Data analysis method and system based on intelligent operation and maintenance management result of numerical control machine tool

Abstract

The invention discloses a data analysis method and system based on an intelligent operation and maintenance management result of a numerical control machine tool, and relates to the technical field of data analysis, and the method comprises the steps: inputting an alarm signal of the numerical control machine tool into a local operation and maintenance database for retrieval, and obtaining a sample maintenance case set; when the sample data volume meets a preset index, carrying out fault type proportion statistics, and determining a first trigger fault type; fault detection is carried out, and if the first detection result is normal, fault occurrence prediction is carried out on the remaining N-1 associated fault types; optimizing and outputting N-1 optimization fault correlation degrees, and determining a second trigger fault type; and carrying out fault detection on the numerical control machine tool according to a second trigger fault type, and if a second detection result is normal, continuing to carry out iterative circulation of association degree adjustment-trigger fault type screening-machine tool fault detection until the detection result is abnormal. According to the invention, the technical problems of high blindness and low efficiency of fault detection in the prior art are solved.

Description

Technical Field

[0001] This invention relates to the field of data analysis technology, specifically to a data analysis method and system based on the results of intelligent operation and maintenance management of CNC machine tools. Background Technology

[0002] As core equipment in modern manufacturing, fault diagnosis is a crucial aspect of intelligent operation and maintenance for CNC machine tools. Existing technologies typically follow a fixed fault diagnosis sequence, such as searching a list of fault types corresponding to the same signal before proceeding with repairs. However, this approach has significant drawbacks: firstly, historical case databases cannot fully cover complex fault scenarios, potentially leading to diagnostic biases; secondly, traditional methods exhibit considerable randomness and haphazardness in fault diagnosis, resulting in low operation and maintenance efficiency. Summary of the Invention

[0003] This application provides a data analysis method and system based on the results of intelligent operation and maintenance management of CNC machine tools, which is used to address the technical problems of high blindness and low efficiency in fault detection in the prior art.

[0004] In view of the above problems, this application provides a data analysis method and system based on the results of intelligent operation and maintenance management of CNC machine tools.

[0005] Firstly, this application provides a data analysis method based on the results of intelligent operation and maintenance management of CNC machine tools, the method comprising: The alarm signals of the CNC machine tool are entered into the local maintenance database for retrieval to obtain a sample maintenance case set; If the sample data volume of the sample repair case set meets the preset index, the proportion of fault types is statistically analyzed based on the sample repair case set, N associated fault types and N fault correlation degrees are output, and the first triggering fault type is determined. The CNC machine tool is subjected to fault detection according to the first triggered fault type. If the first detection result is normal, the remaining N-1 associated fault types are predicted based on the first triggered fault type to obtain N-1 predicted fault probabilities. Based on the N-1 predicted fault probabilities, N-1 correlation adjustment coefficients are calculated. The N-1 fault correlations of the remaining N-1 associated fault types are optimized and corrected, and N-1 optimized fault correlations are output. The second triggering fault type is then determined. The CNC machine tool is subjected to fault detection according to the second trigger fault type. If the second detection result is normal, the iterative cycle of correlation adjustment-trigger fault type screening-machine tool fault detection continues until the detection result is abnormal.

[0006] Secondly, this application provides a data analysis system based on the results of intelligent operation and maintenance management of CNC machine tools, including: The case acquisition module is used to input the alarm signals of CNC machine tools into the local operation and maintenance database for retrieval and to obtain a sample maintenance case set; The first fault type acquisition module is used to perform fault type proportion statistics based on the sample maintenance case set if the sample data volume of the sample maintenance case set meets the preset index, output N associated fault types and N fault correlation degrees, and determine the first triggering fault type. The fault probability prediction module is used to perform fault detection on the CNC machine tool according to the first triggered fault type. If the first detection result is normal, the module performs fault occurrence prediction on the remaining N-1 associated fault types based on the first triggered fault type to obtain N-1 predicted fault probabilities. The second fault type acquisition module is used to calculate N-1 correlation adjustment coefficients based on the N-1 predicted fault probabilities, optimize and correct the N-1 fault correlations of the remaining N-1 associated fault types, output N-1 optimized fault correlations, and determine the second trigger fault type. The loop detection module is used to perform fault detection on the CNC machine tool according to the second triggered fault type. If the second detection result is normal, the iterative loop of correlation adjustment-triggered fault type screening-machine tool fault detection continues until the detection result is abnormal.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages: This application proposes a data analysis method and system based on the results of intelligent operation and maintenance management of CNC machine tools. By dynamically correcting the association weights of remaining faults based on real-time predicted probabilities and iteratively updating the troubleshooting path, the efficiency and accuracy of CNC machine tool fault diagnosis are significantly improved. Compared with traditional methods, the technical solution provided in this application significantly overcomes the problem of high fault randomness under static rules, achieving the technical effect of replacing static linear troubleshooting with dynamic prediction, and realizing efficient and adaptive diagnosis of CNC machine tool faults. Attached Figure Description

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

[0009] Figure 1 This is a flowchart illustrating the data analysis method based on the intelligent operation and maintenance management results of CNC machine tools provided in this application embodiment.

