Gear matching surface machining quality online prediction method and related equipment

Abstract

The application discloses a gear matching surface machining quality online prediction method and related equipment, comprising acquiring spindle vibration data and servo current data of a complete machining process of a part, calculating time domain characteristics and frequency domain characteristics of the spindle vibration data, calculating time domain characteristics and frequency domain characteristics of the servo current, using PCA to fuse the time domain and frequency domain characteristics of the spindle vibration data and the time domain and frequency domain characteristics of the servo current, extracting the fused characteristics as input characteristics, inputting the input characteristics into a pre-trained machining quality prediction model of a gear matching surface, obtaining the machining quality of the gear matching surface, the machining quality prediction model of the gear matching surface is obtained through training and testing of multiple parts, the machining quality prediction model of the gear matching surface adopts a random forest model, can improve the noise anti-interference ability of industrial data, has fast model training speed, strong anti-overfitting ability, and is suitable for machining process quality online control analysis.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of part detection, in particular to a gear matching surface machining quality online prediction method and related equipment. BACKGROUND

[0002] With the automobile power system has fully turned to electrification, intelligent, two-stage speed reduction system speed is above 10000RPM, the requirement of NVH is very high. Vibration noise is one of the important factors affecting the automobile NVH, and it will also have a serious impact on the performance of the transmission system, and vibration mainly comes from gear machining error and gear deformation. The precision of traditional gear process cannot meet the high precision requirement of new energy power system. The surface quality of gear reference surface as the key link in gear forming process is particularly important for improving the surface quality of gear in the later process such as gear grinding.

[0003] At present, in order to meet the high precision requirement, the tool is frequently replaced in the production site, and more intensive sampling is adopted to ensure the machining quality, and in the face of frequent production change, multiple measuring tools need to be developed, which makes the machining cost high and the efficiency is not high, and the existing offline detection efficiency is low. SUMMARY

[0004] The present application aims to provide a kind of gear matching surface machining quality online prediction method and related equipment, to overcome the problems of low efficiency of offline detection and long design and manufacture period of special measuring tool in prior art during part machining process.

[0005] To achieve the above-mentioned purpose, the present application provides the following technical scheme:

[0006] A kind of gear matching surface machining quality online prediction method, comprising:

[0007] Obtain the spindle vibration data and servo current data of complete machining process of part;

[0008] Calculate the time domain feature and frequency domain feature of spindle vibration data;Calculate the time domain feature and frequency domain feature of servo current;

[0009] Fuse the time domain and frequency domain features of spindle vibration data and the time domain and frequency domain features of servo current using PCA, and extract the fused features as input features;

[0010] Input the input features into the pre-trained gear matching surface machining quality prediction model to obtain the machining quality of gear matching surface;

[0011] Wherein, the gear matching surface machining quality prediction model is the gear matching surface machining quality prediction model obtained by training and testing of multiple parts, and the gear matching surface machining quality prediction model adopts random forest model.

[0012] Preferably, the top 5 components are extracted as input features.

[0013] Preferably, the gear matching surface machining quality prediction model comprises the following specific training and testing processes:

[0014] Collect spindle vibration and servo current of the complete machining process of multiple parts;

[0015] Collect the key quality parameters of the parts corresponding to each part at the offline detection station of the process;

[0016] Calculate the time domain and frequency domain features of the machining spindle vibration data of each part; calculate the time domain and frequency domain features of the servo current;

[0017] Calculate the correlation between each feature and the key quality parameters of the parts, filter the features according to the correlation, and obtain the filtered features;

[0018] Fuse the filtered features using PCA, extract the top 5 components, and use them as input features; according to the key quality parameters of the parts, divide the key dimensions and surface quality data into qualified and unqualified, and mark them as 1 and -1 as output features;

[0019] According to the input features and corresponding output features, construct a data set, and split the data set into a training set and a test set;

[0020] Construct a random forest model, use the training set as the input and output of the random forest model, train the model parameters, and obtain the trained random forest model; input the input data of the test set into the trained random forest model, obtain the output of the test set, and finally obtain the gear matching surface machining quality prediction model.

