Data separation method and separation system

Abstract

The present invention provides a data separation method and a separation system, wherein the data separation method comprises the following steps: obtaining a full-cycle processing curve during workpiece processing, wherein the full-cycle processing curve comprises a plurality of single-cycle processing curves; selecting one of the single-cycle processing curves in the full-cycle processing curve as a template curve, and the other single-cycle processing curves as processing curves to be matched; determining a standard sample curve in the template curve; judging whether the standard sample curve matches each processing curve to be matched; judging whether to update the standard sample curve based on the matching results between n adjacent processing curves to be matched that match the standard sample curve and the standard sample curve; and separating the template curve before the update and all processing curves to be matched that match the standard sample curve in the template curve.

Description

Technical Field

[0001] The present invention relates to the field of field data processing, and in particular to a data separation method and a separation system for periodically processed data. Background Art

[0002] In industrial production, machine tool processing data is crucial for analyzing the quality of finished products. Therefore, it's necessary to collect and analyze data from multiple production stages to determine whether the finished product meets production requirements. However, in actual production, many machine tools are unable to distinguish between workpieces and their corresponding processing steps, and instead continuously transmit and store all processing data. Therefore, extracting the actual processing data of individual workpieces from this stored data is crucial.

[0003] In addition, how to extract single processing cycle data in real time during the processing process, and how to properly deal with the impact of tool wear as the processing progresses, are issues that urgently need to be solved today. Summary of the Invention

[0004] In order to solve the problems existing in the background technology, in a first aspect, the present invention provides a data separation method, which includes the following steps: obtaining a full-cycle processing curve during workpiece processing, the full-cycle processing curve including multiple single-cycle processing curves; selecting one of the single-cycle processing curves in the full-cycle processing curve as a template curve, and the other single-cycle processing curves as processing curves to be matched; determining a standard sample curve in the template curve; judging whether the standard sample curve matches each processing curve to be matched; judging whether to update the standard sample curve based on the matching results between n adjacent processing curves to be matched that match the standard sample curve and the standard sample curve; separating the template curve before updating and all processing curves to be matched that match the standard sample curve in the template curve.

[0005] Furthermore, the standard sample curve includes the starting point of the template curve.

[0006] Furthermore, the standard sample curve is part or all of the template curve.

[0007] Furthermore, determining a standard sample curve from the template curve includes the following steps: selecting at least a portion of the template curve as a sample curve; determining a sample sequence in the sample curve; and calculating the correlation between the sample sequence in the sample curve and each data sequence in the template curve. If there is only one data sequence in the template curve that is related to the sample sequence in the sample curve, then the sample curve is the standard sample curve. Otherwise, reselecting a sample curve from the template curve until there is only one data sequence in the template curve that is related to the sample sequence in the sample curve, at which point the selected sample curve is the standard sample curve.

[0008] Furthermore, determining the sample sequence in the sample curve includes the steps of: judging whether the sample curve is the first sample curve or the second sample curve based on the sample quantity in the template curve; if the sample curve is the first sample curve, all the data in the sample curve constitute the sample sequence in the sample curve; if the sample curve is the second sample curve, finding all the feature point data in the template curve, and at least part of the feature point data in the second sample curve constitutes the sample sequence in the sample curve.

[0009] Furthermore, the characteristic points are peak points and / or trough points.

[0010] Furthermore, determining whether the standard sample curve matches each of the to-be-matched processed curves includes the steps of: calculating the correlation between the sample sequence in the standard sample curve and the corresponding data sequence in each of the to-be-matched processed curves; comparing the correlation result with a preset threshold; if the correlation result is not less than the threshold, the match is successful; otherwise, the match fails.

[0011] Furthermore, the threshold is determined during a learning phase prior to separation, specifically by the following method: obtaining a full-cycle machining curve for workpiece machining, the full-cycle machining curve comprising multiple single-cycle machining curves; selecting one of the single-cycle machining curves from the full-cycle machining curve as a template curve; determining a standard sample curve from the template curve; calculating the correlation between the standard sample curve and the corresponding data sequence in each of the machining curves to be matched, to obtain a correlation calculation result curve; if the difference between the minimum value of the highest peak in the correlation calculation result curve and the maximum value of the secondary peak in each machining cycle is greater than 0.1, or if there is no secondary peak, then using -0.05 of the minimum value of the highest peak in each machining cycle as the threshold. Conversely, using the average of the minimum value of the highest peak and the maximum value of the secondary peak in the correlation calculation result curve as the threshold.

