A Fault Prediction and Diagnosis System and Method for CNC Machine Tools Based on Deep Learning

Abstract

The present invention relates to the technical field of fault prediction and diagnosis of numerical control machine tools, and specifically provides a numerical control machine tool fault prediction and diagnosis system and method based on deep learning, including: obtaining first target data of a numerical control machine tool, processing and analyzing the first target data, and constructing a standard attenuation model; optimizing the standard attenuation model by using a convergence analysis method to improve the accuracy of the output of the standard attenuation model; calculating the real-time standard attenuation amount of the numerical control module through the standard attenuation model and inputting it into the fault prediction model to output the safety status and maintenance suggestions of the numerical control module. By monitoring data and predicting faults of the numerical control machine tool in real time, the present invention can discover potential fault hazards in advance and put forward maintenance suggestions before the faults occur, avoiding long-term downtime caused by sudden faults, thereby improving production efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of fault prediction and diagnosis of numerical control machine tools, and particularly relates to a numerical control machine tool fault prediction and diagnosis system and method based on deep learning. Background Technique

[0002] A numerical control machine tool is an automated device that controls the movement and machining process of the machine tool through digital information. It uses computer programs and preset instructions to precisely control the relative position and movement of the tool and the workpiece, thereby achieving high-precision and high-efficiency machining operations.

[0003] Due to the high precision of numerical control machine tool machining, when some functional components in the machine tool experience wear or aging, etc., it may have a great impact on the machining accuracy of the numerical control machine tool, and the numerical control machine tool needs to be maintained; traditional regular maintenance may not be able to detect structural component failures of the numerical control machine tool in a timely manner, affecting the overall quality of machining. At the same time, it may also lead to unnecessary maintenance work and increase maintenance costs; therefore, there is an urgent need for a numerical control machine tool fault prediction and diagnosis system and method based on deep learning to discover potential fault hazards in advance and propose maintenance suggestions before a fault occurs, avoiding long-term downtime caused by sudden faults, thereby improving production efficiency. Summary of the Invention

[0004] The object of the present invention is to address the problems in the background technique and propose a numerical control machine tool fault prediction and diagnosis system and method based on deep learning.

[0005] The technical solution of the present invention: A numerical control machine tool fault prediction and diagnosis system and method based on deep learning includes the following steps:

[0006] S1. Obtain the first target data of the numerical control machine tool, process and analyze the first target data, construct a standard attenuation model, and obtain the degree of influence of the wear and aging of the relevant structural components of the numerical control machine tool on the functions of the numerical control machine tool through the standard attenuation model;

[0007] S2. Optimize the standard attenuation model using the convergence analysis method to improve the accuracy of the output of the standard attenuation model;

[0008] S3. Obtain the second target data of the numerical control machine tool, process the second target data, determine the first maintenance value and the second maintenance value, calculate the real-time standard attenuation amount of the numerical control module through the standard attenuation model, and input it into the fault prediction model to output the safety status and maintenance suggestions of the numerical control module.

[0009] Preferably, for S1, the method of obtaining the first target data of the numerical control machine tool, processing and analyzing the relevant data, and constructing a standard attenuation model includes the following steps:

[0010] S11. Obtain the full modular information of the CNC machine tool, determine the numerical control module of the CNC machine tool, and the corresponding module functions, structural components and component sub-functions of the numerical control module;

[0011] S12. Determine the function attenuation information according to the module function, conduct in-depth analysis based on the component sub-functions, determine each possible factor that causes the function attenuation information, and combine all possible factors as the attenuation mapping set of the numerical control module;

[0012] S13. Obtain the first operation data of the CNC machine tool, evaluate the CNC machine tool based on the first operation data, and obtain the function attenuation tolerance range;

[0013] S14. Obtain the structural component data of the CNC machine tool, calculate the attenuation influence weight of each possible factor in the attenuation mapping set based on the function attenuation tolerance range and the structural component data, obtain the attenuation weight set, and establish the standard attenuation model of the CNC machine tool.