[0010] Figure 2A schematic diagram of the structure of a data analysis system based on the intelligent operation and maintenance management results of CNC machine tools, provided in an embodiment of this application.

[0011] The components represented by each number in the attached diagram are explained below: Case acquisition module 100, first fault type acquisition module 200, fault probability prediction module 300, second fault type acquisition module 400, and loop detection module 500. Detailed Implementation

[0012] This application provides a data analysis method and system based on the results of intelligent operation and maintenance management of CNC machine tools, which is used to address the technical problems of high blindness and low efficiency in fault detection in the prior art.

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

[0014] It should be noted that the terms "comprising" and "having" are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to these processes, methods, products, or devices.

[0015] Example 1, as Figure 1 As shown, this application provides a data analysis method based on the results of intelligent operation and maintenance management of CNC machine tools, wherein the method includes: S10: Input the alarm signals of the CNC machine tool into the local maintenance database for retrieval and obtain a sample maintenance case set.

[0016] In this embodiment, alarm signals from the CNC machine tool are input into a local maintenance database for retrieval to obtain a sample repair case set. The alarm signals from the CNC machine tool include various types, such as electrical alarm signals (overvoltage alarm signals, grounding fault alarm signals, etc.); mechanical alarm signals (spindle overheat alarm signals, mechanical jamming alarm signals, etc.); system alarm signals (program error alarm signals, memory fault alarm signals, etc.); and feed system alarm signals (feed motor alarm signals, feed out of control alarm signals, etc.). The local maintenance database is established by mapping historical alarm signals to historical repair solutions, where one alarm signal may correspond to multiple repair solutions. Inputting the CNC machine tool alarm signal into the local maintenance database retrieves the corresponding sample repair cases. For example, an overvoltage alarm signal may correspond to repairs such as power supply repair or short circuit repair.

[0017] By targeting and retrieving data from the local maintenance database, a sample set of repair cases is obtained, fully leveraging the strong relevance of the local database. Compared to global retrieval, this method prioritizes cases that highly match the current machine tool attributes and maintenance environment, significantly improving the accuracy of the sample set of repair cases.

[0018] S20: If the sample data volume of the sample repair case set meets the preset index, perform fault type proportion statistics based on the sample repair case set, output N associated fault types and N fault correlation degrees, and determine the first triggering fault type.

[0019] Traditional fault statistics methods force the output of correlation results when the sample size is insufficient, leading to distorted fault correlation. When the local sample size is small, directly introducing external data will introduce bias and dilute the value of the local data.

[0020] Step S20 in the method provided in this application embodiment includes: If the sample data volume of the sample repair case set is greater than or equal to the preset index, the proportion of the same fault type is statistically analyzed according to the sample repair case set, and the fault type proportion is set as the fault correlation degree, resulting in N related fault types and N fault correlation degrees, where the preset index is 200, and N is an integer greater than or equal to 1. If the sample data volume of the sample maintenance case set is less than the preset index, then the attribute information of the CNC machine tool is used as the equipment constraint, and the alarm signal and N associated fault types are used as the condition constraint. Based on big data, a multi-dimensional maintenance case set is obtained. If the sample data volume of the sample repair case set is less than the preset index, calculate the difference between the two to obtain the data volume difference, and set the product of the data volume difference and 10 as the preset retrieval data volume; Using the attribute information of the CNC machine tool as equipment constraints and the alarm signal as conditional constraints, a retrieval is performed based on big data until the retrieval data reaches the preset retrieval data volume, thereby obtaining a multi-dimensional maintenance case set.

