A method and system for monitoring processing parameters of aluminum plate cutting

Abstract

The present invention relates to the field of data processing, and in particular to a method and system for monitoring processing parameters for aluminum plate cutting. The method includes the following steps: obtaining a number of process parameter data points for aluminum plate cutting; obtaining a number of feasible directions, combining a preset number of mutually perpendicular feasible directions to obtain a number of direction combinations, projecting the process parameter data points to each feasible direction in the direction combination, and calculating the degree of abnormal differentiation of any feasible direction; performing dimensional transformation on the process parameter data points according to each feasible direction in the direction combination with the largest difference in the degree of abnormal differentiation; selecting the transformed dimensions corresponding to some feasible directions with large degrees of abnormal differentiation in the transformed dimensions as the reserved dimensions; constructing an isolation forest based on the reserved dimensions using all process parameter data points to assist in monitoring process parameter abnormalities. Effectively improve the accuracy of abnormality detection, thereby improving the accuracy of process parameter monitoring.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and in particular to a method and system for monitoring processing parameters of aluminum plate cutting. Background Art

[0002] Aluminum plate cutting machines mostly use unmanned cutting control, so aluminum plate cutting machines generally monitor their own quality during the aluminum plate cutting process. During the aluminum plate cutting process, the operating status of the aluminum plate cutting machine is generally judged by analyzing whether there are any abnormalities in the processing parameters.

[0003] There are differences between the process parameters generated when the aluminum plate cutting machine operates normally and the process parameters generated when it operates abnormally. Therefore, the operating status of the aluminum plate cutting machine can be judged by analyzing the outliers of the process parameters. As a commonly used anomaly detection algorithm, the isolation forest algorithm constructs an isolation tree by randomly selecting segmentation points in the values ​​of each dimension to segment the data set. The outlier data is far away from other data, and the outlier data is isolated after several segmentations. The isolation forest algorithm implements anomaly detection based on this feature. Therefore, the greater the distance difference between normal data and abnormal data, the higher the detection accuracy of the isolation forest algorithm. Therefore, how to improve the distance difference between abnormal data and normal data becomes the research focus of this invention.

[0004] The patent application document with publication number CN118710254A discloses a monitoring and maintenance method and system for steel plate laser cutting equipment. The method in the patent application document uses a neural network to analyze operating parameters to determine the operating status of the cutting equipment. The method in the patent application document does not involve an isolation forest algorithm or content that increases the distance difference between abnormal data and normal data. Therefore, the method in the patent application document cannot solve the technical problem of this solution. Summary of the invention

[0005] In order to solve the problem of how to improve the distance difference between abnormal data and normal data, the present invention provides a processing parameter monitoring method and system for aluminum plate cutting.

[0006] In a first aspect, the present invention provides a method for monitoring processing parameters for aluminum plate cutting, which adopts the following technical solution:

[0007] A method for monitoring processing parameters for aluminum plate cutting, comprising the steps of:

[0008] Obtain several process parameter data points of aluminum plate cutting;

[0009] Obtain several feasible directions, combine a preset number of mutually perpendicular feasible directions to obtain several direction combinations, project the process parameter data points to each feasible direction in the direction combination, and calculate the degree of abnormal distinction of any feasible direction , obtain the abnormality degree of each process parameter data point, and cluster the process parameter data points into two categories according to the abnormality degree. The category with a large abnormality degree is recorded as the first category, and the category with a small abnormality degree is recorded as the second category. represents the distance between the projection point of the i-th process parameter data point in the first category and the adjacent projection point in the feasible direction, It represents the distance between the projection point of the i-th process parameter data point in the second category and the adjacent projection point in the feasible direction. Represent the abnormality degree of the i-th process parameter data point in the first category and the second category respectively, , Respectively represent the number of process parameter data points in the first category and the second category; perform dimension transformation on the process parameter data points according to each feasible direction in the direction combination with the largest difference in abnormality differentiation degree;

[0010] The transformed dimensions corresponding to some feasible directions with a large degree of abnormality distinction are selected from the transformed dimensions as retained dimensions; an isolation forest is constructed based on the retained dimensions using all process parameter data points to assist in monitoring process parameter anomalies.