[0014] Preferably, for S13, the method for obtaining the first operation data of the CNC machine tool is:

[0015] Determine the function dimension of the numerical control module according to the function attenuation information;

[0016] Within the target time period, obtain the function parameter change values under different function dimensions of the numerical control module, and generate the first operation data of the structural components of the numerical control module based on the function parameter change values.

[0017] Preferably, for S13, the method for evaluating the CNC machine tool based on the first operation data to obtain the function attenuation tolerance range is:

[0018] Perform data cleaning on the first operation data, remove abnormal data and low-frequency data, and optimize the data structure;

[0019] Perform normalization processing on the first operation data;

[0020] Evaluate the functionality of the numerical control module through the following formula and obtain the function attenuation characteristic value:

[0021] ; (Equation 1)

[0022] In the formula, is the function attenuation characteristic value of the CNC machine tool; u is the number of the CNC machine tool, and u is a positive integer; is the function parameter change value; is the quantization coefficient based on the function dimension; i is the function dimension number, and i is a positive integer;

[0023] Set the parameter tolerance threshold R, and construct the functional attenuation tolerance range Au based on the parameter tolerance threshold R and the functional attenuation eigenvalue AT, such that Au satisfies:

[0024] ; (Equation 2)

[0025] In the formula, Au is the functional attenuation tolerance range of the CNC machine tool, u is the serial number of the CNC machine tool, and u is a positive integer.

[0026] Preferably, for S14, the method for obtaining the structural component data is as follows:

[0027] Number and label the structural components of the numerical control module as j;

[0028] Count the number of possible factors causing the functional attenuation of the numerical control module and label it as Q(j);

[0029] During the target time period, obtain the parameter change values of the structural components in the numerical control module under different possible factors, and use them as the first time period eigenvalues of the possible factors.

[0030] Preferably, for S14, the method for calculating the attenuation influence weight of each possible factor in the attenuation mapping set to obtain the attenuation weight set is as follows:

[0031] Solve the attenuation influence weights of different state dimensions of each structural component through the following set of relational expressions:

[0032] ; (Equation 3)

[0033] In the formula, is the first time period eigenvalue of the possible factor; j is the serial number of the structural component in the numerical control module, j ∈ [1, n], and n is the total number of structural components in the numerical control module; h is the state dimension of the structural component, h is a positive integer, h ∈ [1, Q(j)]; A1, A2,..., are the functional attenuation tolerance ranges of different CNC machine tools; is the attenuation influence weight of the possible factor.

[0034] Preferably, for S14, the method for establishing the standard attenuation model of the CNC machine tool is as follows:

[0035] During the target time period, obtain the time period eigenvalues corresponding to all possible factors in the attenuation mapping set, and establish a time period eigenvalue sequence ;

[0036] Among them, is the sequence element of the time period eigenvalue sequence of the xth possible factor;

[0037] Establish an attenuation influence weight sequence according to the attenuation influence weights of all possible factors in the attenuation mapping set ;

[0038] Among them, is the sequence element of the x-th attenuation influence weight sequence;

[0039] The standard attenuation model expression is as follows:

[0040] ; (Equation 4)

[0041] In the formula, is the standard attenuation amount, is the sequence element of the time period feature sequence, is the sequence element of the attenuation influence weight sequence.

[0042] Preferably, for S3, the method for processing the second target data to determine the first maintenance value and the second maintenance value is as follows:

[0043] Mark the faulty module of the numerically controlled machine tool that has had a fault as the target module;

[0044] Mark the previous target time period when the target module had a fault as the abnormal time period, obtain the parameter change values of all possible factors in the abnormal time period and use them as the second time period feature values, and calculate the standard attenuation amount of the target module in the second time period based on the standard attenuation model and mark it as the target attenuation amount;

