A titanium alloy casting defect detection system based on laser ultrasound

Abstract

The invention discloses a titanium alloy casting defect detection system based on laser ultrasound, which relates to the technical field of material detection. The method comprises: a module for constructing a three-dimensional model of a titanium alloy casting containing surface coordinates and casting control parameters, a module for partitioning the casting surface coordinates according to the casting control parameters, a module for traversing the partition queue to predict the defect probability and obtain a probability queue, a module for selecting a target partition that meets a threshold from the partition queue based on the probability queue, a module for detecting a first signal of the target partition according to preset parameters using a laser ultrasonic transmitter, a module for extracting a second signal pre-stored according to preset parameters of the target partition, and a module for analyzing the detection result of the first signal according to the second signal. The system solves the technical problems that the existing titanium alloy casting defect detection does not perform partition processing, resulting in blind detection, inefficiency, insufficient precision and unreasonable resource allocation, and achieves the technical effect of improving detection efficiency and accuracy and ensuring casting quality and safety.

Description

Technical Field

[0001] The present application relates to the technical field of material detection, and in particular to a titanium alloy casting defect detection system based on laser ultrasound. Background Art

[0002] In the defect detection scenario of titanium alloy castings, the problem of ensuring the quality of castings is particularly prominent, and the contradiction between the demand for accurate and efficient detection is relatively more prominent. Realizing accurate detection in different areas and better meeting the detection needs of different areas have become a crucial link in solving the quality of titanium alloy casting detection. Traditional defect detection of titanium alloy castings is often relatively general and lacks specificity. It only relies on general detection methods, lacks overall control over the inside of the casting, lacks in-depth analysis and utilization of casting processes and material properties, and is difficult to comprehensively and accurately assess the defect risk status of each area. There are unreasonable situations in the allocation of detection resources, resulting in insufficient detection of some key areas and excessive detection of some areas. The formulation of detection plans is relatively simple and fixed, and cannot cope well with the complex and diverse actual conditions of titanium alloy castings.

[0003] In the current relevant technologies, there are technical problems that the defect detection of titanium alloy castings is not partitioned, resulting in blind, inefficient, insufficiently precise detection and unreasonable resource allocation. Summary of the invention

[0004] The present application provides a titanium alloy casting defect detection system based on laser ultrasound, which adopts a module for constructing a three-dimensional model of a titanium alloy casting containing surface coordinates and casting control parameters, a module for partitioning the casting surface coordinates according to the casting control parameters, a module for traversing the partition queue to predict the defect probability to obtain a probability queue, a module for selecting a target partition that meets the threshold from the partition queue based on the probability queue, a module for detecting a first signal of the target partition according to preset parameters using a laser ultrasonic transmitter, a module for extracting a second signal pre-stored in the target partition according to preset parameters, and a module for analyzing the detection result of the first signal according to the second signal. The system achieves the technical effect of accurately locating titanium alloy casting defects and optimizing detection resource allocation through partition detection, thereby improving detection efficiency and accuracy and ensuring the quality and safety of castings.

[0005] The present application provides a titanium alloy casting defect detection system based on laser ultrasound, comprising:

[0006] A three-dimensional model construction module, the three-dimensional model construction module is used to construct a three-dimensional model of a titanium alloy casting, wherein the three-dimensional model of the titanium alloy casting has surface coordinates of the titanium alloy casting and casting control parameters; a partitioning module, the partitioning module is used to partition the surface coordinates of the titanium alloy casting according to the casting control parameters to obtain a titanium alloy casting partition queue; a defect distribution probability queue acquisition module, the defect distribution probability queue acquisition module is used to traverse the titanium alloy casting partition queue to perform defect probability prediction and obtain a defect distribution probability queue; a target partition sorting module, the target partition sorting module is used to select the target partition from the titanium alloy casting based on the defect distribution probability queue. A partition queue selects target partitions of titanium alloy castings that meet the distribution probability threshold; a first ultrasonic detection signal acquisition module, the detection signal acquisition module is used to detect the target partition of the titanium alloy casting through a laser ultrasonic transmitter according to preset control parameters to obtain a first ultrasonic detection signal; a second ultrasonic detection signal acquisition module, the second ultrasonic detection signal acquisition module is used to extract the second ultrasonic detection signal pre-stored in the target partition of the titanium alloy casting according to the preset control parameters; a defect detection result acquisition module, the defect detection result acquisition module is used to perform defect analysis on the first ultrasonic detection signal according to the second ultrasonic detection signal to obtain a defect detection result.

[0007] A titanium alloy casting defect detection system based on laser ultrasound proposed in this application first constructs a module of a three-dimensional model of a titanium alloy casting containing surface coordinates and casting control parameters, a module of a partition queue for partitioning the casting surface coordinates according to the casting control parameters, a module for traversing the partition queue to predict the defect probability to obtain a probability queue, a module for selecting a target partition that meets the threshold from the partition queue based on the probability queue, a module for detecting a first signal of the target partition according to preset parameters using a laser ultrasonic transmitter, a module for extracting a second signal pre-stored in the target partition according to preset parameters, and a module for analyzing the detection result of the first signal according to the second signal. Through partition detection, the defects of titanium alloy castings can be accurately located and the allocation of detection resources can be optimized, thereby achieving the technical effect of improving detection efficiency and accuracy and ensuring the quality and safety of castings. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In order to more clearly illustrate the technical solution of the embodiment of the present invention, the accompanying drawings of the embodiment of the present invention will be briefly introduced below. A flow chart is used in the present application to illustrate the operations performed by the system according to the embodiment of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps can be processed in reverse order or simultaneously as needed. At the same time, other operations can also be added to these processes, or one or more operations can be removed from these processes.

[0009] Figure 1A schematic diagram of the structure of a titanium alloy casting defect detection system based on laser ultrasound provided in an embodiment of the present application;

[0010] Figure 2 A schematic diagram of the structure of a defect detection result acquisition module of a titanium alloy casting defect detection system based on laser ultrasound provided in an embodiment of the present application.