[0011] The present invention uses dimensionality transformation to make the distance difference between process parameter data points with a larger degree of abnormality and process parameter data points with a smaller degree of abnormality in some dimensions larger, so that some dimensions are used as anomaly detection analysis of the isolation forest algorithm to improve the accuracy of anomaly detection; further, in the process of dimensionality transformation, by constructing different dimensionality transformation methods, the distance difference of the adjusted dimension obtained under each dimensionality transformation method is analyzed to accurately screen out the best dimensionality transformation method; further, when screening the best dimensionality transformation method, accurate evaluation results are obtained by analyzing the distance difference of process parameter data points with different degrees of abnormality in the adjusted dimension obtained under each dimensionality transformation method, providing a basis for accurately screening the best dimensionality transformation method.

[0012] Preferably, the obtaining of several feasible directions includes:

[0013] The geometric center of the process parameter data point is obtained, and the surrounding area is evenly divided into a preset number of directions with the geometric center as the center and recorded as feasible directions.

[0014] The method for obtaining a feasible direction according to the present invention is relatively simple and has a higher implementation efficiency.

[0015] Preferably, the combining of a preset number of mutually perpendicular feasible directions to obtain a plurality of direction combinations includes:

[0016] A preset number of dimensions of process parameter data points is set, and a preset number of mutually perpendicular feasible directions are randomly selected and combined together to obtain a plurality of direction combinations.

[0017] The present invention combines mutually perpendicular directions so as to enable the dimensions to be perpendicular after subsequent adjustment, eliminate the correlation between the dimensions, and reduce waste of information description.

[0018] Preferably, obtaining the abnormality degree of each process parameter data point includes:

[0019] Arrange the values ​​of all process parameter data points in any dimension in time series to obtain a value sequence of the dimension, take the value of any process parameter data point in the dimension as the center, and obtain a data segment of a preset length in the value sequence as a reference segment of the process parameter data point in the dimension;

[0020] The mean of the correlations of all two-dimensional value sequences is recorded as the overall correlation; the mean of the correlations of the process parameter data point in all two-dimensional reference segments is recorded as the local correlation of the process parameter data point;

[0021] The degree of abnormality of the process parameter data point is obtained by multiplying the mean of the difference between the value of the process parameter data point in all dimensions and the fitted value by the difference between the overall correlation and the local correlation.

[0022] The present invention reflects the abnormal situation of process parameter data points in time series through correlation and fitting deviation, avoids repeated analysis of abnormal situations in space, and realizes simple and efficient abnormal analysis.

[0023] Preferably, the dimensional transformation of the process parameter data points according to each feasible direction in the direction combination with the largest difference in abnormality differentiation degree includes:

[0024] Calculate the variance of the abnormality discrimination degree of all feasible directions in the direction combination, record it as the abnormality discrimination degree difference of the direction combination, and take each feasible direction in the direction combination with the largest abnormality discrimination degree difference as the adjusted direction;

[0025] The process parameter data points are projected to each adjusted direction, and the dimensions formed by the projection values ​​of all the process parameter data points in each adjusted direction are used as the transformed dimensions.

[0026] Preferably, the step of selecting transformed dimensions corresponding to some feasible directions with a large degree of abnormal differentiation from the transformed dimensions as retained dimensions includes:

[0027] Arrange the abnormality discrimination degrees of all feasible directions in the direction combination with the largest abnormality discrimination degree difference in descending order to obtain an abnormality discrimination degree sequence; start from the first data in the abnormality discrimination degree sequence, calculate the cumulative sum backward in sequence, and take the abnormality discrimination degree whose cumulative sum is greater than the preset degree threshold for the first time as the demarcation abnormality discrimination degree;

[0028] The transformed dimensions corresponding to the demarcation anomaly discrimination degree and all previous anomaly discrimination degrees are taken as the retained dimensions.