[0021] Preferably, the correlation between each feature and the key quality parameters of the parts is calculated according to the following formula:

[0022]

[0023] Wherein, r is the correlation, X is the time domain or frequency domain feature curve, and Y is the key quality parameter curve.

[0024] Preferably, the features with a correlation coefficient less than 0.3 are removed according to the correlation to obtain the filtered features.

[0025] Preferably, the ratio of the training set to the test set is 7:3.

[0026] Preferably, the number of leaves of the random forest model is 5, and the tree parameter is 20.

[0027] A gear matching surface machining quality online prediction system comprises:

[0028] data acquisition module: acquire spindle vibration data and servo current data of the complete machining process of the part;

[0029] calculation module: calculate time domain features and frequency domain features of the spindle vibration data; calculate time domain features and frequency domain features of the servo current;

[0030] extraction module: fuse the time domain and frequency domain features of the spindle vibration data and the time domain and frequency domain features of the servo current using PCA, and extract the fused features as input features;

[0031] prediction module: input the input features into the pre-trained machining quality prediction model of the gear mating surface to obtain the machining quality of the gear mating surface;

[0032] The machining quality prediction model of the gear mating surface is a machining quality prediction model of the gear mating surface obtained by training and testing a plurality of parts, and the machining quality prediction model of the gear mating surface adopts a random forest model.

[0033] A computer device comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the steps of the above-mentioned gear mating surface machining quality online prediction method when executing the computer program.

[0034] A computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the steps of the above-mentioned gear mating surface machining quality online prediction method.

[0035] Compared with the prior art, the present application has the following beneficial effects: the present application provides a gear mating surface machining quality online prediction method, adopts multi-source data to describe the machining process, adopts multi-dimensional features to describe the machining characteristics of each data, and constructs a machining quality online prediction algorithm based on a random forest model. The model has stable precision for describing two labels of qualified and unqualified machining quality, can improve the noise interference resistance of industrial data through random features, has fast model training speed and strong anti-overfitting ability, and is suitable for machining process quality online control analysis.

[0036] Further, the construction of the model establishes a feature screening mechanism based on a linear correlation coefficient with a key quality parameter, avoids the calculation complexity caused by too many input features, picks up data in the machining process, constructs the association relationship between data features and key dimensions, and realizes online analysis and prediction of part machining quality. The present application has the advantages of low algorithm complexity, good robustness, fast calculation speed, good accuracy, can replace part of the offline detection program after machining, shorten the overall beat of the production line, and improve the production efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0037] Figure 1 A flow chart of a gear matching surface machining quality online prediction method of the present application;

[0038] Figure 2 A schematic diagram of a gear matching surface machining quality online prediction system of the present application;

[0039] Figure 3 A flow chart of a gear matching surface machining quality online prediction method of the present application;

[0040] Figure 4 To collect the key dimensions of the detected parts after machining;

[0041] Figure 5 To extract the typical features of the machining process data: effective value, kurtosis and pulse index;

[0042] Figure 6 A comparison chart of the prediction results and actual data in the present application. DETAILED DESCRIPTION

[0043] In order to make the objectives, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described below in connection with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. The components of the embodiments of the present application described and shown in the drawings can be arranged and designed in various different configurations.

[0044] Therefore, the detailed description of the embodiments of the present application provided below in the drawings is not intended to limit the scope of the claimed present application, but only represents selected embodiments of the present application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present application without creative labor are within the scope of protection of the present application.

[0045] It should be noted that: similar reference numerals and letters represent similar items in the following drawings, therefore, once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings.

[0046] In the description of the embodiments of the present application, it should be noted that if the terms "upper", "lower", "horizontal", "inner" and the like indicating the orientation or position relationship are based on the orientation or position relationship shown in the drawings, or the orientation or position relationship when the product of the present application is usually placed, and are only for the convenience of describing the present application and simplifying the description, and therefore cannot be understood as indicating or implying that the device or element referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as limiting the present application. In addition, the terms "first", "second" and the like are only used for differentiation and cannot be understood as indicating or implying relative importance.

[0047] In addition, if the term "horizontal" appears, it does not mean that the component must be absolutely horizontal, but can be slightly inclined. For example, "horizontal" only means that its direction is relatively more horizontal than "vertical", and does not mean that the structure must be completely horizontal, but can be slightly inclined.