[0011] Explanation of the accompanying drawings: three-dimensional model construction module 10, partition module 20, defect distribution probability queue acquisition module 30, target partition sorting module 40, first ultrasonic detection signal acquisition module 50, second ultrasonic detection signal acquisition module 60, defect detection result acquisition module 70. DETAILED DESCRIPTION

[0012] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below.

[0013] In order to make the purpose, technical scheme and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present application. In the following description, "some embodiments" are involved, which describe a subset of all possible embodiments, but it can be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments, and can be combined with each other without conflict. The term "first\second" involved is only to distinguish similar objects and does not represent a specific ordering for objects. The terms "including" and "having" and any variations are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units does not have to be limited to those steps or units that are clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by technicians in the technical field of the present application. The terms used herein are only for the purpose of describing the embodiments of the present application.

[0014] The present application embodiment provides a titanium alloy casting defect detection system based on laser ultrasound, such as Figure 1 As shown, the method includes:

[0015] The three-dimensional model construction module 10 is used to construct a three-dimensional model of a titanium alloy casting, wherein the three-dimensional model of the titanium alloy casting has the surface coordinates and casting control parameters of the titanium alloy casting. Specifically, the three-dimensional model construction module 10 performs a comprehensive characteristic analysis on the original data of the titanium alloy casting, identifies key elements such as the shape characteristics, surface characteristics, and casting control parameters of the casting, uses professional data processing and model construction technology, extracts various details of the titanium alloy casting, and analyzes the association and dependency between them. Based on the results of the characteristic analysis, the key information for model construction is determined, and the key information includes information such as the acquisition method of the surface coordinates, the integration method of the casting control parameters, and the special processing methods that may be required (such as surface fitting, precision control, etc.), which are used to construct an accurate three-dimensional model. The model features of the obtained key information are extracted and classified, and the model features refer to the key factors that affect the effect of the three-dimensional model construction, such as coordinate accuracy, parameter integrity, model complexity, etc., and the information with similar model features is classified into one category through a classification algorithm. According to the classification results and model requirements, the corresponding 3D model construction method is configured. Each construction method represents a group of operations with the same or similar model features, and they share the same construction parameters and strategies.

[0016] Partitioning module 20, the partitioning module 20 is used to partition the surface coordinates of the titanium alloy casting according to the casting control parameters to obtain the titanium alloy casting partition queue. Specifically, the partitioning module 20 performs a detailed characteristic analysis on the titanium alloy casting, identifies key information such as the surface coordinates and casting control parameters of the casting, uses data processing and analysis technology to extract the characteristics of each region of the casting, and analyzes the association and difference between them. Based on the results of the characteristic analysis, the partition information is determined, and the partition information includes the partition range, partition boundary, partition order, and special processing methods that may be required (such as key focus areas, mergeable areas, etc.) and other information for subsequent detection operations. The key features of the obtained partition information are extracted and clustered, and the key features refer to the key factors that affect the partition effect, such as casting parameter differences, surface complexity, etc. Through the clustering algorithm, the regions with similar key features are classified into one category, and the corresponding partition categories are configured according to the clustering results and the characteristics of the regions. Each partition category represents a group of regions with the same or similar features.

[0017] In a possible implementation, the partitioning module 20 includes: a record configuration unit, which is used to traverse the casting control attribute set and configure the casting control attribute partition threshold based on the casting defect historical detection record. Specifically, the record configuration unit traverses the casting control attribute set one by one, analyzes and reviews each casting control attribute in the set, and configures the partition threshold for each casting control attribute based on the casting defect historical detection record, which contains the relevant control attribute data when the defect is found in the past detection, and based on the rich historical experience and data accumulation, the threshold plays a key role, which determines that only when the specific casting control attribute value exceeds this threshold, the corresponding area will be identified as requiring subsequent partitioning operations.

[0018] A partition result acquisition unit is used to traverse the casting control parameters and the casting control attribute partition thresholds, partition the surface coordinates of the titanium alloy castings, and obtain a number of titanium alloy casting partition results. Specifically, the partition result acquisition unit analyzes the casting control parameters, combines the parameters with the casting control attribute partition thresholds determined by the previous record configuration unit, and performs partitioning work on the surface coordinates of the titanium alloy castings, and finally obtains a number of titanium alloy casting partition results.

[0019] A partition queue acquisition unit is used to perform regional fusion segmentation on the partition results of the plurality of titanium alloy castings to obtain the partition queue of the titanium alloy castings. Specifically, the partition queue acquisition unit receives the plurality of titanium alloy casting partition results generated by the partition result acquisition unit, and performs regional fusion segmentation processing on the partition results. For example, if there is a situation in which the division of some regions overlaps or can be merged in different partition results, a corresponding fusion or segmentation operation will be performed, and finally a unified and reasonable titanium alloy casting partition queue is obtained, which provides an accurate and effective data basis for subsequent detection and analysis.

[0020] In a possible implementation, the record configuration unit includes: a coefficient set acquisition subunit, the coefficient set acquisition subunit is used to perform casting defect correlation analysis on the casting control attribute set to obtain a casting control attribute correlation coefficient set. Specifically, the coefficient set acquisition subunit analyzes the casting control attribute set, accurately calculates the correlation between each casting control attribute and the casting defect, for example, through statistical analysis, regression analysis and other means, considers how the change of a certain attribute affects the probability of occurrence of casting defects, and finally obtains a casting control attribute correlation coefficient set, each coefficient in the set clearly reflects the closeness or correlation strength between the corresponding casting control attribute and the casting defect. An attribute sorting result subunit, the attribute sorting result subunit is used to sort the casting control attribute sorting results that meet the correlation coefficient threshold from the casting control attribute set according to the casting control attribute correlation coefficient set. Specifically, the attribute sorting result subunit sets a reasonable correlation coefficient threshold based on the casting control attribute correlation coefficient set obtained by the coefficient set acquisition subunit. The role of the threshold is to screen out casting control attributes with significant correlation, and compare the size relationship between each coefficient in the set and the threshold. When the correlation coefficient of a casting control attribute exceeds the set threshold, it will be selected and included in the casting control attribute sorting results. The sorting operation aims to select those attributes that are highly correlated with casting defects. Only these attributes are more valuable and meaningful, and can play a key role in the subsequent partitioning process, thereby effectively reducing the amount of calculation, avoiding the interference of irrelevant or weakly correlated attributes, and significantly improving the accuracy and effectiveness of partitioning.