[0029] The present invention increases the difference in the degree of abnormality distinction in different dimensions, thereby making the distance between data points with a larger degree of abnormality and data points with a smaller degree of abnormality larger in some dimensions, thereby improving the accuracy of abnormality detection.

[0030] Preferably, constructing an isolation forest based on the retained dimension using all process parameter data points includes:

[0031] The reserved dimension is used as the dimension selected for constructing the isolation forest, and all process parameter data points are used to construct the isolation forest.

[0032] Preferably, the auxiliary process parameter abnormality monitoring includes:

[0033] The newly collected process parameter data points are input into the isolation forest to obtain the anomaly detection results of the newly collected process parameter data points. If the anomaly detection results of a preset first number of newly collected process parameter data points appear to be abnormal continuously, an early warning is issued.

[0034] In a second aspect, the present invention provides a processing parameter monitoring system for aluminum plate cutting, which adopts the following technical solution:

[0035] A processing parameter monitoring system for aluminum plate cutting includes: a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned processing parameter monitoring method for aluminum plate cutting is implemented.

[0036] By adopting the above technical solution, the above-mentioned method for monitoring processing parameters for aluminum plate cutting is generated into a computer program and stored in a memory so as to be loaded and executed by a processor, thereby making a terminal device based on the memory and the processor for easy use.

[0037] The present invention has the following technical effects:

[0038] The present invention uses dimensional transformation to make the distance difference between process parameter data points with a larger degree of abnormality and process parameter data points with a smaller degree of abnormality in some dimensions larger, so that some dimensions are used as anomaly detection analysis of the isolation forest algorithm to improve the accuracy of anomaly detection;

[0039] Furthermore, in the process of dimensional transformation, by constructing different dimensional transformation methods, the distance difference of the adjusted dimension obtained under each dimensional transformation method is analyzed to accurately select the best dimensional transformation method;

[0040] Furthermore, when screening the best dimensional transformation method, accurate evaluation results are obtained by analyzing the distance differences of process parameter data points with different abnormal degrees in the adjusted dimension under each dimensional transformation method, providing a basis for accurately screening the best dimensional transformation method. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] By reading the detailed description below with reference to the accompanying drawings, the above and other purposes, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-restrictive manner, and the same or corresponding reference numerals represent the same or corresponding parts.

[0042] Figure 1 It is a flow chart of a method in a method for monitoring processing parameters for aluminum plate cutting according to an embodiment of the present invention;

[0043] Figure 2 A schematic diagram of the distribution of process parameter data points provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0044] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0045] It should be understood that when the terms "first", "second", etc. are used in the claims, descriptions, and drawings of the present invention, they are only used to distinguish different objects, rather than to describe a specific order. The terms "include" and "comprise" used in the description and claims of the present invention indicate the presence of the described features, wholes, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components, and / or their collections.

[0046] The embodiment of the present invention discloses a method for monitoring processing parameters for aluminum plate cutting, referring to Figure 1 , comprising steps S1 to S3:

[0047] S1: Obtain several process parameter data points of aluminum plate cutting.

[0048] Specifically, each type of process parameter of the aluminum plate cutting machine at each moment in the cutting process is obtained, and data points consisting of all types of process parameters at each moment are recorded as process parameter data points.

[0049] The types of process parameters include but are not limited to the following: laser power, cutting speed, focus deviation, gas pressure, temperature, and pulse frequency.

[0050] S2: Obtain several feasible directions, combine a preset number of mutually perpendicular feasible directions together to obtain several direction combinations, project the process parameter data points to each feasible direction in the direction combination, and calculate the abnormality distinction degree of any feasible direction; perform dimensional transformation on the process parameter data points according to each feasible direction in the direction combination with the largest difference in abnormality distinction degree.