[0048] In the description of the embodiments of the present application, it should be noted that unless otherwise explicitly specified and limited, if the terms "arrangement", "installation", "connection", "connection" appear, they should be understood in a broad sense, for example, they can be fixedly connected, or can be detachably connected, or integrally connected; can be mechanically connected, or can be electrically connected; can be directly connected, or can be indirectly connected through an intermediate medium; can be the communication inside two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0049] The present application will be described in further detail below in conjunction with the accompanying drawings:

[0050] As shown in Figure 1 The present application provides a gear matching surface machining quality online prediction method, which comprises:

[0051] S101 acquiring spindle vibration data and servo current data of a complete machining process of a part;

[0052] S102 calculating time domain features and frequency domain features of the spindle vibration data; calculating time domain features and frequency domain features of the servo current;

[0053] S103 using PCA to fuse the time domain and frequency domain features of the spindle vibration data and the time domain and frequency domain features of the servo current, and extracting the fused features as input features;

[0054] S104 inputting the input features into a pre-trained machining quality prediction model of the gear matching surface to obtain the machining quality of the gear matching surface;

[0055] The gear matching surface machining quality prediction model is a gear matching surface machining quality prediction model obtained by training and testing a plurality of parts, and the gear matching surface machining quality prediction model adopts a random forest model.

[0056] The training and detection process of the gear matching surface machining quality prediction model is specifically as follows:

[0057] S1: Collect spindle vibration, servo current and other high-frequency data of the complete machining process of each part at a time;

[0058] S2: Collect the key quality parameters of the part corresponding to each part at the offline detection station of the process;

[0059] S3: Calculate the time domain and frequency domain features of the machining spindle vibration data of each part; calculate the time domain and frequency domain data of the current data;

[0060] S4: Calculate the correlation between each feature and the surface quality change rule of the part, and remove the features with little correlation;

[0061] S5: The above selected features are fused by using PCA, and the top 5 components are extracted as input features; according to the tolerance of the part machining size, the key size and surface quality data of the part are divided into qualified and unqualified, and marked as 1 and -1 as output features;

[0062] S6: Construct a data set containing input features and corresponding output features, and use 70% of the data set as a training set and the remaining 30% as a test set;

[0063] S7: Construct an improved random forest model, use the training set data as the input and output of the random forest model, train the model parameters, and input the test set input data into the trained random forest model to obtain the test set output.

[0064] S8: Determine whether the predicted value is accurate according to the error tolerance range of the key size.

[0065] In the machining quality online prediction method of the improved random forest model driven by multi-source data fusion in S1, the spindle vibration and servo current data of the machining process from when the part enters the machine tool to when it comes out are collected each time, the spindle speed is suddenly increased during machining, which causes the servo current to rise, and a trigger point is set in the data collection software according to this, the automatic collection time is set according to the machining time, and the machining process data is recorded repeatedly in this way;

[0066] The sensitive parameters of the detection process corresponding to S2 include key size parameters such as inner diameter and surface quality parameters;

[0067] In S3, the expression of the time domain index is as follows:

[0068]

[0069]

[0070] The frequency domain index expression is as follows,

[0071]

[0072] S4 adopts the correlation coefficient of the linear coefficient to describe the correlation coefficient of the key quality parameter and the feature in step S3, wherein the expression of the correlation coefficient is as follows,

[0073]

[0074] Wherein X is the time domain and frequency domain feature change curve, and Y is the key quality parameter curve

[0075] The random forest in S7 is based on the decision tree, and the random feature selection is introduced in the decision tree training process, and the information entropy of the data set D is defined as follows

[0076]

[0077] The information gain formula is

[0078]

[0079] For the feature selection decision tree (CART) based on the Gini coefficient,

[0080]

[0081] Then

[0082] Then for the random forest model, the mean of all decision trees is taken as the output of the random forest

[0083]

[0084] In step S8, according to the upper and lower errors of the inner hole size and the end face runout error within the allowable range, it is determined whether the predicted machining quality is accurate.