[0012] Furthermore, according to the matching results between the adjacent n processing curves to be matched that match the standard sample curve and the standard sample curve, it is determined whether to update the standard sample curve, including the steps of: sequentially numbering the matching results between the standard sample curve and the processing curve to be matched that matches the standard sample curve, and recording them as H m , m is a positive integer; calculate the H m-n Correlation calculation results to H m The average value of the correlation calculation results is recorded as K, where n is a positive integer less than m; take the values ​​from H1 to H m-1 The result of correlation calculation forms a sequence Q; if K≤95th percentile of Q, then extract the sequence Q m All or part of the to-be-matched processing curve corresponding to the correlation calculation result is used as a new standard sample curve, and the new standard sample curve is used for subsequent matching; if K>95th percentile of Q, there is no need to update the standard sample curve, and the subsequent matching continues.

[0013] In a second aspect, the present invention also provides a separation system, which includes: a data acquisition module, a selection module, a determination module, a judgment module, an update module and a separation module; the data acquisition module acquires the full-cycle processing curve during workpiece processing, and the full-cycle processing curve includes multiple single-cycle processing curves; the selection module is used to select one of the single-cycle processing curves in the full-cycle processing curve as the template curve, and the other single-cycle processing curves are processing curves to be matched; the determination module is used to determine the standard sample curve in the template curve; the judgment module is used to judge whether the standard sample curve matches each processing curve to be matched; the update module is used to judge whether to update the standard sample curve based on the matching results between the adjacent n processing curves to be matched that match the standard sample curve and the standard sample curve; the separation module is used to separate all processing curves to be matched that match the template curve and the standard sample curve in the template curve before updating.

[0014] The beneficial effects of the present invention are as follows: the data separation method of the present invention matches the standard sample curve and the full-cycle processing curve to obtain a single-cycle curve that meets the matching requirements; and by setting a standard sample curve update step, it overcomes the abnormal matching failure caused by the deviation of the processing curve as the processing progresses, thereby improving the accuracy of the separation method of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 is a flow chart of the data separation method of the present invention. DETAILED DESCRIPTION

[0016] To make the purpose, technical solutions, and advantages of the embodiments of this application more clear, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0017] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application belongs; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" and any variations thereof in the specification and claims of this application and the above-mentioned figure descriptions are intended to cover non-exclusive inclusions.

[0018] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute an independent or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0019] The data separation method of the present invention will be described in detail below with reference to the accompanying drawings. Figure 1 As shown, the following steps S1-S6 are included:

[0020] S1. Obtain the full-cycle processing curve during workpiece processing, wherein the full-cycle processing curve includes multiple single-cycle processing curves; S2. Select one of the single-cycle processing curves in the full-cycle processing curve as the template curve, and the other single-cycle processing curves are the processing curves to be matched; S3. Determine a standard sample curve in the template curve; S4. Determine whether the standard sample curve matches each processing curve to be matched; S5. Determine whether to update the standard sample curve based on the matching results between the adjacent n processing curves to be matched that match the standard sample curve and the standard sample curve; S6. Separate the template curve before the update and all processing curves to be matched that match the standard sample curve in the template curve. The data separation method of the present invention obtains a single-cycle curve that meets the matching requirements by matching the standard sample curve with the full-cycle processing curve; and by setting a standard sample curve update step, it overcomes the abnormal matching failure caused by the offset of the processing curve as the processing progresses, thereby improving the accuracy of the separation method of the present invention.

[0021] The following is a detailed description of the above steps:

[0022] In step S2, one of the single-cycle processing curves in the full-cycle processing curve is selected as the template curve, and the other single-cycle processing curves are the processing curves to be matched; the data length of the single-cycle processing curve is recorded as T, and T lengths of data of a single cycle are intercepted on the full-cycle curve, which can be intercepted manually.