[0021] Partition threshold configuration subunit, the partition threshold configuration subunit is used to configure the casting control attribute partition threshold for the casting control attribute sorting result according to the casting defect historical detection record. Specifically, the partition threshold configuration subunit focuses on the casting defect historical detection record. The record is a treasure trove of valuable experience and data accumulated in past detection work. In-depth research and analysis of the record obtain detailed information about the occurrence of defects. Based on historical data, the casting control attribute sorting results screened out by the attribute sorting result subunit are configured with partition thresholds. For example, according to the correspondence between specific attributes and defect occurrence in historical records, determine within which attribute value range the partition operation needs to be performed. Through threshold configuration, the area that needs to be partitioned can be more accurately delineated, further improving the accuracy and efficiency of partitioning, and providing a reliable decision-making basis for the entire partitioning work.

[0022] In a possible implementation, the coefficient set acquisition subunit includes: a first casting control attribute extraction microunit, which is used to traverse the casting control attribute set and extract the first casting control attribute. Specifically, the first casting control attribute extraction microunit checks each attribute in the casting control attribute set one by one, and selects the first casting control attribute that needs to be analyzed in depth. For example, assuming that the casting control attribute set includes attributes such as "casting temperature", "casting pressure", and "casting time", the first casting control attribute extraction microunit may first extract "casting temperature" as the first casting control attribute.

[0023] The frequency recording micro-unit is used to take the deviation between the monitoring characteristic value and the expected characteristic value of the first casting control attribute as the only variable, and after determining the characteristic value of the variable, take the first casting control attribute deviation and the titanium alloy casting model as constraints, collect a preset number of defect monitoring logs, count the defect detection record frequencies of the preset number of defect monitoring logs, and add them to the first casting control attribute deviation sequence and the first defect detection record frequency sequence, wherein the first casting control attribute deviation corresponds to the first defect detection record frequency one by one. Specifically, the deviation between the monitoring characteristic value and the expected characteristic value of the first casting control attribute extracted by the frequency recording micro-unit is set as the only variable, and after determining the specific characteristic value of the variable, take the deviation and the titanium alloy casting model as constraints, collect a preset number of defect monitoring logs, for example, set to collect 100 defect monitoring logs, count the logs, calculate the record frequency of defects detected in the monitoring data, and add the calculated frequencies to the first casting control attribute deviation sequence and the first defect detection record frequency sequence in sequence, and ensure that each deviation value can accurately correspond to a defect detection record frequency.

[0024] The analysis function constructs a micro-unit, and the analysis function constructs a micro-unit for constructing a correlation coefficient analysis function:

[0025] ,in Characterize the correlation coefficient of any attribute, represents the i-th deviation value, represents the frequency of the i-th detected record, and N represents the number of sequences. Specifically, the analysis function constructs a micro-unit for constructing a mathematical function = Represents the correlation coefficient of any attribute, represents the i-th deviation value, represents the frequency of the i-th detected record, N represents the number of data in the sequence, and the function first calculates each and The difference between the two values ​​and their respective means, multiply the differences and sum them, square the sum of the squares of the differences respectively, and multiply them as the denominator. The final ratio is the correlation coefficient. It is used to measure the degree of linear correlation between two variables (deviation value and detection record frequency), providing strong mathematical support and guarantee for the subsequent accurate calculation of the correlation between casting control attributes and defect detection record frequency.

[0026] The first casting control attribute correlation coefficient acquisition micro unit is used to perform correlation analysis on the first casting control attribute deviation sequence and the first defect detection record frequency sequence according to the correlation coefficient analysis function to obtain the first casting control attribute correlation coefficient. Specifically, the first casting control attribute correlation coefficient acquisition micro unit uses the constructed analysis function to substitute the data in the first casting control attribute deviation sequence and the first defect detection record frequency sequence into the function for calculation and analysis, and finally obtains the correlation coefficient that can quantitatively describe the degree of correlation between the first casting control attribute and the defect detection record frequency.

[0027] The micro-unit for adding the correlation set is used to add the first casting control attribute correlation coefficient into the casting control attribute correlation set. Specifically, the micro-unit for adding the correlation set adds the first casting control attribute correlation coefficient obtained previously to the casting control attribute correlation set, which is used to store and manage various casting control attribute correlation coefficients. By continuously adding new correlation coefficients, a comprehensive data basis is provided for the subsequent comprehensive evaluation of the correlation between multiple casting control attributes and defects.

[0028] In a possible implementation, the partition threshold configuration subunit includes: a basic data analysis microunit, which is used for basic data analysis that the first casting control attribute deviation sequence and the defect detection record frequency sequence belong to the correlation coefficient analysis basic data. Specifically, the basic data analysis microunit receives two sets of data, the first casting control attribute deviation sequence and the defect detection record frequency sequence, and uses statistical analysis, trend analysis, etc. to analyze the intrinsic characteristics of the two sets of data, for example, calculating the mean, variance, standard deviation and other statistical quantities of the data, observing whether the distribution form of the data is normal distribution or skewed distribution, and checking whether the data has a periodic or seasonal change trend. Through analysis, a solid data foundation and feature description are provided for the subsequent processing of other units.

[0029] The attribute selection micro-unit is used to extract the first casting control selected attribute according to the casting control attribute sorting result. Specifically, the attribute selection micro-unit obtains the casting control attribute sorting result generated previously, including many attributes that have been preliminarily screened, and accurately extracts the first casting control selected attribute from the attributes according to specific rules or conditions pre-set by the user. The selected attribute is more closely related to the quality of the casting, the formation of defects, etc., and plays a key role in the subsequent partition threshold determination process.