[0045] Judge whether the target attenuation amount is not greater than the maximum value of the functional attenuation tolerance range; if the target attenuation amount is not greater than the maximum value of the functional attenuation tolerance range, update the target duration, and based on the updated target duration, update the functional attenuation tolerance range and the standard attenuation model once, and use the once-updated standard attenuation model to update and cover the target attenuation amount;

[0046] Perform a single trace of the target attenuation amount and the functional attenuation tolerance range. If the target attenuation amount is still not greater than the maximum value of the functional attenuation tolerance range, perform a secondary update of the target duration, and perform a secondary trace of the target attenuation amount and the functional attenuation tolerance range until the target duration satisfies that the target attenuation amount is greater than the maximum value of the functional attenuation tolerance range, then mark the maximum value of the functional attenuation tolerance range as the first maintenance value Y1 and mark the target attenuation amount as the second maintenance value Y2;

[0047] Obtain the real-time standard attenuation amount of the numerical control module of the numerically controlled machine tool and mark it as SA, and input the real-time standard attenuation amount SA into the fault prediction model.

[0048] Preferably, for S3, the prediction logic of the fault prediction model is as follows:

[0049] If the real-time standard attenuation amount SA satisfies: SA ≤ Y1, it is determined that the numerical control module is in a safe state and no maintenance is required;

[0050] If the real-time standard attenuation amount SA satisfies: Y1 ≤ SA ≤ Y2, it is determined that the numerical control module is in a warning state and maintenance suggestions are output;

[0051] If the real-time standard attenuation amount SA satisfies: Y2 ≤ SA, it is determined that the numerical control module is in a dangerous state and timely maintenance is required.

[0052] The present invention also discloses a numerical control machine tool fault prediction and diagnosis system based on deep learning, which applies the above-mentioned numerical control machine tool fault prediction and diagnosis method based on deep learning, and specifically includes:

[0053] A standard attenuation model construction module, which is used to obtain the first target data of the numerical control machine tool, process and analyze the first target data, construct a standard attenuation model, and obtain the influence degree of the wear and aging of the relevant structural components of the numerical control machine tool on the function of the numerical control machine tool through the standard attenuation model;

[0054] A standard attenuation model optimization module, which is used to optimize the standard attenuation model by using a convergence analysis method to improve the accuracy of the output of the standard attenuation model;

[0055] A fault prediction model construction module, which is used to obtain the second target data of the numerical control machine tool, process the second target data, determine the first maintenance value and the second maintenance value, calculate the real-time standard attenuation amount of the numerical control module through the standard attenuation model, and input it into the fault prediction model to output the safe state and maintenance suggestions of the numerical control module.

[0056] Compared with the prior art, the above technical solutions of the present invention have the following beneficial technical effects:

[0057] (1) By obtaining the first target data of the numerical control machine tool, processing and analyzing the first target data, constructing a standard attenuation model, and effectively showing the influence degree of the wear and aging of the relevant structural components of the numerical control machine tool on the function of the numerical control machine tool through the standard attenuation model;

[0058] (2) Using a convergence analysis method to optimize the standard attenuation model, improving the accuracy of the output of the standard attenuation model, making the standard attenuation amount of the numerical control machine tool output by it more in line with the data, and enhancing the prediction ability of the fault prediction model;

[0059] (3) Calculate the real-time standard attenuation of the numerical control module through the standard attenuation model, and input it into the fault prediction model to output the safety status and maintenance suggestions of the numerical control module, so as to facilitate real-time data monitoring and fault prediction of the numerically controlled machine tool, discover potential fault hazards in advance, and put forward maintenance suggestions before the occurrence of faults, avoiding long-term downtime caused by sudden faults, thereby improving production efficiency. Brief Description of the Drawings

[0060] Figure 1 It is a flowchart of the method of Embodiment 1 proposed by the present invention.