[0023] In step S3, determining a standard sample curve from the template curve includes the following steps:

[0024] S31. Select at least a portion of the template curve as a sample curve; for example, select 1 / 3 of the length of the template curve including the starting point as the sample curve for comparison, or select a curve whose length including the starting point is less than the length of the template curve. Similarly, the complete template curve data can be directly used for correlation calculation. Preferably, the standard sample curve includes the starting point of the template curve, and a portion of the template curve including the starting point is selected for matching. This simplifies the calculation on the one hand; on the other hand, it facilitates the determination of the starting point of the single-cycle processing curve in subsequent steps, thereby facilitating the separation of a single processing cycle curve and the extraction of the single-cycle curve.

[0025] S32. Determine the sample sequence in the sample curve (ie, the data points of the sample curve);

[0026] Wherein, step 32 determines the sample sequence in the sample curve, including the following steps:

[0027] S321. Determine whether the sample curve is the first sample curve or the second sample curve based on the sample size (data length) in the template curve; select different calculation benchmarks (continuous sample curves or feature points) in subsequent steps based on the sample size in the template curve, so as to reduce the subsequent correlation calculation processing volume and improve the separation processing efficiency.

[0028] As one approach, in step S321, the following method can be used to determine whether the sample curve is the first sample curve or the second sample curve. The sample size (number of data points, N) of the template curve (single machining cycle curve) is used to determine whether the first sample curve or the second sample curve is used. If N*N / (CPU main frequency (Hz)*0.8) < sampling frequency (Hz), the curve is the first sample curve, where the sampling frequency is a system parameter of the machine tool. Otherwise, the curve is the second sample curve.

[0029] S322. If the sample curve is the first sample curve, all template curve data in the first sample curve constitute the sample sequence in the sample curve; that is, all or part of the template curve, and proceed to step S33.

[0030] If the sample curve is the second sample curve, all feature point data in the template curve is found. At least some of the feature point data in the second sample curve constitutes the sample sequence in the sample curve, and the process proceeds to step S33. The feature point data includes at least feature point location information and measurement parameter information (such as power, feed rate, coordinates, and other low-frequency varying data).

[0031] Preferably, the characteristic points are peak points and / or trough points. Among them, the method for determining peaks and troughs in the peak and trough extraction method can refer to existing algorithms, such as the find_peaks method in scipy.singal. In addition to the above-mentioned selection of peak points and / or trough points as characteristic points, uniform downsampling (uniform downsampling means selecting sample points at equal distances as characteristic points in the selected sample curve to form a sample sequence, and then using the sample sequence to perform correlation calculation with the processing curve to be matched.) can also be used for correlation calculation. The uniform downsampling method has the same purpose as the peak / trough extraction, both of which are to reduce the amount of calculation on the basis of completing the processing cycle matching.

[0032] S33. Calculate the correlation between the sample sequence in the sample curve and each data sequence in the template curve; S33A. If the template curve contains only one data sequence related to the sample sequence in the sample curve, then the sample curve is the standard sample curve; specifically, if the calculated result curve has only one highest peak, and the difference between the highest peak and the second highest peak is greater than or equal to 0.1, then the sample curve can be used as the standard sample curve; then the matching determination of step S4 is performed using this standard sample curve.

[0033] S33B. On the contrary (i.e., the difference between the highest peak value and the second highest peak value in the calculated result curve is less than 0.1), a new sample curve is selected from the template curve until there is only one data sequence in the template curve that is correlated with the sample sequence in the sample curve (correlation here refers to the highest correlation). At this time, the selected sample curve is the standard sample curve, and steps S31-33 are repeated.

[0034] In step S4, determining whether the standard sample curve matches each of the to-be-matched processing curves comprises the following steps:

[0035] S41. Calculate the correlation between the sample sequence in the standard sample curve and the corresponding data sequence in each of the processed curves to be matched respectively; the correlation coefficient obtained by the correlation calculation is an indicator reflecting the degree of linear correlation between two variables.