[0030] A calibration deviation threshold micro-unit is used to calibrate the deviation threshold of the first casting control selected attribute for the first casting control selected attribute through the user terminal. Specifically, the calibration deviation threshold micro-unit receives user-defined settings through the user terminal. The user determines a deviation threshold for the first casting control selected attribute based on actual production experience, process requirements, and expectations for casting quality. The threshold is usually regarded as a tolerance deviation, which means that if the deviation between the actual measured value and the ideal value is less than this threshold, it will be considered to be within an acceptable range and regarded as having the same characteristic value, and no special processing or partitioning is required.

[0031] The positive partition threshold selection micro-unit is used to perform positive deviation analysis on the first casting control selected attribute based on the first casting control selected attribute deviation threshold, and obtain the first casting control selected attribute positive partition threshold. Specifically, the positive partition threshold selection micro-unit is based on the deviation threshold calibrated by the user, and analyzes the positive deviation of the first casting control selected attribute. The positive deviation refers to the situation where the actual value is greater than the ideal value. Since the positive deviation may have different effects on the quality, performance or defects of the casting, it needs to be studied separately. By calculating and comparing the defect detection record frequencies corresponding to different positive deviation values, a suitable positive partition threshold is determined. Positive deviations exceeding the threshold need to be divided into different areas for special processing.

[0032] The negative partition threshold selects the micro-unit, and the negative partition threshold selects the micro-unit for performing negative deviation analysis on the first casting control selected attribute based on the first casting control selected attribute deviation threshold, and obtains the negative partition threshold of the first casting control selected attribute. Specifically, the negative partition threshold selects the micro-unit similar to the positive partition threshold selects the micro-unit, for the negative deviation of the first casting control selected attribute, that is, the actual value is less than the ideal value, because the negative deviation may lead to different results from the positive deviation, such as it may affect the strength, durability and other aspects of the casting, so it also needs to be discussed and analyzed separately by classification, and the negative partition threshold is obtained by calculation to distinguish the negative deviation that requires special treatment.

[0033] A negative partition threshold storage micro-unit is used to store the first casting control selected attribute positive partition threshold and the first casting control selected attribute negative partition threshold as the first casting control selected attribute partition threshold, and add them into the casting control attribute partition threshold. Specifically, the negative partition threshold storage micro-unit organizes and stores the positive partition threshold and the negative partition threshold obtained previously, merges the two thresholds into the partition threshold of the first casting control selected attribute, and adds them to the set of casting control attribute partition thresholds. In subsequent partition operations and related quality control, defect detection and other processes, these partition thresholds can be conveniently called and applied to ensure the accuracy and effectiveness of partitioning and improve the quality and efficiency of casting production.

[0034] In a possible implementation, the positive partition threshold selection micro-unit includes: a gradient sequence construction nano-unit, the gradient sequence construction nano-unit is used to take the positive deviation between the monitored characteristic value and the expected characteristic value of the first casting control selected attribute as the only variable, and based on the first casting control selected attribute deviation threshold, construct a positive deviation gradient sequence. Specifically, the gradient sequence construction nano-unit obtains the monitored characteristic value and the expected characteristic value of the first casting control selected attribute, calculates the positive deviation between them, that is, the part where the monitored characteristic value is greater than the expected characteristic value, and takes the positive deviation as the core variable. Assuming that the first casting control selected attribute deviation threshold set by the user is T, the gradient sequence construction nano-unit will start from the smallest positive deviation and gradually increase with a certain step size until it reaches or exceeds the deviation threshold T, and construct a series of ordered positive deviation values ​​to form a positive deviation gradient sequence. For example, the initial positive deviation may be 0.1, with a step size of 0.05, up to a deviation value of 1.5.

[0035] The second defect detection record frequency sequence calibration unit is used to traverse the forward deviation gradient sequence and calibrate the second defect detection record frequency sequence, wherein the second defect detection record frequency sequence is configured in the same manner as the first defect detection record frequency sequence. Specifically, the second defect detection record frequency sequence calibration unit sequentially checks each forward deviation value in the forward deviation gradient sequence, and for each forward deviation value, the second defect detection record frequency sequence calibration unit determines the defect detection record frequency corresponding to this deviation value according to specific rules and existing data, traverses the entire forward deviation gradient sequence, and calibrates a corresponding defect detection record frequency for each deviation value, thereby constructing a second defect detection record frequency sequence, wherein the configuration method is the same as the first defect detection record frequency sequence, which means that when determining the defect detection record frequency corresponding to each deviation value, the method used, the data source based on it, and the calculation method, etc. are all consistent, so as to ensure the consistency and comparability of the data.

[0036] A sequence interception unit, the sequence interception unit is used to intercept the positive deviation gradient sequence from the second defect detection record frequency sequence based on the defect detection record frequency threshold to obtain the first casting control selected attribute positive partition threshold. Specifically, the sequence interception unit finds a key segmentation point from the second defect detection record frequency sequence based on a preset defect detection record frequency threshold, and then determines the interception position of the positive deviation gradient sequence to obtain the first casting control selected attribute positive partition threshold.

[0037] The detection record frequency extraction unit is used to extract the kth detection record frequency and the k+1th detection record frequency according to the second defect detection record frequency sequence. Specifically, the detection record frequency extraction unit first extracts two adjacent frequencies, namely, the kth detection record frequency and the k+1th detection record frequency, from the second defect detection record frequency sequence.

[0038] A positive deviation gradient sequence interception unit is used to intercept the positive deviation gradient sequence based on the k+1th sequence number detection record frequency when the k+1th sequence number detection record frequency is greater than or equal to the defect detection record frequency threshold, and the kth sequence number detection record frequency is less than the defect detection record frequency threshold, so as to obtain the first casting control selected attribute positive partition threshold. Specifically, the positive deviation gradient sequence interception unit will compare these two frequencies with the defect detection record frequency threshold. If the k+1th detection record frequency is greater than or equal to the defect detection record frequency threshold, and the kth detection record frequency is less than the threshold, then the positive deviation value corresponding to the k+1th detection record frequency is used as a reference to intercept in the positive deviation gradient sequence. For example, if the 5th frequency is less than the threshold and the 6th frequency is greater than or equal to the threshold, then the positive deviation value corresponding to the 6th frequency is used as the interception point. The positive deviation value represented by this interception point is the final positive partition threshold of the first casting control selected attribute. The threshold can help accurately distinguish situations that require different processing methods and provide a clear basis for subsequent work.