[0051] It should be noted that since the isolation forest algorithm performs segmentation analysis on each dimension separately to achieve anomaly detection, in other words, as long as the distance difference between abnormal data and normal data in some dimensions is large, the abnormal data can be segmented faster and more accurately.

[0052] It needs to be further explained that if Figure 2 As shown in the schematic diagram of the distribution of process parameter data points, the distance difference between abnormal data points and normal data points can be reflected in most dimensions in the current dimensional space, but no certain dimensions can obviously highlight the distance difference between abnormal data points and normal data points. In order to increase the distance difference between abnormal data and normal data in certain dimensions, it is necessary to transform the dimensions of the process parameter data points so that the distance difference between abnormal data and normal data in certain dimensions is larger. When the isolation forest algorithm is used for anomaly detection, the dimension with a larger distance difference can be selected as the dimension used for anomaly detection, thereby improving the accuracy of anomaly detection.

[0053] S20: Acquire a plurality of feasible directions, combine a preset number of mutually perpendicular feasible directions to obtain a plurality of direction combinations, and project the process parameter data points to each feasible direction in the direction combination.

[0054] Preferably, as an example, a plurality of feasible directions are obtained, a preset number of mutually perpendicular feasible directions are combined together to obtain a plurality of direction combinations, and the process parameter data points are projected onto each feasible direction in the direction combination, including:

[0055] The geometric center of the process parameter data point is obtained, and the surrounding is evenly divided into a preset number of directions with the geometric center as the center, and recorded as feasible directions. The preset number of the present invention is 360 as an example for description, and other embodiments may take other values, which are not specifically limited in this embodiment.

[0056] A preset number of dimensions of process parameter data points is set, and a preset number of mutually perpendicular feasible directions are randomly selected and combined together to obtain a plurality of direction combinations.

[0057] Project the process parameter data points to each feasible direction in the direction combination to obtain the corresponding projection points.

[0058] It is understandable that currently it is unknown which transformation method can increase the distance between abnormal data points and normal data points, so by constructing multiple transformation methods and analyzing the distance between each transformation method, the best transformation method is obtained. The direction combination is a direction combination corresponding to a transformation method.

[0059] S21: Calculate the degree of abnormal differentiation in any feasible direction.

[0060] It should be noted that in order to obtain the best transformation method, the interval distance between data under each transformation method needs to be evaluated.

[0061] Preferably, as an example, calculating the degree of abnormal differentiation in any feasible direction includes:

[0062]

[0063] The abnormality degree of each process parameter data point is obtained, and the process parameter data points are clustered into two categories according to the abnormality degree. The category with a large abnormality degree is recorded as the first category, and the category with a small abnormality degree is recorded as the second category. represents the distance between the projection point of the i-th process parameter data point in the first category and the adjacent projection point in the feasible direction, It represents the distance between the projection point of the i-th process parameter data point in the second category and the adjacent projection point in the feasible direction. represents the abnormality of the i-th process parameter data point in the first category, represents the abnormality of the i-th process parameter data point in the second category, represents the number of process parameter data points in the first category, Represents the number of process parameter data points in the second category.

[0064] Understandably, the degree of abnormal As a weight to distance The weighted sum can reflect the distance between the process parameter data points with a larger degree of abnormality and the surrounding process parameter data points in the feasible direction. As a weight to distance The weighted sum can reflect the distance between the process parameter data points with less abnormality and the surrounding process parameter data points in the feasible direction. It reflects the difference in distance between process parameter data points with a large degree of abnormality and those with a small degree of abnormality. The larger the value is, the greater the difference in distance between process parameter data points with a large degree of abnormality and those with a small degree of abnormality in the feasible direction. Therefore, transformation in this feasible direction can better increase the distance between process parameter data points with different degrees of abnormality.

[0065] The above embodiments involve the degree of abnormality and the distance between the projection point of the process parameter data point in the feasible direction and the adjacent projection point. The following describes a method for obtaining the degree of abnormality and the distance between the projection point of the process parameter data point in the feasible direction and the adjacent projection point.