[0085] Embodiment:

[0086] As Figure 3 The application provides a gear matching surface machining quality online prediction method:

[0087] S1: Collect spindle vibration and servo current data from the time when the part enters the machine tool to the time when it comes out of the machine tool. The current data is collected at a frequency of 1024 Hz, and the vibration data is collected at a frequency of 10240 Hz. According to the spindle speed, the current will rise suddenly during processing, so in the data collection software, the current exceeding 100 A is set as the trigger condition. The automatic collection time is set according to the processing time, and the processing data is recorded repeatedly in this way;

[0088] S2: Since the inner hole surface and the end face are the mating surfaces of the gear, the key size parameters such as the inner diameter of each part in S1 at the offline detection station are collected, and the surface quality parameters are collected, as shown in Figure 4

[0089] S3: Calculate the time domain and frequency domain features of the spindle vibration data of each part; calculate the time domain and frequency domain data of the current data; some typical features are shown in Figure 5

[0090] S4: Use the correlation coefficient to describe the correlation between the key quality parameters and the features in step S3, and the expression of the correlation coefficient is as follows,

[0091]

[0092] where X is the time domain and frequency domain feature curve, and Y is the key quality parameter curve. Remove the features with a correlation coefficient less than 0.3.

[0093] S5: Fuse the above-mentioned screened features using PCA, and extract the top 5 components as input features;

[0094] S6: Use the fused features in step S5 as input features, and use the inner hole size and end face runout as output features to construct a data set, and use 70% of the data set as a training set and the remaining 30% as a test set;

[0095] S7: Construct a random forest model, use the training set data as the input and output of the random forest model, and train the model parameters; input the test set input data into the trained random forest model to get the test set output. In the decision tree training process, introduce random feature selection, and define the information entropy of the data set D as follows

[0096]

[0097] The information gain formula is

[0098]

[0099] For the CART based on the Gini coefficient,

[0100] ​​

[0101] Then

[0102] Then for the random forest model, take the mean of all decision trees as the output of the random forest

[0103]

[0104] In the random forest training process, the number of leaves of the random forest is 5, and the tree parameter is 20. After the model is trained, the prediction result of the inner hole size is as shown in the following table. Figure 6

[0105] Step S9 determines the accuracy of the predicted value according to the accuracy error of the inner hole size, from Figure 6 It can be seen that the total number of key quality data of 17 parts is predicted this time, and the upper and lower lines of the inner hole size are marked out Figure 6 There are 2 prediction errors, false positives, and the online prediction accuracy of the model proposed by the application for the machining quality is 88.25%.

[0106] The application also provides a gear matching surface machining quality online prediction system, comprising:

[0107] A data acquisition module is configured to acquire spindle vibration data and servo current data of a complete machining process of a part.

[0108] A calculation module is configured to calculate time domain features and frequency domain features of the spindle vibration data, and calculate time domain features and frequency domain features of the servo current.

[0109] An extraction module is configured to fuse the time domain and frequency domain features of the spindle vibration data and the time domain and frequency domain features of the servo current using PCA, and extract the fused features as input features.

[0110] A prediction module is configured to input the input features into a pre-trained gear matching surface machining quality prediction model to obtain the machining quality of the gear matching surface.

[0111] The gear matching surface machining quality prediction model is a gear matching surface machining quality prediction model obtained by training and testing a plurality of parts, and the gear matching surface machining quality prediction model adopts a random forest model.

[0112] The terminal device provided by the embodiment of the application. The terminal device of the embodiment includes a processor, a memory, and a computer program stored in the memory and executable on the processor. The processor executes the computer program to implement the steps in each of the above method embodiments. Alternatively, the processor executes the computer program to implement the functions of each module / unit in each of the above device embodiments. ​

[0113] The computer program can be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present application.

[0114] The terminal device can be a desktop computer, a notebook computer, a palm computer, a cloud server and other computing devices. The terminal device can include, but is not limited to, a processor and a memory.

[0115] The processor can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0116] The memory can be used to store the computer program and / or modules. The processor realizes various functions of the terminal device by running or executing the computer program and / or modules stored in the memory, and calling the data stored in the memory.