[0061] Figure 2 It is a flowchart of the method of S1 in Embodiment 1 proposed by the present invention. Detailed Embodiments

[0062] Embodiment 1, as Figure 1 shown, a method for fault prediction and diagnosis of a numerically controlled machine tool based on deep learning proposed by the present invention includes the following steps:

[0063] S1. Obtain the first target data of the numerically controlled machine tool, process and analyze the first target data, construct a standard attenuation model, and obtain the influence degree of the wear and aging of the relevant structural components of the numerically controlled machine tool on the function of the numerically controlled machine tool through the standard attenuation model;

[0064] For S1, the method of obtaining the first target data of the numerically controlled machine tool, processing and analyzing the relevant data, and constructing a standard attenuation model includes the following steps:

[0065] S11. Obtain the full modular information of the numerically controlled machine tool, determine the numerical control module of the numerically controlled machine tool, and the corresponding module functions, structural components and component sub-functions of the numerical control module;

[0066] S12. Determine the function attenuation information according to the module function, conduct in-depth analysis based on the component sub-functions, determine each possible factor that causes the function attenuation information, and combine all possible factors as the attenuation mapping set of the numerical control module;

[0067] The method for determining the function attenuation information according to the module function is:

[0068] Obtain the name of the numerical control module, and retrieve it on the network using the AI large model based on the name to obtain the module function of the numerical control module and the results generated by the attenuation of the component sub-functions, and integrate the text to obtain the function attenuation information;

[0069] Exemplarily, for a vertical machining module of a numerically controlled machine tool, its function is the small-batch and multi-variety production of small precision parts, and its function attenuation information is reduced machining accuracy, reduced machining efficiency, abnormal machining line, machining abnormal noise, etc.;

[0070] The method for analyzing each possible factor that causes the generation of function attenuation information is as follows:

[0071] Using the recurrent neural network technology, perform natural language processing on the function attenuation information and component sub-functions, and list the processing results to obtain each possible factor; it should be noted that the recurrent neural network is a prior art and will not be elaborated here;

[0072] Exemplarily, the vertical machining module includes a spindle structure and a tool structure. The function of the spindle structure is to ensure the cooperation and connection stability with components such as bearings and tools;

[0073] Then the possible factors for the spindle system to cause the generation of function attenuation information in the CNC machine tool include insufficient spindle accuracy, damaged spindle bearings, poor sealing, etc.;

[0074] The function of the tool system is to output an accurate machining trajectory based on the machining program;

[0075] Then the possible factors for the tool system to cause the generation of function attenuation information in the CNC machine tool include tool wear, improper tool installation, etc.;

[0076] S13. Obtain the first operating data of the CNC machine tool, and evaluate the CNC machine tool based on the first operating data to obtain the function attenuation tolerance range;

[0077] Regarding S13, the method for obtaining the first operating data of the CNC machine tool is as follows:

[0078] Determine the function dimension of the numerical control module according to the function attenuation information;

[0079] Exemplarily, for the function attenuation information of the above vertical machining module, the function dimension of the vertical machining module is 4;

[0080] During the target time period, obtain the function parameter change values of the numerical control module under different function dimensions, and generate the first operating data of the structural components of the numerical control module based on the function parameter change values;

[0081] Exemplarily, at the beginning of the target time period, obtain the first function parameters under four function dimensions respectively. At the end of the target time period, obtain the second function parameters under four function dimensions respectively, and calculate the difference between the first function parameters and the second function parameters to obtain the function parameter change values under different function dimensions;

[0082] It should be particularly noted that the target time period is a time period based on the operating duration of the CNC machine tool, that is, the data for constructing the standard attenuation model is collected based on CNC machine tools with the same operating duration and normal operation, that is, to eliminate the influence of the variable of the operating duration of the CNC machine tool on the accuracy of the standard attenuation model;

[0083] Exemplarily, if the target period is 100 hours, then numerically controlled machine tools of the same batch that have been running for more than 100 hours, and numerically controlled machine tools with the same running time and normal operation are selected as the acquisition objects, and the first target data is collected from these numerically controlled machine tools;

[0084] For S13, the method for evaluating the numerically controlled machine tool based on the first operation data to obtain the functional attenuation tolerance range is as follows:

