode, whether the shape of the propagation path is stable, that is, by comparing the graph edit distance of each node in the flat propagation path, in the current abnormal mode, the graph edit distance of multiple propagation paths is small, and the graph edit distance standard deviation is low. In the scene, it is indicated that the abnormal situation existing in the current propagation path is relatively stable, which belongs to the relatively stable abnormal event obtained by the current analysis; as for the judgment standard of small graph edit distance and low graph edit distance standard deviation, the average value calculated in the current abnormal mode in the historical data is taken as the comparison standard, that is, when the graph edit distance and the graph edit distance standard deviation are less than the average value in the historical data, it is considered that the high consistency of the path consistency is high. If these values are greater than the average value of the historical data, it is indicated that the path consistency is low, and the specific values are marked and output according to the situation.

[0081] As for the trend consistency, it indicates that the current correlation value belongs to high correlation, and the change trend of the two key indicators calculated is consistent, such as both greater than 0 or both less than 0; it is indicated that the current correlation value belongs to high correlation, and the calculated value of the Pearson correlation coefficient is greater than 0.7, the output data is relatively small and there is high correlation; the key indicators meeting the trend consistency will represent a clear change direction of the current abnormal event, and a relatively accurate result can be obtained according to the trend change; otherwise, the trend change will be relatively chaotic, and the situation of low consistency will be marked. For example, the slope of the transmission efficiency is less than 0 and the slope of the single event score is less than 0, indicating that the duration of the abnormal event may be shortened, and the value of the single abnormal event may be relatively small. These data will represent the relative relationship of the abnormal event under multiple transmissions, indicating the specific situation of the current smelting event when the abnormal event occurs.

[0082] As for the transmission efficiency stability, the standard deviation and the coefficient of variation of the transmission efficiency of the corresponding key indicators are required to be small, and the coefficient of variation is represented as the ratio of the standard deviation and the average value. When the standard deviation and the coefficient of variation are small, it is indicated that the characteristics of each fault propagation are very consistent, which belongs to a high consistency abnormality. If the values of the two are large, it is indicated that the abnormality is difficult to predict, which belongs to a part with high risk, and belongs to a low consistency abnormality. These data are output according to the consistency, as the transmission relationship at this time.

[0083] In the current processing process, the processing methods of the single parameter and the parameter pair are the same, and the key indicators in the batch-index association matrix are processed. The abnormality degree value of the parameter pair is calculated by the vector modulus method, and the abnormality degree values of the two parameters in the parameter pair are calculated to more accurately reflect the abnormality degree of the corresponding parameter pair in the current smelting event.

[0084] Step S5 obtains data of parameter combination-event sequence-quality result through the transmission relationship of abnormal events, which represents the parameter combination under the related abnormality in smelting, so as to facilitate the monitoring of the occurrence of smelting events and improve the processing effect of titanium alloy smelting.

[0085] The implementation mode of step S5 includes: determining the combination form of single parameters and parameter pairs corresponding to smelting process parameters based on the transmission relationship of abnormal events in multiple smeltings, and taking the corresponding single parameters and parameter pairs as the target parameters of output according to the abnormal events pointed by the smelting process parameters.

[0086] At this time, the single parameters and parameter pairs corresponding to the current abnormal event are found out through simple backtracking, and these data are taken as the target parameters for subsequent monitoring, so as to complete the intelligent analysis and monitoring of the smelting effect of titanium alloy in multiple batches, thereby improving the subsequent processing efficiency of titanium alloy.

[0087] As shown in Figure 5 The present application also provides a titanium alloy smelting effect intelligent analysis system, which comprises a smelting event module, a smelting correlation module, a propagation path module, a transmission analysis module and a parameter output module; wherein the output end of the smelting event module is connected with the smelting correlation module, the output end of the smelting correlation module is connected with the propagation path module, the output end of the propagation path module is connected with the transmission analysis module, and the output end of the transmission analysis module is connected with the parameter output module.

[0088] The smelting event module is used for acquiring smelting process parameters, mapping the smelting process parameters to a smelting device, and recording smelting events of the smelting device in a preset time period.

[0089] The smelting correlation module is used for recording abnormal events in each smelting in the form of combination of single parameters and parameter pairs, forming a correlation parameter chain.

[0090] The propagation path module is used for performing mode clustering on the correlation parameter chain according to the combined single parameters and parameter pairs, and determining the propagation path after clustering.