[0039] The defect distribution probability queue acquisition module 30 is used to traverse the titanium alloy casting partition queue to predict the defect probability and obtain the defect distribution probability queue. Specifically, the defect distribution probability queue acquisition module 30 performs a comprehensive feature analysis on the titanium alloy casting partition queue, identifies the key features in the partition queue, such as the size, shape, position and other information of the partition, extracts the specific characteristics of each partition by using data processing and analysis technology, and analyzes the association and difference between them. Based on the result of feature analysis, the key information for defect probability prediction is determined, and the key information includes the area where defects may occur, the defect type and the corresponding probability, etc., which is used to generate the defect distribution probability queue. The probability features of the obtained key information are extracted and classified, and the probability features refer to the key factors that affect the defect distribution probability, such as casting process parameters, material properties, etc. Through the classification algorithm, the partitions with similar probability features are classified into one category, and the corresponding defect distribution probability categories are configured according to the classification results and the characteristics of the partitions. Each category represents a group of partitions with the same or similar defect distribution probabilities.

[0040] In one possible implementation, the defect distribution probability queue acquisition module 30 includes: a first titanium alloy casting partition extraction unit, which is used to extract the first titanium alloy casting partition according to the titanium alloy casting partition queue. Specifically, the first titanium alloy casting partition extraction unit receives a titanium alloy casting partition queue containing rich information, covering details such as the specific location coordinates, area size, shape characteristics and material properties of the partition, and operates according to pre-set screening rules or strictly follows the clear instructions given by the user. If the screening rule is to select according to the order of the position of the partition in the casting, then starting from the starting position of the queue, the position identifier of each partition will be checked in turn, so as to pick out the partition ranked first and determine it as the first titanium alloy casting partition. If the rule is based on the area size of the partition, then the unit will perform a screening on each partition in the queue. The area data of each partition are accurately calculated and compared, and the partition with the largest area is finally locked as the first titanium alloy casting partition. If the partition is extracted according to a specific identification mark or number specified by the user, the unit will carefully search for the partition that completely matches the mark or number in the entire queue, and select it as the first titanium alloy casting partition. After clarifying the screening rules, the first titanium alloy casting partition extraction unit compares and screens each partition in the titanium alloy casting partition queue, and through judgment and verification, it finally successfully and accurately extracts the first titanium alloy casting partition that fully meets the established conditions, laying the foundation for subsequent in-depth analysis and processing.

[0041] A deviation vector matrix construction unit, the deviation vector matrix construction unit is used to extract the expected values ​​of the casting control parameters and the monitoring values ​​of the casting control parameters of the first titanium alloy casting partition, and construct a normalized casting control deviation vector matrix. Specifically, after receiving the first titanium alloy casting partition provided by the first titanium alloy casting partition extraction unit, the deviation vector matrix construction unit first analyzes the expected values ​​of the casting control parameters closely related to the partition. The expected values ​​are ideal values ​​that are carefully set in advance during the design and planning stage of the product based on rigorous theoretical calculations, strict process requirements, and rich experience accumulated in the past. The control parameter standards that should be achieved by the partition under the most ideal production conditions are then collected. The monitoring values ​​of the casting control parameters obtained by real-time monitoring of the partition in the actual production process are first calculated. In order to Eliminate the possible impact of dimensional differences between different control parameters, perform fine normalization on the differences, and finally, construct the processed and adjusted differences into a normalized casting control deviation vector matrix with clear structure and accurate content in an orderly manner according to a specific matrix arrangement method planned in advance. Each element in the matrix accurately represents the degree of deviation of a specific casting control parameter from the ideal state, and can comprehensively display the deviation of the partition from the ideal state in the actual production process in an intuitive, clear and easy-to-understand way, providing key and powerful data support for subsequent quality assessment and defect prediction.

[0042] A probability queue distribution unit is used to perform historical sample retrieval on the titanium alloy casting model according to the normalized casting control deviation vector matrix, obtain the defect sample ratio, set it as the first titanium alloy casting partition defect distribution probability, and add it into the defect distribution probability queue. Specifically, after the probability queue distribution unit obtains the normalized casting control deviation vector matrix from the deviation vector matrix construction unit, it retrieves historical samples of the same titanium alloy casting model, accesses the database or data warehouse of historical sample data, and accurately finds historical samples with highly similar deviations based on the unique characteristics presented by the received deviation vector matrix. After successfully locating these similar historical samples, statistics are carried out to count and record the number of samples with defects one by one. After completing the statistics, a simple and accurate proportion calculation operation is performed, that is, the number of defective samples is divided by the total number of samples, so as to obtain the proportion of defective samples in all samples. The obtained proportion value is clearly and accurately set as the defect distribution probability of the first titanium alloy casting partition, and is added to the defect distribution probability queue according to a predetermined format. By continuously accumulating and updating the defect distribution probability information about different partitions, it provides crucial, comprehensive, detailed and valuable data support and decision-making basis for subsequent more accurate and in-depth evaluation of the quality status of the entire titanium alloy casting, accurate prediction of possible defects, and scientific and reasonable formulation of corresponding improvement and optimization measures.