[0066] Preferably, as an example, the method for calculating the degree of abnormality includes:

[0067] Arrange the values ​​of all process parameter data points in any dimension in time series to obtain a value sequence of the dimension, take the value of any process parameter data point in the dimension as the center, and obtain a data segment of a preset length in the value sequence as a reference segment of the process parameter data point in the dimension;

[0068] Calculate the correlation of the value sequence of each two dimensions, and record the average of the correlations of all the value sequences of the two dimensions as the overall correlation; calculate the correlation of the process parameter data point in the reference segment of the two dimensions, and record the average of the correlations of the process parameter data point in all the reference segments of the two dimensions as the local correlation of the process parameter data point;

[0069] The absolute value of the difference between the value of the process parameter data point in each dimension and the fitting value is calculated, and the average of the absolute values ​​of the difference between the value of the process parameter data point in all dimensions and the fitting value is multiplied by the difference between the overall correlation and the local correlation to obtain the abnormality degree of the process parameter data point.

[0070] It is understandable that the variation law of the process parameter data points in time series when the cutting machine is operating abnormally is different from the variation law when it is operating normally, and most cutting machines are operating normally, so the fitting law of all data points in time series mainly reflects the variation law during normal operation, so the abnormality of the process parameter data points can be reflected by the difference between the values ​​of each dimension of the process parameter data points and the fitting value; when the cutting machine is operating normally, there will be a certain correlation between the different dimensions of the process parameter data points, and when the cutting machine is operating abnormally, the correlation between the different dimensions of the process parameter data points will become smaller, so the abnormality of the process parameter data points can be reflected by the difference between the correlation of the local area of ​​each process parameter data point and the overall correlation. For example, when the cutting machine is operating normally, the laser power and temperature have a certain correlation, that is, the greater the laser power, the greater the temperature. When the cutting machine is operating abnormally, a large amount of cutting debris is generated, and these debris melt under the laser, causing the temperature at the incision position to rise, thereby causing the correlation between power and temperature to deteriorate.

[0071] It should be added that the methods for obtaining fitting values ​​include:

[0072] The values ​​of all process parameter data points in each dimension are arranged in time series to obtain the value sequence of each dimension. The value sequence of each dimension is fitted with a polynomial using the least squares method to obtain the fitting relationship of each dimension. The fitting relationship of each dimension is used to fit the fitting value of each process parameter data point in that dimension.

[0073] Preferably, as an example, a method for obtaining the distance between a projection point of a process parameter data point in the feasible direction and an adjacent projection point includes:

[0074] Calculate the distance between the projection point of the process parameter data point in the feasible direction and the adjacent projection points on each side, and take the maximum value of the distance between the projection point of the process parameter data point in the feasible direction and the connected projection points on both sides as the distance between the projection point of the process parameter data point in the feasible direction and the adjacent projection points.

[0075] S22: Dimensional transformation of the process parameter data points is performed according to each feasible direction in the direction combination with the largest difference in abnormality differentiation degree.

[0076] It should be noted that in order to highlight the distance difference between abnormal data points and normal data points in some dimensions, the distance difference between abnormal data points and normal data points in some dimensions needs to be larger, and the distance difference in other dimensions needs to be smaller.

[0077] Preferably, as an example, the dimension transformation of the process parameter data points is performed according to each feasible direction in the direction combination with the largest difference in the degree of abnormal differentiation, including:

[0078] Calculate the variance of the abnormality discrimination degree of all feasible directions in the direction combination, record it as the abnormality discrimination degree difference of the direction combination, and take each feasible direction in the direction combination with the largest abnormality discrimination degree difference as the adjusted direction;

[0079] The process parameter data points are projected to each adjusted direction, and the dimensions formed by the projection values ​​of all the process parameter data points in each adjusted direction are used as the transformed dimensions.