[0021] Based on the sample repair case set and the multivariate repair case set, the proportion of the same fault type is statistically analyzed, and the proportion of fault type is set as the fault correlation degree, resulting in N related fault types and N fault correlation degrees. The statistical analysis of the proportion of the same fault type based on the sample repair case set and the multi-factor repair case set includes: Obtain the complete set of fault types corresponding to the alarm signal, wherein the complete set of fault types includes all possible fault types when the alarm signal occurs; The ratio of N to the number of fault types in the comprehensive fault type set is set as the sample generalization coefficient; The sample weight adjustment coefficient is calculated based on the sample generalization coefficient and the data volume difference, and the ratio of the sample weight adjustment coefficient to the initial sample data weight is set as the adapted sample data weight, wherein the initial sample data weight is 2, and the sample weight adjustment coefficient is positively correlated with the sample generalization coefficient and the data volume difference; Configure the weights of multi-data elements for big data retrieval, where the weight of multi-data elements is 0.1; Based on the weights of the adapted sample data and the weights of the multivariate data, the proportion of the same fault type is statistically analyzed according to the sample repair case set and the multivariate repair case set. The associated fault type with the highest fault correlation degree among the N associated fault types is set as the first triggering fault type.

[0022] In this embodiment, if the sample data size of the sample repair case set is greater than or equal to a preset index, the proportion of the same fault type is statistically analyzed based on the sample repair case set. The fault type proportion = the number of the same fault type ÷ the total number of faults in the sample repair case set. The fault type proportion is then set as the fault correlation degree, resulting in N associated fault types and N fault correlation degrees. The preset index is 200, where N is an integer greater than or equal to 1. For example, if the sample size in the sample repair case set is 300 and the number of feed motor alarm faults is 120, then the feed motor alarm fault correlation degree = 120 ÷ 300 = 0.4.

[0023] If the sample data size of the sample repair case set is less than the preset index, calculate the difference between the two to obtain the data size difference, and set the product of the data size difference and 10 as the preset retrieval data size. For example, if the sample repair case set has 100 cases, the data size difference = 200 - 100 = 100, and the preset retrieval data size = 100 × 10 = 1000.

[0024] Using the attribute information of CNC machine tools as equipment constraints and alarm signals as conditional constraints, a search is performed based on big data until the preset amount of search data is reached, resulting in a multi-dimensional maintenance case set. The results of big data retrieval are not as accurate or adaptable as those from the sample maintenance case set; therefore, weights need to be set based on the sample maintenance case set.

[0025] Specifically, obtain the complete set of fault types corresponding to the alarm signal, wherein the complete set of fault types includes all possible fault types when the alarm signal occurs.

[0026] The ratio of N to the number of fault types in the total fault type set is set as the sample generalization coefficient. For example, if the total fault type set has 30 fault types and N is 21, then the sample generalization coefficient = N ÷ number of fault types = 21 ÷ 30 = 0.7.

[0027] The sample weight adjustment coefficient is calculated based on the sample generalization coefficient and the difference in data volume. The ratio of the sample weight adjustment coefficient to the initial sample data weight is set as the adapted sample data weight. The initial sample data weight is 2. The sample weight adjustment coefficient is positively correlated with the sample generalization coefficient and the difference in data volume. For example, the standard generalization coefficient is set to 50%, and the standard data volume difference is set to 100. When the sample generalization coefficient is 0.7 and the data volume difference is 80, each is assigned a weight of 0.5, so the adjustment coefficient = (0.7 ÷ 0.5) × 0.5 + (80 ÷ 100) × 0.5 = 1.1. For example, if the initial sample data weight is set to 0.6, then the adapted sample data weight = sample weight adjustment coefficient ÷ initial sample data weight = 1.1 × 2 = 2.2. Since the adaptability and reference value of the sample repair case set data are stronger than those of the multivariate repair case set, the more comprehensive the fault types in the sample data and the smaller the sample data volume, the greater the weight.

[0028] Configure the weights of multi-data elements for big data retrieval, where the weight of multi-data elements is 0.1; Based on the weights of the adapted sample data and the weights of the multivariate data, the proportion of the same fault type is statistically analyzed according to the sample maintenance case set and the multivariate maintenance case set. The proportion of the fault type is set as the fault correlation degree, resulting in N associated fault types and N fault correlation degrees. For example, if the sample maintenance case set has 100 data points and 40 feed motor alarm faults, and the multivariate maintenance case set has 1000 data points and 300 feed motor alarm faults, then the correlation degree of feed motor alarm faults in the sample maintenance case set = 40 ÷ 100 = 0.4, and the correlation degree of feed motor alarm faults in the multivariate maintenance case set = 300 ÷ 1000 = 0.3. Combining the weights, the weight of the sample maintenance case set is 2.2, and the weight of the multivariate data is 0.1. Therefore, the correlation degree of feed motor alarm faults = 2.2 × 0.4 + 0.1 × 0.3 = 0.91.