[0043] The target partition sorting module 40 is used to sort the target partitions of the titanium alloy casting that meet the distribution probability threshold from the titanium alloy casting partition queue based on the defect distribution probability queue. Specifically, the target partition sorting module 40 performs in-depth data analysis on the defect distribution probability queue, identifies the defect distribution probability characteristics of each titanium alloy casting partition, extracts the probability value of each partition by using data processing and analysis technology, and analyzes the differences and associations between them. Based on the results of the data analysis, the partition sorting criteria are determined, and the partition sorting criteria include information such as distribution probability thresholds and partition characteristics, which are used to screen target partitions. The obtained partition sorting criteria are extracted and compared with features, and the features refer to key factors that affect partition screening, such as defect severity, partition location, etc. Through a comparison algorithm, partitions with similar features are classified into one category, and the target partitions of the titanium alloy casting that meet the distribution probability threshold are sorted according to the comparison results and the characteristics of the partitions. Each target partition represents a group of partitions with the same or similar importance, requiring the same or similar subsequent processing methods.

[0044] The first ultrasonic detection signal acquisition module 50 is used to detect the target partition of the titanium alloy casting according to the preset control parameters through the laser ultrasonic transmitter to obtain the first ultrasonic detection signal. Specifically, the first ultrasonic detection signal acquisition module 50 conducts a detailed characteristic study on the target partition of the titanium alloy casting, identifies the key information such as the material characteristics, shape structure, size and size of the target partition, uses the laser ultrasonic emission technology to emit a specific laser pulse to the target partition, and analyzes the interaction between the laser pulse and the target partition. Based on the results of the detection preparation work, the preset control parameters of the ultrasonic detection are determined. The preset control parameters include the emission power, frequency, pulse width and other information of the laser ultrasonic transmitter, which are used to guide the detection operation of the laser ultrasonic transmitter. The signal characteristics of the obtained preset control parameters are extracted and planned. The signal characteristics refer to the key factors affecting the ultrasonic detection signal, such as the detection angle, the detection time interval, etc. Through the planning algorithm, the target partition of the titanium alloy casting is detected according to the preset control parameters to obtain the first ultrasonic detection signal. Each first ultrasonic detection signal represents the feedback information of the target partition under specific detection conditions, which provides basic data support and judgment basis for subsequent analysis and processing.

[0045] The second ultrasonic detection signal acquisition module 60 is used to extract the second ultrasonic detection signal pre-stored in the target partition of the titanium alloy casting according to the preset control parameters. Specifically, the second ultrasonic detection signal acquisition module 60 conducts an in-depth parameter study on the target partition of the titanium alloy casting, identifies key information such as the material composition, shape characteristics, and size specifications of the target partition, uses data storage and extraction technology to locate the relevant data pre-stored according to the preset control parameters, and analyzes the association and difference between these data. Based on the results of the parameter study, the extraction range of the second ultrasonic detection signal is determined, and the extraction range includes information such as the type, storage location, and time stamp of the signal, which is used to accurately obtain the required detection signal. The signal characteristics of the obtained extraction range information are extracted and screened, and the signal characteristics refer to the key factors that affect the availability of the ultrasonic detection signal, such as signal strength, frequency range, noise level, etc. Through the screening algorithm, the second ultrasonic detection signal that meets the requirements is extracted from the pre-stored data, and the second ultrasonic detection signal is sorted and optimized according to the extraction results and the characteristics of the target partition. Each second ultrasonic detection signal represents the acoustic feedback information of the target partition under specific conditions.

[0046] The defect detection result acquisition module 70 is used to perform defect analysis on the first ultrasonic detection signal according to the second ultrasonic detection signal to obtain defect detection results. Specifically, the defect detection result acquisition module 70 performs a comprehensive feature analysis on the first ultrasonic detection signal and the second ultrasonic detection signal, identifies key information such as the amplitude characteristics, frequency characteristics, phase changes, and propagation time laws of the signals, uses advanced signal processing technology to extract key parameters in the signals, and analyzes the differences and associations between them. Based on the results of the feature analysis, the direction of the defect analysis is determined. The direction of the defect analysis includes information such as areas and types where defects may exist, which is used to accurately detect defects. The key features of the obtained defect analysis direction information are extracted and classified. The key features refer to key factors that affect defect judgment, such as abnormal signal amplitude, frequency mutation, phase reversal, etc. Through the classification algorithm, signal areas with similar features are classified into one category. According to the classification results and the characteristics of the signal, the defect detection results are obtained. Each defect detection result represents the quality status of the target partition of the titanium alloy casting under specific detection conditions.

[0047] In one possible implementation, Figure 2 The defect detection result acquisition module 70 includes: a defect sample sampling unit, which is used to collect ultrasonic detection signals of defective samples and ultrasonic detection signals of non-defective samples according to a preset titanium alloy casting model. Specifically, the defect sample sampling unit determines the preset titanium alloy casting model, and adopts multiple ways to collect samples. On the production line, the titanium alloy castings that have been determined to have defects after preliminary detection, as well as castings that are defect-free in appearance and preliminary detection, are screened. From the inventory, representative defective and non-defective samples are obtained based on detailed records and markings. In an experimental environment, by controlling manufacturing conditions, samples with specific types and degrees of defects and completely defect-free samples are deliberately manufactured, and these samples are detected using ultrasonic detection equipment. For each sample, detection is performed from multiple angles and positions to obtain comprehensive and accurate ultrasonic detection signals. For defective samples, special attention is paid to defective areas to ensure that signals that can reflect defect characteristics are captured. For non-defective samples, multi-directional detection is also performed to obtain typical signals under normal conditions.

[0048] The identification output unit is used to configure a defect detection output identification for the ultrasonic detection signal of the defective sample, and to configure a non-defect output identification for the ultrasonic detection signal of the non-defective sample. Specifically, the identification output unit receives the ultrasonic detection signal from the defective sample sampling unit, and carefully configures a unique defect detection output identification for the ultrasonic detection signal from the defective sample. The identification is a specific code, mark or label that clearly indicates that the signal originates from a titanium alloy casting sample with defects. For the ultrasonic detection signal of the non-defective sample, a clear non-defect output identification is also given. The identification should not only be prominently reflected in the data, but also establish a close and unchangeable association with the corresponding ultrasonic detection signal to ensure that there will be no errors in the mismatch between the identification and the signal in subsequent processing and analysis.