[0091] The transmission analysis module is used for performing single event scoring on abnormal events of each batch based on the propagation path, calculating the transmission efficiency, single event scoring and path characteristics of each batch based on the correlation parameter chain, constructing a batch-index correlation matrix, recording the change trend of each key index in the batch-index correlation matrix in multiple smeltings, and obtaining the transmission relationship of abnormal events in multiple smeltings by analyzing the correlation of the change trends of the key indexes.

[0092] The parameter output module is used for tracing the smelting events based on the transmission relationship of abnormal events in multiple smeltings, determining the target parameters of each batch, and outputting the target parameters.

[0093] Although the embodiments of the present application have been shown and described above, it is understood that the above-described embodiments are exemplary and are not to be construed as limiting the present application, and that changes, modifications, substitutions and variations can be made by those skilled in the art without departing from the scope of the present application, and still fall within the protection scope of the present application.

Claims

1. A titanium alloy smelting effect intelligent analysis method, characterized in that, Comprise: S1, obtain smelting process parameters, map the smelting process parameters to the smelting equipment, record the smelting events of the smelting equipment in a preset time period; S2, record the abnormal events of each smelting in the form of single parameter and parameter pair combination, form the associated parameter chain; S3, the associated parameter chain is clustered according to the form of combined single parameter and parameter pair, and the propagation path after clustering is determined; S4, based on the propagation path, the single event score of each batch of abnormal events is calculated, the transmission efficiency, single event score and path feature of each batch are calculated based on the associated parameter chain, and the batch-index correlation matrix is constructed; record the change trend of each key index in the batch-index correlation matrix in multiple smelting, and obtain the transmission relationship of abnormal events in multiple smelting by analyzing the correlation of the change trend of each key index; S5, based on the transmission relationship of abnormal events in multiple smelting, the smelting events are traced back, the target parameters of each batch are determined, and the target parameters are output.

2. The method for intelligent analysis of titanium alloy smelting effect according to claim 1, characterized in that The implementation of step S1 comprises: S11, based on the mapped smelting equipment, determine the smelting boundary of the current smelting process parameters, and the risk exposure time and the risk mitigation time of the smelting process parameters reaching the smelting boundary on each smelting equipment; S12, the average value and standard deviation of the smelting process parameters in the risk exposure time and the risk mitigation time are calculated, the stability of the smelting process parameters in each smelting is determined, and the smelting event is extracted when the stability of the smelting process parameters meets the requirements; S13, when the stability of the smelting process parameters does not meet the requirements, the smelting process parameters are divided into multiple stages, the smelting process parameters are associated according to the stages, the correlation of the smelting process parameters in different stages is determined, and the smelting process parameters with the maximum correlation are output as the extracted smelting event.

3. The method according to claim 2, wherein The implementation of step S12 further comprises: S121, based on the smelting demand information of titanium alloy, the preset smelting stability threshold is used to analyze the current smelting process parameters; compare the preset smelting stability threshold with the average value and standard deviation of the current smelting process parameters according to the processing flow of titanium alloy smelting; S122, if any one of the average value or the standard deviation of the current smelting process parameters exceeds the upper and lower limit range of the preset smelting stability threshold, it is considered that the stability of the smelting process parameters does not meet the requirements; S123, only when the average value and the standard deviation of the current smelting process parameters are both within the upper and lower limit range of the preset smelting stability threshold, it is considered that the stability of the smelting process parameters meets the requirements.

4. The method for intelligent analysis of titanium alloy smelting effect according to claim 2, characterized in that, The implementation of step S13 further comprises: The smelting process parameters in each stage are paired to form parameter pairs, the Pearson correlation coefficients of all parameter pairs are calculated, and the correlation of the parameter pairs in each stage is obtained; For each stage, find the parameter pair with the maximum correlation, which is considered as the correlation result of the corresponding stage; Compare the correlation results of all stages to determine the global maximum correlation parameter pair and output it as the smelting event.