[0029] Set the associated fault type with the highest fault correlation degree among the N associated fault types as the first trigger fault type. For example, if the highest fault correlation degree is 0.91 and the corresponding associated fault type is feed motor alarm fault, then the first trigger fault type is feed motor alarm fault.

[0030] The statistical strategy is dynamically selected based on the sample size. When the data volume meets the standard, local data is focused on to ensure statistical specificity. When the data volume is insufficient, external data is retrieved through device attribute constraints, and a weighting mechanism is designed to integrate multi-source cases and eliminate sample bias. Finally, the fault with the highest correlation is selected as the first detection target, so that the initial detection is targeted, balancing data quality and statistical efficiency, and providing a reliable starting point for subsequent iterations.

[0031] S30: Perform fault detection on the CNC machine tool according to the first triggered fault type. If the first detection result is normal, predict the occurrence of faults for the remaining N-1 associated fault types based on the first triggered fault type to obtain N-1 predicted fault probabilities.

[0032] When the first detected fault does not occur, traditional methods continue to detect the remaining faults according to the initial correlation order, ignoring the dynamic impact of the first result on the probability of the remaining faults. For example, if the first highly correlated fault does not occur, the probability of other strongly correlated faults occurring may increase, but the static correlation degree cannot reflect this change.

[0033] Step S30 in the method provided in this application embodiment includes: If the first detection result is abnormal, the subsequent analysis shall be stopped, and the CNC machine tool shall be repaired based on the first triggered fault type. If the first detection result is normal, the CNC machine tool's attribute information is used as the equipment constraint, and the alarm signal is used as the condition constraint. Based on big data, the proportion of historical events in which the remaining N-1 associated fault types occur simultaneously when the first triggered fault type occurs is statistically analyzed, and the proportion of events with the same frequency is set as the predicted fault probability to obtain N-1 predicted fault probabilities.

[0034] In this embodiment of the application, the CNC machine tool is fault detected according to the first triggered fault type. For example, if the first triggered fault type is a feed motor alarm fault, the operating status of the feed motor of the CNC machine tool is detected to obtain the first detection result.

[0035] If the initial detection result is abnormal, such as detecting an abnormal operation of the feed motor, subsequent analysis is stopped, and the CNC machine tool is repaired based on the first triggered fault type. The repair plan is selected from the multiple repair plans corresponding to the feed motor alarm fault in the local maintenance database, based on the highest percentage of each plan. For example, if there are three repair plans corresponding to the feed motor alarm fault in the local maintenance database: 12 for adjusting operating data, 8 for cleaning the motor drive shaft, and 10 for replacing the bearing, then the plan to adjust operating data is used for repair, followed by another fault detection. If the fault detection result is normal, the repair ends. If the fault detection result is abnormal, the repair plan is selected from the multiple repair plans corresponding to the feed motor alarm fault in the local maintenance database, in descending order of the highest percentage of each plan, for repair.

[0036] If the first detection result is normal, using the CNC machine tool's attribute information as equipment constraints and the alarm signal as a conditional constraint, based on big data, the percentage of historical events where the remaining N-1 associated fault types occur simultaneously when the first triggered fault type occurs is statistically analyzed. This percentage of events with the same frequency is then set as the predicted fault probability, resulting in N-1 predicted fault probabilities. For example, if the first detection result is normal, then based on big data, historical events where the CNC machine tool's attributes and alarm signals are the same as the current CNC machine tool are retrieved. If the first triggered fault type is a feed motor alarm fault, then the percentage of historical events where other faults occur simultaneously with the feed motor alarm fault is statistically analyzed. For example, if the number of overvoltage faults among faults occurring simultaneously with the feed motor alarm fault accounts for 0.2% of all fault events, and the number of spindle overheating faults accounts for 0.15% of all fault events, then the predicted fault probability for overvoltage faults is 0.2%, and the predicted fault probability for spindle overheating faults is 0.15%.

[0037] After the initial normal detection, the probability of remaining faults is predicted in real time based on the triggered fault type. By statistically analyzing the historical co-occurrence frequency of other faults when the fault occurs, dynamic probability is incorporated into the correlation assessment. Compared to static correlation ranking, this method enables the system to perceive the conditional dependencies between faults, improving prediction accuracy.

[0038] S40: Calculate N-1 correlation adjustment coefficients based on the N-1 predicted fault probabilities, optimize and correct the N-1 fault correlations of the remaining N-1 associated fault types, output N-1 optimized fault correlations, and determine the second triggering fault type.

[0039] Traditional methods cannot update fault weights based on real-time detection feedback, resulting in highly unpredictable fault detection and low detection efficiency.