[0085] Clean the first operation data, remove abnormal data and low-frequency data, and optimize the data structure;

[0086] Normalize the first operation data;

[0087] Evaluate the functionality of the numerical control module through the following formula to obtain the functional attenuation characteristic value:

[0088] ; (Equation 1)

[0089] In the formula, is the functional attenuation characteristic value of the numerically controlled machine tool; u is the number of the numerically controlled machine tool, and u is a positive integer; is the change value of the function parameter; is the quantization coefficient based on the function dimension; i is the function dimension number, and i is a positive integer;

[0090] Set the parameter tolerance threshold R, and construct the functional attenuation tolerance range Au based on the parameter tolerance threshold R and the functional attenuation characteristic value AT, such that Au satisfies:

[0091] ; (Equation 2)

[0092] In the formula, Au is the functional attenuation tolerance range of the numerically controlled machine tool, u is the number of the numerically controlled machine tool, and u is a positive integer;

[0093] S14. Obtain the structural component data of the numerically controlled machine tool, calculate the attenuation influence weight of each possible factor in the attenuation mapping set based on the functional attenuation tolerance range and the structural component data, obtain the attenuation weight set, and establish the standard attenuation model of the numerically controlled machine tool;

[0094] For S14, the method for obtaining the structural component data is as follows:

[0095] Number and label the structural components of the numerical control module as j;

[0096] Count the number of possible factors that cause the functional attenuation of the numerical control module by the structural component and label it as Q(j);

[0097] In the target period, obtain the parameter change values of the structural components in the numerical control module under different possible factors, and use them as the first period characteristic values of the possible factors;

[0098] For S14, the method for calculating the attenuation influence weight of each possibility factor in the attenuation mapping set to obtain the attenuation weight set is as follows:

[0099] Solve the attenuation influence weight of different state dimensions of each structural component through the following relational expression group:

[0100] ; (Equation Three)

[0101] In the formula, is the first period eigenvalue of the possibility factor; j is the structural component number in the numerical control module, j ∈ [1, n], and n is the total number of structural components in the numerical control module; h is the state dimension of the structural component, h is a positive integer, h ∈ [1, Q(j)]; A1, A2, ……, is the functional attenuation tolerance range of different numerical control machine tools; is the attenuation influence weight of the possibility factor;

[0102] For S14, the method for establishing the standard attenuation model of the numerical control machine tool is as follows:

[0103] Within the target period, obtain the corresponding period eigenvalues of all possibility factors in the attenuation mapping set to establish a period eigenvalue sequence ;

[0104] Among them, is the sequence element of the period eigenvalue sequence of the xth possibility factor;

[0105] Establish an attenuation influence weight sequence according to the attenuation influence weights of all possibility factors in the attenuation mapping set ;

[0106] Among them, is the sequence element of the xth attenuation influence weight sequence;

[0107] The standard attenuation model expression is as follows:

[0108] ; (Equation Four)

[0109] In the formula, is the standard attenuation amount, is the sequence element of the period eigenvalue sequence, is the sequence element of the attenuation influence weight sequence;

[0110] S2. Use the convergence analysis method to optimize the standard attenuation model to improve the accuracy of the output of the standard attenuation model;

[0111] For S2, the method for optimizing the standard attenuation model using the convergence analysis method is as follows:

[0112] Set the reduction radius r of the parameter tolerance threshold R, and based on the formula reset the parameter tolerance threshold R and update the functional attenuation tolerance range Au; where m is the reduction multiple and m is a positive integer;

[0113] Judge the attenuation influence weight of the possibility factor and the satisfaction relationship with the updated functional attenuation tolerance range Au;

[0114] If the attenuation influence weight of the possibility factor does not satisfy the updated functional attenuation tolerance range Au, then the satisfaction relationship is yes, and adjust the attenuation influence weight If the attenuation influence weight of the possibility factor satisfies the updated functional attenuation tolerance range Au, then the satisfaction relationship is no;