[0036] The specific correlation calculation process is as follows:

[0037] The selected standard sample curve is used to perform sliding calculations on the processing curve to be matched. The sliding distance each time is the distance of one data point (that is, the interval between two data points, specifically the interval between the horizontal coordinates of the processing cycle curve to be matched). After each sliding, the correlation calculation is performed between the sample sequence corresponding to the same data point and the data sequence corresponding to the processing curve to be matched.

[0038] When calculating correlation using a feature point set, the specific steps are as follows:

[0039] The distance between adjacent feature points (the interval between the subscripts (abscissas)) is calculated using the acquired feature point information (position information, i.e., the horizontal coordinate value of the processing curve). When performing correlation calculations, after each slide, the distance between the feature points is used to determine the corresponding positions of other data points, and the measurement parameter values ​​of the processing curve to be matched corresponding to the data points are extracted. Correlation calculations are performed using the feature point values ​​and the measurement parameter values ​​corresponding to the periodic curve to be matched.

[0040] The calculation formula used in the correlation calculation is as follows:

[0041] Pearson correlation coefficient: Pearson correlation coefficient, a coefficient used to measure the similarity between two continuous random variables.

[0042] Spearman correlation coefficient: The Spearman correlation coefficient is a rank correlation coefficient. It is calculated based on the order of the original data and is also defined as the Pearson correlation coefficient between ranked variables.

[0043] Both methods reflect the direction and degree of change between two variables, with values ​​ranging from -1 to 1. Values ​​close to 1 indicate a strong positive correlation; values ​​close to -1 indicate a strong negative correlation; and values ​​close to 0 indicate a low correlation.

[0044] Among them, the Pearson correlation coefficient (PearsonCorrelationCoefficient) is applicable when two variables are normally distributed continuous variables and there is a linear relationship between the two. In this case, Pearson can be used to calculate the correlation coefficient;

[0045] The calculation formula is as follows:

[0046]

[0047] The meaning of related parameters:

[0048] 1. Correlation coefficient |ρ X,Y |, which quantitatively characterizes the variables X(x1, x2...x i ) and variables Y(y1, y2...yi ), namely |ρ X,Y The larger the |, the greater the correlation; X,Y |=0 corresponds to the lowest correlation;

[0049] 2. X and Y are perfectly correlated in the sense that they have a linear relationship with probability 1, so |ρ X,Y | is a quantity that can characterize the closeness of the linear relationship between X and Y. When |ρ X,Y When | is large, it means that X and Y are well correlated; when |ρ X,Y When | is small, it means that X and Y are less correlated.

[0050] The following example uses power values ​​for calculation:

[0051] The calculation formula is shown in Pearson correlation coefficient formula (1).

[0052] Standard sample sequence:

[0053] x=[0.0,0.0,0.0,0.0..........433.919,444.078,436.335,409.304],

[0054] Measurement data to be matched to the machining cycle:

[0055] y=[0.0,0.0,0.0,0.0..........444.078,436.335,409.304,386.507],

[0056] Calculate the average of x and y and use x with the same order i and y i The correlation coefficient can be obtained by bringing it into the calculation; when the correlation calculation is performed between the sample sequence and the processing curve data to be matched, the method is the same as above.

[0057] The Spearman correlation coefficient, also known as the rank correlation coefficient, measures correlation based on the ranks of random variables rather than their raw values. It does not require the distribution of the original variables, making it more widely applicable. The Spearman rank correlation coefficient can be used to measure correlations between variables that do not follow a normal distribution, or between categorical or ordinal variables.

[0058] The calculation formula is as follows:

[0059]

[0060] When calculating, first rank the values ​​of the two variables in pairs from small to large (or from large to small), R i Represents x i The rank, Qi Represents y i The rank, R i -Q i is x i 、y i The difference is , n is the sample size.