[0040] Step S40 in the method provided in this application embodiment includes: The N-1 predicted fault probabilities are summed by adding 1 to each of them to obtain N-1 correlation adjustment coefficients; The N-1 fault correlations are optimized and corrected based on the N-1 correlation adjustment coefficients, and N-1 optimized fault correlations are output. The associated fault type corresponding to the maximum optimized fault correlation is selected as the second trigger fault type.

[0041] In this embodiment, the N-1 predicted fault probabilities are summed with 1 to obtain N-1 correlation adjustment coefficients. For example, the predicted fault probability of overvoltage fault is 0.2, and the predicted fault probability of spindle overheating fault is 0.15. Then, the correlation adjustment coefficient of overvoltage fault = 1 + 0.2 = 1.2, and the correlation adjustment coefficient of spindle overheating fault = 1 + 0.15 = 1.15.

[0042] Based on N-1 correlation adjustment coefficients, the N-1 fault correlations are optimized and corrected, resulting in N-1 optimized fault correlations. Optimized fault correlation = fault correlation × correlation adjustment coefficient, where the fault correlation is obtained using the same method as described in the previous steps. For example, the fault correlation of an overvoltage fault is 0.25, and the correlation adjustment coefficient is 1.2; the predicted fault probability of a spindle overheating fault is 0.18, and the correlation adjustment coefficient is 1.15. Therefore, the optimized correlation of an overvoltage fault is 0.25 × 1.2 = 0.3, and the optimized correlation of a spindle overheating fault is 0.18 × 1.15 = 0.207. The associated fault type corresponding to the maximum optimized fault correlation is selected and designated as the second trigger fault type.

[0043] The predicted failure probability is converted into a correlation adjustment coefficient, which is used to correct the original correlation of the remaining failures in real time. The coefficient design retains the initial statistical weights while incorporating the effects of real-time feedback. The optimized correlation dynamically reflects the most likely failure sequence, and the failure with the second highest optimized correlation is selected as the second detection target.

[0044] S50: Perform fault detection on the CNC machine tool according to the second triggered fault type. If the second detection result is normal, continue the iterative cycle of correlation adjustment - triggered fault type screening - machine tool fault detection until the detection result is abnormal.

[0045] Complex faults often require multiple rounds of testing to locate, but traditional methods lack iterative control mechanisms, which may lead to missed detections due to insufficient testing or waste resources due to excessive testing.

[0046] Step S50 in the method provided in this application embodiment includes: According to the second triggering fault type, the CNC machine tool is fault detected. If the second detection result is normal, the proportion of historical events in which the remaining N-2 associated fault types occur simultaneously when the first triggering fault type and the second triggering fault type occur is statistically analyzed. The proportion of events with the same frequency is set as the predicted fault probability to obtain N-2 predicted fault probabilities. N-2 correlation adjustment coefficients are then calculated. The N-2 correlation coefficients are used to optimize and correct the correlation of the N-2 faults, and the third trigger fault type is determined.

[0047] In this embodiment of the application, the CNC machine tool is fault detected according to the second trigger fault type. If the second detection result is normal, the same method as the above steps is used to count the proportion of historical events in which the remaining N-2 related fault types occur simultaneously when the first trigger fault type and the second trigger fault type occur. The proportion of events with the same frequency is set as the predicted fault probability to obtain N-2 predicted fault probabilities and N-2 correlation adjustment coefficients are calculated.

[0048] Based on N-2 correlation adjustment coefficients, the N-2 optimized fault correlations are optimized and corrected, and the third trigger fault type is determined. Then, the fault detection of the CNC machine tool is performed.

[0049] Repeat the iterative cycle of correlation adjustment, fault type screening, and machine tool fault detection until an anomaly is found in the detection results. Based on the anomaly detection results and the local maintenance database, the CNC machine tool is repaired.

[0050] This application constructs an iterative loop architecture. After each normal detection, the remaining fault probability is re-predicted based on the eliminated fault types, the correlation is corrected, and the next detection target is triggered. The loop progressively accumulates detected faults as conditions, making the prediction more accurate while ensuring that the loop exits immediately after fault location. By progressively narrowing the investigation scope, the diagnosis is completed with the fewest possible detections, avoiding ineffective loops and improving detection efficiency.