[0115] According to the satisfaction relationship being yes, gradually adjust the reduction multiple m until the preset parameter tolerance threshold R is equal to the standard parameter tolerance threshold R0, and obtain the convergence solution set of the attenuation influence weights of different possibility factors; it should be noted that the accuracy of the standard parameter tolerance threshold R0 should be the same as the minimum accuracy of the parameters required by the CNC machine tool;

[0116] S3. Obtain the second target data of the CNC machine tool, process the second target data, determine the first maintenance value and the second maintenance value, calculate the real-time standard attenuation amount of the numerical control module through the standard attenuation model, and input it into the fault prediction model to output the safety state and maintenance suggestions of the numerical control module;

[0117] For S3, the method for processing the second target data to determine the first maintenance value and the second maintenance value is as follows:

[0118] Mark the faulty module of the CNC machine tool that has had a fault as the target module;

[0119] Mark the previous target period when the target module had a fault as the abnormal period, obtain the parameter change values of all possibility factors in the abnormal period as the second period feature values, and calculate the standard attenuation amount of the target module in the second period based on the standard attenuation model and mark it as the target attenuation amount;

[0120] As the exemplary explanation of the target duration above, similarly, select the CNC machine tool that has the same running duration as the CNC machine tool used as the acquisition object of the first running data and has had a fault as the acquisition object of the second target data to eliminate the influence of the running duration of the CNC machine tool on the fault prediction model;

[0121] Determine whether the target attenuation amount is not greater than the maximum value of the functional attenuation tolerance range; if the target attenuation amount is not greater than the maximum value of the functional attenuation tolerance range, update the target duration, and based on the updated target duration, update the functional attenuation tolerance range and the standard attenuation model once, and use the once-updated standard attenuation model to update and cover the target attenuation amount;

[0122] Conduct a one-time trace of the target attenuation amount and the functional attenuation tolerance range. If the target attenuation amount is still not greater than the maximum value of the functional attenuation tolerance range, perform a secondary update on the target duration, and conduct a secondary trace of the target attenuation amount and the functional attenuation tolerance range until the target duration satisfies that the target attenuation amount is greater than the maximum value of the functional attenuation tolerance range. Then, mark the maximum value of the functional attenuation tolerance range as the first maintenance value Y1, and mark the target attenuation amount as the second maintenance value Y2;

[0123] Obtain the real-time standard attenuation amount of the numerical control module of the CNC machine tool and mark it as SA, and input the real-time standard attenuation amount SA into the fault prediction model;

[0124] Regarding S3, the prediction logic of the fault prediction model is as follows:

[0125] If the real-time standard attenuation amount SA satisfies: SA ≤ Y1, it is determined that the numerical control module is in a safe state and no maintenance is required;

[0126] If the real-time standard attenuation amount SA satisfies: Y1 ≤ SA ≤ Y2, it is determined that the numerical control module is in a warning state and maintenance suggestions are output;

[0127] If the real-time standard attenuation amount SA satisfies: Y2 ≤ SA, it is determined that the numerical control module is in a dangerous state and needs to be maintained in a timely manner.

[0128] Embodiment 2. A CNC machine tool fault prediction and diagnosis system based on deep learning proposed by the present invention is applied to a CNC machine tool fault prediction and diagnosis method proposed in Embodiment 1, and specifically includes:

[0129] A standard attenuation model construction module, which is used to obtain the first target data of the CNC machine tool, process and analyze the first target data, construct a standard attenuation model, and obtain the influence degree of the wear and aging of the relevant structural components of the CNC machine tool on the function of the CNC machine tool through the standard attenuation model;

[0130] A standard attenuation model optimization module, which is used to optimize the standard attenuation model by using a convergence analysis method to improve the accuracy of the output of the standard attenuation model;

[0131] The fault prediction model construction module is used to obtain the second target data of the CNC machine tool, process the second target data, determine the first maintenance value and the second maintenance value, calculate the real-time standard attenuation amount of the numerical control module through the standard attenuation model, and input it into the fault prediction model to output the safety status and maintenance suggestions of the numerical control module.