[0061] S42. Compare the correlation result with the preset threshold; S42A. If the correlation result is not less than the threshold, the match is successful, and step S5 is performed; S42B. Otherwise, the match fails, and the correlation calculation is continued, i.e., steps S41-42 are repeated. Preferably, after obtaining the calculation result that meets the threshold requirements, (TK) (where T is the data length of the cycle, and K can be max(T / 10,10)) data points are skipped to continue the sliding calculation. This avoids the separation failure caused by multiple calculation results that meet the threshold in the same cycle. Specifically, the data points corresponding to the calculation results that meet the threshold requirements are retained (the points of the processing cycle data curve to be matched corresponding to the starting point of the sample curve). It is advisable to number the calculation results that meet the threshold requirements (i.e., can be extracted as a single-cycle curve) that are matched each time (i.e., the number of the single-cycle curve, such as Hm).

[0062] The threshold in step S42 is determined in the model learning phase before data separation is performed, and is specifically performed as follows:

[0063] Acquire a full-cycle machining curve during workpiece machining, wherein the full-cycle machining curve includes a plurality of single-cycle machining curves;

[0064] Selecting one of the single-cycle processing curves in the full-cycle processing curve as a template curve;

[0065] Determine a standard sample curve from the template curve; (specifically, this may be performed according to the above steps S1-3);

[0066] The correlation between the standard sample curve and the corresponding data sequence in each of the processing curves to be matched is calculated respectively to obtain a correlation calculation result curve; the threshold is determined based on the highest peak and the second highest peak in each single-cycle processing curve in the result curve (when the template data is slid in the single processing cycle to be matched to calculate the correlation and obtain the result, the correlation calculation result forms a correlation calculation result curve distribution. If there is only one highest peak in the correlation calculation result curve, it is considered that a single-cycle curve is matched; or several single-cycle curves can be manually separated and then the correlation calculation is performed on the separated single-cycle curves);

[0067] If the difference between the minimum value of the highest peak in the correlation calculation result curve and the maximum value of the secondary peak in each processing cycle is greater than 0.1, or if there is no secondary peak, then the minimum value of the highest peak in each processing cycle -0.05 is used as the threshold;

[0068] On the contrary (the so-called "on the contrary" refers to other situations different from the above two possible situations), the average value of the minimum value of the highest peak in the correlation calculation result curve and the maximum value of the second highest peak in each processing cycle is used as the threshold.

[0069] In the actual processing, as the processing progresses, the tool will continue to wear, the processing signal data will also change accordingly, and then the processing curve will also change. At this time, if the original standard sample curve is still used for matching, the result will be biased. In this case, if Figure 1 As shown, the separation method of the present invention includes a template data updating step S5: a set cycle threshold is set. Once the cycle threshold is less than the threshold, the template data is updated, and the updated template data is used to match the next processing cycle. This process is repeated in each stage of the actual processing process, and the accuracy of the separation processing cycle is improved.

[0070] In step S5, based on the matching results between the adjacent n processing curves to be matched and the standard sample curve, it is determined whether to update the standard sample curve. The template data updating step is set after each correlation calculation that meets the threshold requirements, specifically including:

[0071] Step S51: The matching results between the standard sample curve and the matching processing curve to be matched are numbered in sequence and recorded as H m , m is a positive integer;

[0072] Step S52. Calculate the Hth m-n Correlation calculation results to H m The average value of the correlation calculation results (the correlation value here is the highest peak in the correlation result curve) is recorded as K, where n is a positive integer less than m; n = 1 / 10 of a tool wear cycle (wear cycle).

[0073] Step S53. Take H1 to H m-1 The correlation calculation results form a sequence Q (Q is a sequence containing m-1 calculation results);

[0074] Step S54. If K≤95th percentile of Q, extract the value corresponding to the Hth m The processing curve to be matched corresponding to the correlation calculation result (numbered H mThe whole or part of the single-cycle processing curve) is used as a new standard sample curve, and the new standard sample curve is used for subsequent matching, that is, steps 3-4 are repeated;

[0075] Step S55: If K>95th percentile of Q, there is no need to update the standard sample curve, and continue with subsequent matching, that is, continue with step S4.

[0076] For example, when performing the matching calculation for the 29th cycle, the mean value (K) of the correlation coefficients of the first n cycles (for example, 3, i.e., the 27th, 28th, and 29th cycles) including the 29th cycle is selected and compared with the 95th percentile value of the correlation calculation results (Q) in the first 28 cycles to determine whether to replace the standard sample curve.