[0051] Example 2, as Figure 2As shown, based on the same inventive concept as the data analysis method for intelligent operation and maintenance management results of CNC machine tools provided in Embodiment 1, this embodiment of the invention also provides a data analysis system for intelligent operation and maintenance management results of CNC machine tools, including: The case acquisition module 100 is used to input the alarm signals of CNC machine tools into the local operation and maintenance database for retrieval and to obtain a sample maintenance case set; The first fault type acquisition module 200 is used to perform fault type proportion statistics based on the sample maintenance case set if the sample data volume of the sample maintenance case set meets the preset index, output N associated fault types and N fault correlation degrees, and determine the first triggering fault type. The fault probability prediction module 300 is used to perform fault detection on the CNC machine tool according to the first triggered fault type. If the first detection result is normal, it performs fault occurrence prediction on the remaining N-1 associated fault types based on the first triggered fault type to obtain N-1 predicted fault probabilities. The second fault type acquisition module 400 is used to calculate N-1 correlation adjustment coefficients based on the N-1 predicted fault probabilities, optimize and correct the N-1 fault correlations of the remaining N-1 associated fault types, output N-1 optimized fault correlations, and determine the second trigger fault type. The loop detection module 500 is used to perform fault detection on the CNC machine tool according to the second triggered fault type. If the second detection result is normal, the iterative loop of correlation adjustment-triggered fault type screening-machine tool fault detection continues until the detection result is abnormal.

[0052] In one embodiment, the first fault type acquisition module 200 is further configured to: If the sample data volume of the sample repair case set is greater than or equal to the preset index, the proportion of the same fault type is statistically analyzed according to the sample repair case set, and the fault type proportion is set as the fault correlation degree, resulting in N related fault types and N fault correlation degrees, where the preset index is 200, and N is an integer greater than or equal to 1. If the sample data volume of the sample maintenance case set is less than the preset index, then the attribute information of the CNC machine tool is used as the equipment constraint, and the alarm signal and N associated fault types are used as the condition constraint. Based on big data, a multi-dimensional maintenance case set is obtained. If the sample data volume of the sample repair case set is less than the preset index, calculate the difference between the two to obtain the data volume difference, and set the product of the data volume difference and 10 as the preset retrieval data volume; Using the attribute information of the CNC machine tool as equipment constraints and the alarm signal as conditional constraints, a retrieval is performed based on big data until the retrieval data reaches the preset retrieval data volume, thereby obtaining a multi-dimensional maintenance case set.

[0053] Based on the sample repair case set and the multivariate repair case set, the proportion of the same fault type is statistically analyzed, and the proportion of fault type is set as the fault correlation degree, resulting in N related fault types and N fault correlation degrees. The statistical analysis of the proportion of the same fault type based on the sample repair case set and the multi-factor repair case set includes: Obtain the complete set of fault types corresponding to the alarm signal, wherein the complete set of fault types includes all possible fault types when the alarm signal occurs; The ratio of N to the number of fault types in the comprehensive fault type set is set as the sample generalization coefficient; The sample weight adjustment coefficient is calculated based on the sample generalization coefficient and the data volume difference, and the ratio of the sample weight adjustment coefficient to the initial sample data weight is set as the adapted sample data weight, wherein the initial sample data weight is 2, and the sample weight adjustment coefficient is positively correlated with the sample generalization coefficient and the data volume difference; Configure the weights of multi-data elements for big data retrieval, where the weight of multi-data elements is 0.1; Based on the weights of the adapted sample data and the weights of the multivariate data, the proportion of the same fault type is statistically analyzed according to the sample repair case set and the multivariate repair case set. The associated fault type with the highest fault correlation degree among the N associated fault types is set as the first triggering fault type.

[0054] In one embodiment, the fault probability prediction module 300 is further configured to: If the first detection result is abnormal, the subsequent analysis shall be stopped, and the CNC machine tool shall be repaired based on the first triggered fault type. If the first detection result is normal, the CNC machine tool's attribute information is used as the equipment constraint, and the alarm signal is used as the condition constraint. Based on big data, the proportion of historical events in which the remaining N-1 associated fault types occur simultaneously when the first triggered fault type occurs is statistically analyzed, and the proportion of events with the same frequency is set as the predicted fault probability to obtain N-1 predicted fault probabilities.

[0055] In one embodiment, the second fault type acquisition module 400 is further configured to: The N-1 predicted fault probabilities are summed by adding 1 to each of them to obtain N-1 correlation adjustment coefficients; The N-1 fault correlations are optimized and corrected based on the N-1 correlation adjustment coefficients, and N-1 optimized fault correlations are output. The associated fault type corresponding to the maximum optimized fault correlation is selected as the second trigger fault type.