[0132] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited thereto, and various changes can be made without departing from the spirit of the present invention within the scope of knowledge possessed by those skilled in the art to which the present invention pertains.

Claims

1. A method for fault prediction and diagnosis of numerically controlled machine tools based on deep learning, characterized in that, It includes the following steps: S1. Obtain the first target data of the CNC machine tool, process and analyze the first target data, construct a standard attenuation model, and obtain the degree of influence of the wear and aging of the relevant structural components of the CNC machine tool on the functions of the CNC machine tool through the standard attenuation model; Regarding S1, the method of obtaining the first target data of the CNC machine tool, processing and analyzing the relevant data, and constructing a standard attenuation model includes the following steps: S11. Obtain the full modular information of the CNC machine tool, determine the numerical control module of the CNC machine tool, as well as the corresponding module functions, structural components, and component sub-functions of the numerical control module; S12. Determine the function attenuation information according to the module function, conduct in-depth analysis based on the component sub-functions, determine each possible factor that causes the function attenuation information, and combine all possible factors as the attenuation mapping set of the numerical control module; S13. Obtain the first operation data of the CNC machine tool, evaluate the CNC machine tool based on the first operation data, and obtain the function attenuation tolerance range; Regarding S13, the method of evaluating the CNC machine tool based on the first operation data to obtain the function attenuation tolerance range is: Clean the first operation data, remove abnormal data and low-frequency data, and optimize the data structure; Perform normalization processing on the first operation data; Evaluate the functionality of the numerical control module through the following formula and obtain the function attenuation characteristic value: ; (Formula 1) In the formula, is the function attenuation characteristic value of the CNC machine tool; u is the number of the CNC machine tool, and u is a positive integer; is the change value of the function parameter; is the quantization coefficient based on the function dimension; i is the function dimension number, and i is a positive integer; Set the parameter tolerance threshold R, and based on the parameter tolerance threshold R and the functional degradation eigenvalue Construct the functional degradation tolerance range Au such that Au satisfies: ; (Formula II) In the formula, Au is the function attenuation tolerance range of the CNC machine tool, u is the number of the CNC machine tool, and u is a positive integer; S14. Obtain the structural component data of the CNC machine tool, calculate the attenuation influence weight of each possible factor in the attenuation mapping set based on the function attenuation tolerance range and the structural component data, obtain the attenuation weight set, and establish the standard attenuation model of the CNC machine tool; Regarding S14, the method of obtaining the structural component data is: Number and mark the structural components of the numerical control module as j; Count the number of possible factors that cause the function attenuation of the structural components and mark it as Q(j); In the target time period, obtain the parameter change values of the structural components in the numerical control module under different possible factors and use them as the first time period characteristic values of the possible factors; The method of calculating the attenuation influence weight of each possible factor in the attenuation mapping set to obtain the attenuation weight set is: Solve the attenuation influence weight of different state dimensions of each structural component through the following relational expression group: ; (Formula III) In the formula, is the first period eigenvalue of the possibility factor; j is the structural component number in the numerical control module, j ∈ [1, n], where n is the total number of structural components in the numerical control module; h is the state dimension of the structural component, h is a positive integer, h ∈ [1, Q(j)]; A1, A2, ……, is the function attenuation tolerance range of different numerically controlled machine tools; is the attenuation influence weight of the possibility factor; The method of establishing the standard attenuation model of the CNC machine tool is: During the target time period, obtain the corresponding time period characteristic values of all possible factors in the attenuation mapping set, and establish a time period characteristic sequence ; Among them, is the sequence element of the time period feature sequence of the x-th possible factor; Establish a decay influence weight sequence according to the decay influence weights of all possible factors in the decay mapping set ; wherein, is the sequence element of the x-th attenuation influence weight sequence; The expression of the standard attenuation model is as follows: ; (Formula IV) In the formula, is the standard attenuation amount, is the sequence element of the time period feature sequence, is the sequence element of the attenuation influence weight sequence; S2. Optimize the standard attenuation model by using the convergence analysis method to improve the accuracy of the output of the standard attenuation model; S3. Obtain the second target data of the CNC machine tool, process the second target data, determine the first maintenance value and the second maintenance value, calculate the real-time standard attenuation amount of the numerical control module through the standard attenuation model, and input it into the fault prediction model to output the safety status and maintenance suggestions of the numerical control module.