[0077] If the mean K is less than or equal to the 95th percentile of the correlation calculation result sequence Q for periods 1-28, the sample curve from the 29th period is selected as the new standard sample curve and the correlation calculation is continued;

[0078] If the mean K is greater than or equal to the 95th percentile of the correlation calculation result sequence Q of the 1-28 period, then the correlation calculation will continue using the current standard sample curve.

[0079] S6. Isolate all matching processing curves that match the template curve and the standard sample curves within the template curve before updating. This step can be performed after step S54 or after all matching processing curves have completed correlation calculation. Specifically, the curves between the calculation results that meet the threshold requirement are separated from the matching processing curves to obtain the single-cycle processing curve. The curve segment between the horizontal coordinates of the corresponding matching processing curves that meet the threshold requirement (greater than or equal to the threshold) twice in a row is extracted as the single-cycle processing curve.

[0080] Specifically, step S6 may be followed by step S7 to determine whether the matching is complete (i.e., the sliding calculation is complete). The sliding calculation ends when the number of sliding operations equals the number of points in the full-cycle data (or when the right endpoint of the standard sample curve moves to the right endpoint of the full-cycle data curve). The separation of the single-cycle processing curve can be performed uniformly after the calculation is completed; or after each new standard sample curve is replaced. Here, S6 and S7 are merely used to distinguish the steps and do not represent the order of priority.

[0081] The present invention also provides a separation system, which includes: a data acquisition module, a selection module, a determination module, a judgment module, an update module and a separation module;

[0082] A data acquisition module, for acquiring a full-cycle processing curve during workpiece processing, wherein the full-cycle processing curve includes a plurality of single-cycle processing curves;

[0083] A selection module is used to select one of the single-cycle processing curves in the full-cycle processing curve as a template curve, and the other single-cycle processing curves are processing curves to be matched;

[0084] A determination module, configured to determine a standard sample curve from the template curve;

[0085] A judgment module, used to judge whether the standard sample curve matches each processing curve to be matched;

[0086] An updating module, configured to determine whether to update the standard sample curve according to a matching result between n adjacent processing curves to be matched that match the standard sample curve and the standard sample curve;

[0087] The separation module is used to separate all the to-be-matched processing curves that match the template curve and the standard sample curves in the template curve before updating.

[0088] The data separation system of the present invention matches the standard sample curve and the full-cycle processing curve to obtain a single-cycle curve that meets the matching requirements; and by setting a standard sample curve update step, it overcomes the abnormal matching failure caused by the deviation of the processing curve as the processing progresses, thereby improving the accuracy of the separation system of the present invention.

Claims

1. A data separation method, characterized in that: The steps include: Acquire a full-cycle machining curve during workpiece machining, wherein the full-cycle machining curve includes a plurality of single-cycle machining curves; One of the single-cycle processing curves in the full-cycle processing curve is selected as the template curve, and the other single-cycle processing curves are processing curves to be matched; Determining a standard sample curve from the template curve; Determine whether the standard sample curve matches each of the to-be-matched processing curves; determining whether to update the standard sample curve according to the matching results between the n adjacent processing curves to be matched that match the standard sample curve and the standard sample curve; Separating the template curve before updating and all the to-be-matched processing curves that match the standard sample curve in the template curve; The standard sample curve includes the starting point of the template curve; Determining a standard sample curve from the template curve comprises the steps of: selecting at least a portion of the template curve as a sample curve; determining a sample sequence in the sample curve; Calculating the correlation between the sample sequence in the sample curve and each data sequence in the template curve; if the template curve contains only one data sequence related to the sample sequence in the sample curve, the sample curve is the standard sample curve; otherwise, reselecting sample curves in the template curve until the template curve contains only one data sequence related to the sample sequence in the sample curve; at this time, the selected sample curve is the standard sample curve; Determining the sample sequence in the sample curve comprises the steps of: Determining whether the sample curve is the first sample curve or the second sample curve according to the sample amount in the template curve; If the sample curve is the first sample curve, all data in the sample curve constitute a sample sequence in the sample curve; If the sample curve is the second sample curve, all feature point data in the template curve are found, and at least part of the feature point data in the second sample curve constitutes a sample sequence in the sample curve.