[0117] The modules / units integrated in the terminal device, if realized in the form of software function units and sold or used as independent products, can be stored in a computer readable storage medium. Based on such understanding, all or part of the processes in the above-mentioned embodiment methods can also be completed by a computer program instructing related hardware, and the computer program can be stored in a computer readable storage medium. When the computer program is executed by a processor, the steps of the above-mentioned various method embodiments can be implemented. The computer program includes computer program code, which can be in the form of source code, object code, executable files or some intermediate forms, etc. The computer readable medium can include any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content included in the computer readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction, for example, in some jurisdictions, according to legislation and patent practice, the computer readable medium does not include electrical carrier signals and telecommunication signals.

[0118] Although the embodiments of the present application are described above in conjunction with the drawings, the present application is not limited to the above-mentioned specific embodiments and application fields, and the above-mentioned specific embodiments are only illustrative and guiding, but not limiting. Those skilled in the art can make many forms under the guidance of the specification without departing from the scope protected by the claims of the present application, and these all belong to the protection of the present application.

Claims

1. A gear-mated surface machining quality online prediction method, characterized in that, The method comprises the following steps: Obtain spindle vibration data and servo current data of the complete machining process of the part; Calculate the time domain features and frequency domain features of the spindle vibration data; Calculate the time domain features and frequency domain features of the servo current; Fuse the time domain and frequency domain features of the spindle vibration data and the time domain and frequency domain features of the servo current using PCA, extract the fused features as input features; Input the input features into a pre-trained machining quality prediction model of the gear mating surface to obtain the machining quality of the gear mating surface; The machining quality prediction model of the gear mating surface is a machining quality prediction model of the gear mating surface obtained by training and testing a plurality of parts, and the machining quality prediction model of the gear mating surface adopts a random forest model.

2. The method of claim 1, wherein, Extract the top 5 order components as input features.

3. The method of claim 1, wherein, The specific training and testing process of the machining quality prediction model of the gear mating surface is as follows: Collect spindle vibration and servo current of the complete machining process of a plurality of parts; Collect the key quality parameters of the parts at the offline detection station; Calculate the time domain and frequency domain features of the spindle vibration data of each part; Calculate the time domain and frequency domain features of the servo current; Calculate the correlation between each feature and the key quality parameters of the part, filter the features according to the correlation, and obtain the filtered features; Fuse the filtered features using PCA, extract the top 5 order components as input features; According to the key quality parameters of the part, the key size and surface quality data are divided into qualified and unqualified, and marked as 1 and -1 as output features; According to the input features and the corresponding output features, a data set is constructed, and the data set is divided into a training set and a test set; Construct a random forest model, use the training set as the input and output of the random forest model, train the model parameters, obtain the trained random forest model, input the input data of the test set into the trained random forest model, obtain the output of the test set, and finally obtain the machining quality prediction model of the gear mating surface.

4. The method of claim 3, wherein, The correlation between each feature and the key quality parameters of the part is calculated according to the following formula: wherein, is a correlation, is a time or frequency domain feature variation curve, and Y is a key quality parameter curve.

5. The method of claim 3, wherein the method is characterized by: According to the correlation, the features with a correlation coefficient less than 0.3 are removed to obtain the filtered features.

6. The method of claim 3, wherein the method is characterized by: The ratio of the training set to the test set is 7:

3.

7. The method of claim 1, wherein, The number of leaves of the random forest model is 5, and the tree parameter is 20.

8. A gear-mated surface machining quality on-line prediction system characterized by, The method comprises the following steps: Data acquisition module: obtain spindle vibration data and servo current data of the complete machining process of the part; Calculation module: calculate the time domain features and frequency domain features of the spindle vibration data; Calculate the time domain features and frequency domain features of the servo current; Extraction module: fuse the time domain and frequency domain features of the spindle vibration data and the time domain and frequency domain features of the servo current using PCA, extract the fused features as input features; Prediction module: input the input features into a pre-trained machining quality prediction model of the gear mating surface to obtain the machining quality of the gear mating surface; The machining quality prediction model of the gear mating surface is a machining quality prediction model of the gear mating surface obtained by training and testing a plurality of parts, and the machining quality prediction model of the gear mating surface adopts a random forest model.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, The computer program is executed by the processor to implement the steps of the gear matching surface machining quality online prediction method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 9. The computer program is executed by the processor to implement the steps of the gear matching surface machining quality online prediction method according to any one of claims 1 to 7.

Images