[0049] A supervision unit is used to mix the ultrasonic detection signal of the defective sample and the ultrasonic detection signal of the non-defective sample, and to supervise with the defect detection output mark and the non-defective output mark to construct a defect detection channel. Specifically, the supervisory unit obtains ultrasonic detection signals of defective samples with labels and ultrasonic detection signals of non-defective samples, and randomly mixes the signals according to a certain proportion and rule. 70% of non-defective sample signals and 30% of defective sample signals can be mixed to simulate various situations that may be encountered in actual detection. During the mixing process, it is strictly ensured that the label of each signal is always closely connected with it without any confusion or loss. Machine learning is used to use the label as supervisory information to construct a defect detection channel. In the initial stage of training, some simple features in the mixed signal, such as the amplitude mean and frequency distribution range of the signal, are preliminarily learned and analyzed. As the training deepens, more complex features are gradually focused on, such as the mutation point of the signal and the distribution law of local peaks. In each iterative training, the current understanding and classification of the signal features are evaluated based on the label. If the predicted label is inconsistent with the actual label, the algorithm will automatically adjust the internal parameters, such as weights, biases, etc., to gradually improve the ability to distinguish between defective and non-defective signals. After multiple iterative trainings, until the algorithm can achieve a high accuracy and recall rate on an independent verification set, it indicates that the defect detection channel has been preliminarily constructed.

[0050] A defect analysis unit is used to perform defect analysis on the second ultrasonic detection signal and the first ultrasonic detection signal through the defect detection channel to obtain the defect detection result. Specifically, the defect analysis unit inputs the first ultrasonic detection signal and the second ultrasonic detection signal to be analyzed into the constructed defect detection channel. The model first preprocesses the input signal, including denoising, normalization and other operations to improve the accuracy of subsequent analysis, extracts deep-level features of the input signal, uses a convolutional neural network layer to capture local features, and uses a recurrent neural network layer to process the time series features of the signal. The extracted features are compared and matched with the defective and non-defective feature patterns learned during the training process. The final classification judgment is performed through the fully connected layer to determine whether the input signal is closer to a defective or non-defective pattern. If the judgment result is that there is a defect, the specific features of the defect are further analyzed, such as type (cracks, holes, etc.), location (by comparing with the defect location features in the training sample), size (based on quantitative analysis of signal strength and features), etc., and new sample data is continuously input into the channel for testing and verification. The model is fine-tuned according to the feedback results to adapt to defect detection of different types and degrees, and the accuracy and reliability of defect detection are continuously improved.

[0051] The embodiment of the present application adopts a module for constructing a three-dimensional model of a titanium alloy casting including surface coordinates and casting control parameters, a module for partitioning the casting surface coordinates into partition queues according to the casting control parameters, a module for traversing the partition queues to predict the defect probability and obtain a probability queue, a module for selecting a target partition that meets a threshold from the partition queue based on the probability queue, a module for detecting a first signal of a target partition according to preset parameters using a laser ultrasonic transmitter, a module for extracting a second signal pre-stored in the target partition according to preset parameters, and a module for analyzing the detection result of the first signal according to the second signal, so as to achieve the technical effect of accurately locating defects of titanium alloy castings and optimizing the allocation of detection resources through partition detection, thereby improving detection efficiency and accuracy and ensuring the quality and safety of castings.

[0052] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules may be used and run on the user terminal and / or server, and the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.

[0053] The above specific implementation manner does not constitute a limitation to the protection scope of the present application. It should be understood by those skilled in the art that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the protection scope of the present application. In some cases, the actions or steps recorded in the present application can be performed in an order different from that in the embodiment and can still achieve the desired results. In addition, the process depicted in the accompanying drawings does not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A titanium alloy casting defect detection system based on laser ultrasound, characterized in that: include: A three-dimensional model building module, wherein the three-dimensional model building module is used to build a three-dimensional model of a titanium alloy casting, wherein the three-dimensional model of the titanium alloy casting has surface coordinates of the titanium alloy casting and casting control parameters; A partitioning module, wherein the partitioning module is used to partition the surface coordinates of the titanium alloy casting according to the casting control parameters to obtain a partition queue of the titanium alloy casting; A defect distribution probability queue acquisition module, wherein the defect distribution probability queue acquisition module is used to traverse the titanium alloy casting partition queue to perform defect probability prediction and obtain a defect distribution probability queue; A target partition sorting module, the target partition sorting module is used to sort the titanium alloy casting target partition that meets the distribution probability threshold from the titanium alloy casting partition queue based on the defect distribution probability queue; A first ultrasonic detection signal acquisition module, the detection signal acquisition module is used to detect the target partition of the titanium alloy casting according to preset control parameters through a laser ultrasonic transmitter to obtain a first ultrasonic detection signal; A second ultrasonic detection signal acquisition module, the second ultrasonic detection signal acquisition module is used to extract the second ultrasonic detection signal pre-stored in the target partition of the titanium alloy casting according to the preset control parameters; a defect detection result acquisition module, the defect detection result acquisition module being used to perform defect analysis on the first ultrasonic detection signal according to the second ultrasonic detection signal to obtain a defect detection result; The partition module comprises: A record configuration unit, the record configuration unit is used to traverse the casting control attribute set and configure the casting control attribute partition threshold based on the casting defect historical detection record; A partition result acquisition unit, the partition result acquisition unit is used to traverse the casting control parameters and the casting control attribute partition thresholds, partition the surface coordinates of the titanium alloy casting respectively, and obtain a plurality of titanium alloy casting partition results; A partition queue acquisition unit, the partition queue acquisition unit is used to perform regional fusion segmentation on the partition results of the plurality of titanium alloy castings to obtain the titanium alloy casting partition queue; The defect distribution probability queue acquisition module includes: A first titanium alloy casting partition extraction unit, the first titanium alloy casting partition extraction unit is used to extract the first titanium alloy casting partition according to the titanium alloy casting partition queue; A deviation vector matrix construction unit, the deviation vector matrix construction unit is used to extract the casting control parameter expected value and the casting control parameter monitoring value of the first titanium alloy casting partition, and construct a normalized casting control deviation vector matrix; A probability queue distribution unit is used to perform historical sample retrieval on the titanium alloy casting model according to the normalized casting control deviation vector matrix, obtain the defect sample ratio, set it as the first titanium alloy casting partition defect distribution probability, and add it into the defect distribution probability queue.