[0056] In one embodiment, the loop detection module 500 is further configured to: According to the second triggering fault type, the CNC machine tool is fault detected. If the second detection result is normal, the proportion of historical events in which the remaining N-2 associated fault types occur simultaneously when the first triggering fault type and the second triggering fault type occur is statistically analyzed. The proportion of events with the same frequency is set as the predicted fault probability to obtain N-2 predicted fault probabilities. N-2 correlation adjustment coefficients are then calculated. The N-2 correlation coefficients are used to optimize and correct the correlation of the N-2 faults, and the third trigger fault type is determined.

[0057] In summary, the embodiments of this application have at least the following technical effects: This application proposes a data analysis method and system based on the results of intelligent operation and maintenance management of CNC machine tools. By dynamically correcting the correlation weights of remaining faults based on real-time predicted probabilities and iteratively updating the troubleshooting path, the system significantly improves the diagnostic efficiency and accuracy of CNC machine tool faults. Specifically, when the sample library is insufficient, the system integrates local and external data through attribute constraints and adaptive weight adjustment mechanisms to ensure the coverage and representativeness of the correlation fault statistics. After the initial fault detection is normal, the correlation adjustment coefficient is dynamically calculated based on historical co-occurrence probabilities to optimize the troubleshooting priority of remaining faults in real time, forming a closed-loop optimization process of "detection-feedback-correction," ensuring that each detection focuses on the most likely fault type. This iterative mechanism can gradually converge the troubleshooting scope, significantly reducing the number of invalid detections and downtime, while also reducing the risk of missed detections due to misjudgment of correlation. Furthermore, through progressive screening and correlation correction of multi-level fault trigger types, the system can quickly locate the real fault source within a limited number of detections, improving the accuracy of fault diagnosis. Compared with traditional methods, the technical solution provided in this application significantly overcomes the problem of high randomness of faults under static rules, and achieves the technical effect of replacing static linear investigation with dynamic prediction to realize efficient and adaptive diagnosis of CNC machine tool faults.

[0058] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0059] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0060] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. A data analysis method based on the results of intelligent operation and maintenance management of CNC machine tools, characterized in that, The methods include: The alarm signals of the CNC machine tool are entered into the local maintenance database for retrieval to obtain a sample maintenance case set; If the sample data volume of the sample repair case set meets the preset index, the proportion of fault types is statistically analyzed based on the sample repair case set, N associated fault types and N fault correlation degrees are output, and the first triggering fault type is determined. The CNC machine tool is subjected to fault detection according to the first triggered fault type. If the first detection result is normal, the remaining N-1 associated fault types are predicted based on the first triggered fault type to obtain N-1 predicted fault probabilities. Based on the N-1 predicted fault probabilities, N-1 correlation adjustment coefficients are calculated. The N-1 fault correlations of the remaining N-1 associated fault types are optimized and corrected, and N-1 optimized fault correlations are output. The second triggering fault type is then determined. The CNC machine tool is subjected to fault detection according to the second trigger fault type. If the second detection result is normal, the iterative cycle of correlation adjustment-trigger fault type screening-machine tool fault detection continues until the detection result is abnormal.

2. The data analysis method based on the intelligent operation and maintenance management results of CNC machine tools according to claim 1, characterized in that, If the sample data volume of the sample repair case set meets the preset index, the proportion of fault types is statistically analyzed based on the sample repair case set, N associated fault types and N fault correlation degrees are output, and the first triggering fault type is determined, including: If the sample data volume of the sample repair case set is greater than or equal to the preset index, the proportion of the same fault type is statistically analyzed based on the sample repair case set, and the fault type proportion is set as the fault correlation degree, resulting in N related fault types and N fault correlation degrees. If the sample data volume of the sample maintenance case set is less than the preset index, then the attribute information of the CNC machine tool is used as the equipment constraint, and the alarm signal and N associated fault types are used as the condition constraint. Based on big data, a multi-dimensional maintenance case set is obtained. Based on the sample repair case set and the multivariate repair case set, the proportion of the same fault type is statistically analyzed, and the proportion of fault type is set as the fault correlation degree, resulting in N related fault types and N fault correlation degrees. The associated fault type with the highest fault correlation degree among the N associated fault types is set as the first triggering fault type.

3. The data analysis method based on the intelligent operation and maintenance management results of CNC machine tools according to claim 2, characterized in that, If the sample data volume of the sample repair case set is less than the preset index, calculate the difference between the two to obtain the data volume difference, and set the product of the data volume difference and 10 as the preset retrieval data volume; Using the attribute information of the CNC machine tool as equipment constraints and the alarm signal as conditional constraints, a retrieval is performed based on big data until the retrieval data reaches the preset retrieval data volume, thereby obtaining a multi-dimensional maintenance case set.