2. The method for fault prediction and diagnosis of a numerically controlled machine tool based on deep learning according to claim 1, characterized in that Regarding S13, the method of obtaining the first operation data of the CNC machine tool is: Determine the function dimension of the numerical control module according to the function attenuation information; During the target time period, obtain the change values of the function parameters under different function dimensions of the numerical control module, and generate the first operation data of the structural components of the numerical control module based on the change values of the function parameters.

3. A fault prediction and diagnosis method for a numerically controlled machine tool based on deep learning according to claim 2, characterized in that, For S3, the method for processing the second target data to determine the first maintenance value and the second maintenance value is as follows: Mark the faulty module of the numerically controlled machine tool that has experienced a fault as the target module; Mark the target time period immediately before the target module fails as the abnormal time period, obtain the parameter change values of all possible factors in the abnormal time period and use them as the second time period characteristic values, and calculate the standard attenuation amount of the target module in the second time period based on the standard attenuation model and mark it as the target attenuation amount; Judge whether the target attenuation amount is not greater than the maximum value of the function attenuation tolerance range; if the target attenuation amount is not greater than the maximum value of the function attenuation tolerance range, update the target duration, and based on the updated target duration, perform an update on the function attenuation tolerance range and the standard attenuation model once, and use the once-updated standard attenuation model to update and overwrite the target attenuation amount; Perform a single trace on the target attenuation amount and the function attenuation tolerance range. If the target attenuation amount is still not greater than the maximum value of the function attenuation tolerance range, perform a secondary update on the target duration, and perform a secondary trace on the target attenuation amount and the function attenuation tolerance range until the target duration meets the condition that the target attenuation amount is greater than the maximum value of the function attenuation tolerance range, then mark the maximum value of the function attenuation tolerance range as the first maintenance value Y1, and mark the target attenuation amount as the second maintenance value Y2.

4. A fault prediction and diagnosis method for numerically controlled machine tools based on deep learning according to claim 3, characterized in that, For S3, the prediction logic of the fault prediction model is as follows: If the real-time standard attenuation amount SA satisfies: SA ≤ Y1, it is determined that the numerical control module is in a safe state and no maintenance is required; If the real-time standard attenuation amount SA satisfies: Y1 ≤ SA ≤ Y2, it is determined that the numerical control module is in a warning state and maintenance suggestions are output; If the real-time standard attenuation amount SA satisfies: Y2 ≤ SA, it is determined that the numerical control module is in a dangerous state and needs to be maintained in a timely manner.

5. A fault prediction and diagnosis system for numerically controlled machine tools based on deep learning, which applies a fault prediction and diagnosis method for numerically controlled machine tools based on deep learning as described in any one of claims 1 to 4 above, characterized in that, Specifically include: A standard attenuation model construction module, which is used to obtain the first target data of the numerically controlled machine tool, process and analyze the first target data, construct a standard attenuation model, and obtain the influence degree of the wear and aging of the relevant structural components of the numerically controlled machine tool on the functions of the numerically controlled machine tool through the standard attenuation model; A standard attenuation model optimization module, which is used to optimize the standard attenuation model by using a convergence analysis method to improve the accuracy of the output of the standard attenuation model; A fault prediction model construction module, which is used to obtain the second target data of the numerically controlled machine tool, process the second target data, determine the first maintenance value and the second maintenance value, calculate the real-time standard attenuation amount of the numerical control module through the standard attenuation model, and input it into the fault prediction model to output the safe state and maintenance suggestions of the numerical control module.

Images