2. The system according to claim 1, characterized in that The record configuration unit includes: A coefficient set acquisition subunit, the coefficient set acquisition subunit is used to perform casting defect correlation analysis on the casting control attribute set to obtain a casting control attribute correlation coefficient set; An attribute sorting result subunit, the attribute sorting result subunit being used to sort casting control attribute sorting results satisfying a correlation coefficient threshold from the casting control attribute set according to the casting control attribute correlation coefficient set; A partition threshold configuration subunit is used to configure the casting control attribute partition threshold for the casting control attribute sorting result according to the casting defect historical detection record.

3. The system according to claim 2, characterized in that The coefficient set acquisition subunit comprises: A first casting control attribute extraction micro-unit, wherein the first casting control attribute extraction micro-unit is used to traverse the casting control attribute set and extract the first casting control attribute; A frequency recording micro-unit, wherein the frequency recording micro-unit is used to take the deviation between the monitoring value of the casting control parameter of the first casting control attribute and the expected value of the casting control parameter as the only variable, determine the variable characteristic value, take the first casting control attribute deviation and the titanium alloy casting model as constraints, collect a preset number of defect monitoring logs, count the defect detection record frequencies of the preset number of defect monitoring logs, and add them into the first casting control attribute deviation sequence and the first defect detection record frequency sequence, wherein the first casting control attribute deviation corresponds to the first defect detection record frequency one by one; The analysis function constructs a micro-unit, and the analysis function constructs a micro-unit for constructing a correlation coefficient analysis function: in, Characterize the correlation coefficient of any attribute, represents the i-th deviation value, represents the i-th detection record frequency, and N represents the number of sequences; a first casting control attribute correlation coefficient acquisition micro unit, wherein the first casting control attribute correlation coefficient acquisition micro unit is used to perform a correlation analysis on the first casting control attribute deviation sequence and the first defect detection record frequency sequence according to the correlation coefficient analysis function to obtain the first casting control attribute correlation coefficient; The association degree set adding micro unit is used to add the first casting control attribute correlation coefficient into the casting control attribute association degree set.

4. The system according to claim 3, characterized in that The partition threshold configuration subunit includes: A basic data analysis micro-unit, the basic data analysis micro-unit is used to analyze the intrinsic characteristics of the first casting control attribute deviation sequence and the defect detection record frequency sequence; An attribute selection micro-unit, wherein the attribute selection micro-unit is used to extract a first casting control selected attribute according to the casting control attribute sorting result; a calibration deviation threshold micro unit, wherein the calibration deviation threshold micro unit is used to calibrate the first casting control selected attribute deviation threshold for the first casting control selected attribute through the user terminal; A positive partition threshold selected micro-unit, wherein the positive partition threshold selected micro-unit is used to perform a positive deviation analysis on the first casting control selected attribute based on the first casting control selected attribute deviation threshold to obtain a first casting control selected attribute positive partition threshold; A negative partition threshold selected micro-unit, wherein the negative partition threshold selected micro-unit is used to perform a negative deviation analysis on the first casting control selected attribute based on the first casting control selected attribute deviation threshold to obtain a negative partition threshold of the first casting control selected attribute; The negative partition threshold storage micro unit is used to store the first casting control selected attribute positive partition threshold and the first casting control selected attribute negative partition threshold as the first casting control selected attribute partition threshold, and add them into the casting control attribute partition threshold.

5. The system according to claim 4, characterized in that The positive partition threshold selection micro-unit comprises: A gradient sequence construction nanounit, wherein the gradient sequence construction nanounit is used to construct a positive deviation gradient sequence based on a deviation threshold of the first casting control selected attribute and a positive deviation between a casting control parameter monitoring value and an expected value of the casting control parameter of the first casting control selected attribute as a unique variable; a second defect detection record frequency sequence calibration unit, the second defect detection record frequency sequence calibration unit being used to traverse the forward deviation gradient sequence and calibrate a second defect detection record frequency sequence, wherein the second defect detection record frequency sequence is configured in the same manner as the first defect detection record frequency sequence; A sequence interception unit, the sequence interception unit is used to intercept the positive deviation gradient sequence from the second defect detection record frequency sequence based on the defect detection record frequency threshold, to obtain the first casting control selected attribute positive partition threshold, including: a detection record frequency extraction unit, the detection record frequency extraction unit being used to extract a k-th detection record frequency and a k+1-th detection record frequency according to the second defect detection record frequency sequence; A positive deviation gradient sequence interception unit is used to intercept the positive deviation gradient sequence based on the k+1th sequence number detection record frequency when the k+1th sequence number detection record frequency is greater than or equal to the defect detection record frequency threshold, and the kth sequence number detection record frequency is less than the defect detection record frequency threshold, to obtain the first casting control selected attribute positive partition threshold.

6. The system according to claim 1, wherein: The defect detection result acquisition module includes: A defect sample sampling unit, wherein the defect sample sampling unit is used to collect defect sample ultrasonic detection signals and non-defect sample ultrasonic detection signals according to a preset titanium alloy casting model; An identification output unit, the identification output unit is used to configure a defect detection output identification for the defective sample ultrasonic detection signal, and to configure a non-defect output identification for the non-defective sample ultrasonic detection signal; A supervision unit, the supervision unit is used to mix the ultrasonic detection signal of the defective sample and the ultrasonic detection signal of the non-defective sample, supervise with the defect detection output mark and the non-defective output mark, and build a defect detection channel; A defect analysis unit is used to perform defect analysis on the second ultrasonic detection signal and the first ultrasonic detection signal through the defect detection channel to obtain the defect detection result.

Images