[0080] S3: Select the transformed dimensions corresponding to some feasible directions with a large degree of abnormality distinction from the transformed dimensions as the retained dimensions; construct an isolation forest based on the retained dimensions using all process parameter data points to assist in monitoring process parameter anomalies.

[0081] S30: Selecting transformed dimensions corresponding to some feasible directions with a large degree of abnormal differentiation from the transformed dimensions as retained dimensions.

[0082] It should be noted that in order to improve the anomaly detection accuracy of the isolation forest algorithm, it is necessary to select dimensions with large distance differences between data points with larger anomalies and data points with smaller anomalies for anomaly detection analysis.

[0083] Preferably, as an example, the transformed dimensions corresponding to some feasible directions with a large degree of abnormal distinction are selected from the transformed dimensions as the retained dimensions, including:

[0084] Arrange the abnormality discrimination degrees of all feasible directions in the direction combination with the largest abnormality discrimination degree difference in descending order to obtain an abnormality discrimination degree sequence; start from the first data in the abnormality discrimination degree sequence, calculate the cumulative sum backward in sequence, and take the abnormality discrimination degree whose cumulative sum is greater than the preset degree threshold for the first time as the demarcation abnormality discrimination degree;

[0085] The transformed dimensions corresponding to the demarcation anomaly discrimination degree and all previous anomaly discrimination degrees are taken as the retained dimensions.

[0086] S31: Based on the retained dimensions, an isolation forest is constructed using all process parameter data points to assist in monitoring process parameter anomalies.

[0087] Preferably, as an example, an isolation forest is constructed based on the reserved dimension using all process parameter data points to assist in monitoring abnormal process parameters, including:

[0088] The reserved dimension is used as the dimension selected for constructing the isolation forest, and all process parameter data points are used to construct the isolation forest.

[0089] The newly collected process parameter data points are input into the isolation forest to obtain the abnormal detection results of the newly collected process parameter data points. If the abnormal detection results of the preset first number of newly collected process parameter data points are abnormal, an early warning is issued to allow relevant personnel to conduct cutting abnormality investigation. This embodiment is described by taking 3 as the preset first number as an example, and other embodiments may take other values, which are not specifically limited in this embodiment.

[0090] It is understandable that according to statistics, if there are three or more process parameter data points in succession, the probability of abnormal operation of the cutting machine is greater than 85%. Therefore, an early warning can be issued at this time to remind relevant personnel to conduct abnormality inspection.

[0091] An embodiment of the present invention also discloses a processing parameter monitoring system for aluminum plate cutting, including a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a processing parameter monitoring method for aluminum plate cutting according to the present invention is implemented.

[0092] The above system also includes other components well known to those skilled in the art, such as a communication bus and a communication interface, and their configuration and functions are known in the art, so they will not be described in detail here.

[0093] In the present invention, the aforementioned memory may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium may be any suitable magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory, a dynamic random access memory, a static random access memory, an enhanced dynamic random access memory, a high bandwidth memory, a hybrid storage cube, etc., or any other medium that can be used to store the required information and can be accessed by an application, a module, or both. Any such computer storage medium may be part of the device or accessible or connectable to the device.