5. The method for intelligent analysis of titanium alloy smelting effect according to claim 1, characterized in that, The implementation of step S2 comprises: S21, based on the abnormal event at each smelting, the smelting process parameters associated with the abnormal event are extracted in the form of single parameters and parameter pairs, and each monitored single parameter and parameter pair is regarded as an abnormal node; S22, for each abnormal node, the time period when the abnormal event occurs is counted, and the continuously occurring abnormal nodes are merged into an abnormal time interval, and the abnormal time interval contains at least one abnormal node; S23, for each abnormal time interval, check whether there is an intersection between the abnormal time interval composed of all other abnormal nodes and the current abnormal time interval, if there is an intersection, the abnormal nodes in the current abnormal time interval are regarded as root nodes, and the other abnormal nodes are regarded as secondary chain nodes of the abnormal nodes in the current abnormal time interval; S24, if there is no intersection, the current abnormal time interval is formed into an output associated parameter chain in the form of time stamp advancement; S25, all secondary chain nodes are counted, and tertiary chain nodes under the secondary chain nodes are found according to the way that the abnormal time intervals have intersections, and the tertiary chain nodes are recursively found step by step until there is no intersection between the abnormal time intervals, and the formed chain nodes and corresponding abnormal nodes are taken as the output associated parameter chain.

6. The method for intelligent analysis of titanium alloy smelting effect according to claim 5, characterized in that, The implementation mode of setting secondary chain nodes in step S23 further includes: determine whether the abnormal time interval of the current parameter pair is earlier than the single abnormal time interval of its constituent parameter, if it is earlier, take the abnormal time interval of the current parameter pair as the root node, and take its constituent parameter as the secondary chain node.

7. The method for intelligent analysis of titanium alloy smelting effect according to claim 1, characterized in that, The implementation mode of step S3 includes: set a node type for each chain node in the associated parameter chain, and record the structure topology of the current associated parameter chain; cluster the abnormal modes according to the node type and structure topology corresponding to the associated parameter chain, and take the multiple chain structures after completing the mode clustering as the output propagation path.

8. The method for intelligent analysis of titanium alloy smelting effect according to claim 1, characterized in that, The implementation mode of step S4 includes: S41, count the propagation paths under each batch, set the weight of each node in the propagation path based on the abnormal mode to which the propagation path belongs, and set the single event score corresponding to the propagation path in the weighted summation mode according to the abnormal degree of each node on the propagation path; S42, integrate the propagation paths of multiple batches, take the average length of the propagation path under the same abnormal mode, the maximum abnormal degree value of each propagation path at the time of smelting, and the abnormal mode label corresponding to the propagation path as the path feature; and set the transmission efficiency by using the ratio of the abnormal degree difference between the root node and the terminal node of the propagation path to the time length of the propagation path, to complete the construction of the batch-index association matrix; S43, use the transmission efficiency, path feature and single event score as the key indicator value of the batch-index association matrix, and calculate the trend along the time sequence to obtain the trend line slope corresponding to each key indicator; S44, calculate the Pearson correlation coefficient of the slope of the trend line corresponding to the plurality of key indicators, and the calculated value is regarded as the correlation of the change trend of each key indicator. When the correlation value of the change trend of each key indicator and the slope of the trend line corresponding to the corresponding key indicator is consistent, the corresponding abnormal event is regarded as the output of the transmission relationship.

9. The method for intelligent analysis of titanium alloy smelting effect according to claim 1, characterized in that, The implementation mode of step S5 includes: Based on the transmission relationship of abnormal events in multiple smelting, determine the single parameter and parameter pair combination form corresponding to the smelting process parameters, and according to the abnormal event pointed by the smelting process parameter, the corresponding single parameter and parameter pair are taken as the output target parameter.

10. A titanium alloy smelting effect intelligent analysis system, characterized in that, Comprise: The smelting event module is used for acquiring the smelting process parameters, mapping the smelting process parameters to the smelting equipment, and recording the smelting events of the smelting equipment in the preset time period; The smelting correlation module is used for recording the abnormal events in each smelting by the combination of single parameter and parameter pair, forming the correlation parameter chain; The propagation path module is used for mode clustering of the correlation parameter chain according to the combined single parameter and parameter pair form, and determining the clustered propagation path; The transmission analysis module is used for single event scoring of each batch of abnormal events based on the propagation path, calculating the transmission efficiency, single event scoring and path characteristics of each batch based on the correlation parameter chain, and constructing the batch-index correlation matrix; record the change trend of each key indicator in the batch-index correlation matrix in multiple smelting, and obtain the transmission relationship of abnormal events in multiple smelting by analyzing the correlation of the change trend of each key indicator; The parameter output module is used for tracing the smelting events based on the transmission relationship of abnormal events in multiple smelting, determining the target parameter of each batch, and outputting the target parameter.

Images