2. The data separation method according to claim 1, wherein: The standard sample curve is part or all of the template curve.

3. The data separation method according to claim 1, wherein: The characteristic points are peak points and / or trough points.

4. The data separation method according to claim 1, wherein: Determining whether the standard sample curve matches each of the to-be-matched processing curves comprises the following steps: respectively calculating the correlation between the sample sequence in the standard sample curve and the corresponding data sequence in each of the processed curves to be matched; Comparing the correlation result with a preset threshold, if the correlation result is not less than the threshold, the match is successful; Otherwise, the match fails.

5. The data separation method according to claim 4, characterized in that: The threshold is determined in the learning phase before separation, specifically in the following manner: Acquire a full-cycle machining curve during workpiece machining, wherein the full-cycle machining curve includes a plurality of single-cycle machining curves; Selecting one of the single-cycle processing curves in the full-cycle processing curve as a template curve; Determining a standard sample curve from the template curve; Calculating the correlation between the standard sample curve and the corresponding data sequence in each of the to-be-matched processed curves to obtain a correlation calculation result curve; If the difference between the minimum value of the highest peak in the correlation calculation result curve and the maximum value of the secondary peak in each processing cycle is greater than 0.1, or if there is no secondary peak, then the minimum value of the highest peak in each processing cycle -0.05 is used as the threshold; On the contrary, the average value of the minimum value of the highest peak and the maximum value of the second highest peak in the correlation calculation result curve is used as the threshold.

6. The data separation method according to claim 1, wherein: According to the matching results between n adjacent processing curves to be matched that match the standard sample curve and the standard sample curve, determining whether to update the standard sample curve includes the following steps: The matching results between the standard sample curve and the matching processing curve to be matched are numbered in sequence and recorded as H m , m is a positive integer; Calculate the H m-n Correlation calculation results to H m The average value of the correlation calculation results is recorded as K, where n is a positive integer less than m; Take H1 to H m-1 The result of correlation calculation forms the sequence Q; If K≤95th percentile of Q, then extract m All or part of the to-be-matched processing curve corresponding to the correlation calculation result is used as a new standard sample curve, and the new standard sample curve is used for subsequent matching; If K>95th percentile of Q, there is no need to update the standard sample curve and continue with subsequent matching.

7. A separation system, characterized in that: include: Data acquisition module, selection module, determination module, judgment module, update module and separation module; A data acquisition module, for acquiring a full-cycle processing curve during workpiece processing, wherein the full-cycle processing curve includes a plurality of single-cycle processing curves; A selection module is used to select one of the single-cycle processing curves in the full-cycle processing curve as a template curve, and the other single-cycle processing curves are processing curves to be matched; a determination module, configured to determine a standard sample curve from the template curve, wherein the standard sample curve includes a starting point of the template curve; Determining a standard sample curve from the template curve comprises the steps of: selecting at least a portion of the template curve as a sample curve, determining a sample sequence in the sample curve, Calculate the correlation between the sample sequence in the sample curve and each data sequence in the template curve. If there is only one data sequence in the template curve that is correlated with the sample sequence in the sample curve, then the sample curve is the standard sample curve. Otherwise, reselect sample curves in the template curve until there is only one data sequence in the template curve that is correlated with the sample sequence in the sample curve. At this time, the selected sample curve is the standard sample curve. The step of determining the sample sequence in the sample curve comprises the steps of: Determine whether the sample curve is the first sample curve or the second sample curve according to the sample amount in the template curve, If the sample curve is the first sample curve, all data in the sample curve constitute the sample sequence in the sample curve. If the sample curve is the second sample curve, all feature point data in the template curve are found, and at least part of the feature point data in the second sample curve constitutes a sample sequence in the sample curve; A judgment module, used to judge whether the standard sample curve matches each processing curve to be matched; An updating module, configured to determine whether to update the standard sample curve according to a matching result between n adjacent processing curves to be matched that match the standard sample curve and the standard sample curve; The separation module is used to separate all the to-be-matched processing curves that match the template curve and the standard sample curves in the template curve before updating.

Images