[0094] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

[0095] The above are all preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Therefore, any equivalent changes made based on the structure, shape, and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for monitoring processing parameters for aluminum plate cutting, characterized in that: Includes steps: Obtain several process parameter data points of aluminum plate cutting; Obtaining several feasible directions, including: obtaining the geometric center of the process parameter data point, taking the geometric center as the center, evenly dividing the surrounding area into a preset number of directions and recording them as feasible directions; Combine a preset number of mutually perpendicular feasible directions to obtain several direction combinations, project the process parameter data points to each feasible direction in the direction combination, and calculate the degree of abnormal differentiation of any feasible direction , obtain the abnormality degree of each process parameter data point, and cluster the process parameter data points into two categories according to the abnormality degree. The category with a large abnormality degree is recorded as the first category, and the category with a small abnormality degree is recorded as the second category. represents the distance between the projection point of the i-th process parameter data point in the first category and the adjacent projection point in the feasible direction, It represents the distance between the projection point of the i-th process parameter data point in the second category and the adjacent projection point in the feasible direction. Represent the abnormality degree of the i-th process parameter data point in the first category and the second category respectively, , denotes the number of process parameter data points in the first and second categories, respectively; The process parameter data points are transformed in dimension according to each feasible direction in the direction combination with the largest difference in abnormality discrimination degree, including: calculating the variance of the abnormality discrimination degree of all feasible directions in the direction combination, recording it as the abnormality discrimination degree difference of the direction combination, and taking each feasible direction in the direction combination with the largest difference in abnormality discrimination degree as the adjusted direction; projecting the process parameter data points to each adjusted direction, and taking the dimension formed by the projection values ​​of all the process parameter data points in each adjusted direction as the transformed dimension; The transformed dimensions corresponding to some feasible directions with a large degree of abnormality distinction are selected from the transformed dimensions as retained dimensions; an isolation forest is constructed based on the retained dimensions using all process parameter data points to assist in monitoring process parameter anomalies.

2. A method for monitoring processing parameters for aluminum plate cutting according to claim 1, characterized in that: The step of combining a preset number of mutually perpendicular feasible directions to obtain a plurality of direction combinations includes: A preset number of dimensions of process parameter data points is set, and a preset number of mutually perpendicular feasible directions are randomly selected and combined together to obtain a plurality of direction combinations.

3. The method for monitoring processing parameters for aluminum plate cutting according to claim 1, characterized in that: The obtaining of the abnormality degree of each process parameter data point includes: Arrange the values ​​of all process parameter data points in any dimension in time series to obtain a value sequence of the dimension, take the value of any process parameter data point in the dimension as the center, and obtain a data segment of a preset length in the value sequence as a reference segment of the process parameter data point in the dimension; The mean value of the correlation of all two-dimensional value sequences is recorded as the overall correlation; the mean value of the correlation of the process parameter data point in all two-dimensional reference segments is recorded as the local correlation of the process parameter data point; The degree of abnormality of the process parameter data point is obtained by multiplying the mean of the difference between the value of the process parameter data point in all dimensions and the fitted value by the difference between the overall correlation and the local correlation.

4. The method for monitoring processing parameters for aluminum plate cutting according to claim 1, characterized in that: The step of selecting transformed dimensions corresponding to some feasible directions with a large degree of abnormal differentiation from the transformed dimensions as retained dimensions includes: Arrange the abnormality discrimination degrees of all feasible directions in the direction combination with the largest abnormality discrimination degree difference in descending order to obtain an abnormality discrimination degree sequence; start from the first data in the abnormality discrimination degree sequence, calculate the cumulative sum backward in sequence, and take the abnormality discrimination degree whose cumulative sum is greater than the preset degree threshold for the first time as the demarcation abnormality discrimination degree; The transformed dimensions corresponding to the demarcation anomaly discrimination degree and all previous anomaly discrimination degrees are taken as the retained dimensions.

5. The method for monitoring processing parameters for aluminum plate cutting according to claim 1, characterized in that: The isolation forest is constructed based on the reserved dimension using all process parameter data points, including: The reserved dimension is used as the dimension selected for constructing the isolation forest, and all process parameter data points are used to construct the isolation forest.

6. The method for monitoring processing parameters for aluminum plate cutting according to claim 1, characterized in that: The auxiliary process parameter abnormality monitoring includes: The newly collected process parameter data points are input into the isolation forest to obtain the anomaly detection results of the newly collected process parameter data points. If the anomaly detection results of a preset first number of newly collected process parameter data points appear to be abnormal continuously, an early warning is issued.

7. A processing parameter monitoring system for aluminum plate cutting, characterized in that: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a processing parameter monitoring method for aluminum plate cutting according to any one of claims 1 to 6 is implemented.

Images