4. The data analysis method based on the intelligent operation and maintenance management results of CNC machine tools according to claim 3, characterized in that, Based on the sample repair case set and the multi-dimensional repair case set, the percentage of the same fault type is statistically analyzed, including: Obtain the complete set of fault types corresponding to the alarm signal, wherein the complete set of fault types includes all possible fault types when the alarm signal occurs; The ratio of N to the number of fault types in the comprehensive fault type set is set as the sample generalization coefficient; The sample weight adjustment coefficient is calculated based on the sample generalization coefficient and the data volume difference, and the ratio of the sample weight adjustment coefficient to the initial sample data weight is set as the adapted sample data weight, wherein the initial sample data weight is 2, and the sample weight adjustment coefficient is positively correlated with the sample generalization coefficient and the data volume difference; Configure the weights of multi-data elements for big data retrieval, where the weight of multi-data elements is 0.1; Based on the weights of the adapted sample data and the weights of the multivariate data, the percentage of the same fault type is statistically analyzed according to the sample repair case set and the multivariate repair case set.

5. The data analysis method based on the intelligent operation and maintenance management results of CNC machine tools according to claim 1, characterized in that, If the first detection result is normal, based on the first triggered fault type, fault occurrence prediction is performed on the remaining N-1 associated fault types to obtain N-1 predicted fault probabilities, including: If the first detection result is abnormal, the subsequent analysis shall be stopped, and the CNC machine tool shall be repaired based on the first triggered fault type. If the first detection result is normal, the CNC machine tool's attribute information is used as the equipment constraint, and the alarm signal is used as the condition constraint. Based on big data, the proportion of historical events in which the remaining N-1 associated fault types occur simultaneously when the first triggered fault type occurs is statistically analyzed, and the proportion of events with the same frequency is set as the predicted fault probability to obtain N-1 predicted fault probabilities.

6. The data analysis method based on the intelligent operation and maintenance management results of CNC machine tools according to claim 1, characterized in that, Based on the N-1 predicted fault probabilities, N-1 correlation adjustment coefficients are calculated. The N-1 fault correlations of the remaining N-1 associated fault types are then optimized and corrected, resulting in N-1 optimized fault correlations. Finally, the second triggering fault type is determined, including: The N-1 predicted fault probabilities are summed by adding 1 to each of them to obtain N-1 correlation adjustment coefficients; The N-1 fault correlations are optimized and corrected based on the N-1 correlation adjustment coefficients, and N-1 optimized fault correlations are output. The associated fault type corresponding to the maximum optimized fault correlation is selected as the second trigger fault type.

7. The data analysis method based on the intelligent operation and maintenance management results of CNC machine tools according to claim 1, characterized in that, According to the second triggering fault type, the CNC machine tool is fault detected. If the second detection result is normal, the proportion of historical events in which the remaining N-2 associated fault types occur simultaneously when the first triggering fault type and the second triggering fault type occur is statistically analyzed. The proportion of events with the same frequency is set as the predicted fault probability to obtain N-2 predicted fault probabilities. N-2 correlation adjustment coefficients are then calculated. The N-2 correlation coefficients are used to optimize and correct the correlation of the N-2 faults, and the third trigger fault type is determined.

8. A data analysis system based on the results of intelligent operation and maintenance management of CNC machine tools, characterized in that, The system is used to implement the data analysis method based on the intelligent operation and maintenance management results of CNC machine tools as described in any one of claims 1-7, and the system comprises: The case acquisition module is used to input the alarm signals of CNC machine tools into the local operation and maintenance database for retrieval and to obtain a sample maintenance case set; The first fault type acquisition module is used to perform fault type proportion statistics based on the sample maintenance case set if the sample data volume of the sample maintenance case set meets the preset index, output N associated fault types and N fault correlation degrees, and determine the first triggering fault type. The fault probability prediction module is used to perform fault detection on the CNC machine tool according to the first triggered fault type. If the first detection result is normal, the module performs fault occurrence prediction on the remaining N-1 associated fault types based on the first triggered fault type to obtain N-1 predicted fault probabilities. The second fault type acquisition module is used to calculate N-1 correlation adjustment coefficients based on the N-1 predicted fault probabilities, optimize and correct the N-1 fault correlations of the remaining N-1 associated fault types, output N-1 optimized fault correlations, and determine the second trigger fault type. The loop detection module is used to perform fault detection on the CNC machine tool according to the second triggered fault type. If the second detection result is normal, the iterative loop of correlation adjustment-triggered fault type screening-machine tool fault detection continues until the detection result is abnormal.

Images