Blowing flow adjusting system in metal powder machining process

Abstract

The invention relates to the technical field of data processing, in particular to a blowing flow adjusting system in the metal powder machining process. The system comprises the steps of collecting flow data, obtaining a mutation factor of each data in a main body flow data sequence, and obtaining a noise factor of each data in the main body flow data sequence according to the mutation factor of each data in the main body flow data sequence; based on the noise factor, adaptively obtaining a judgment threshold value of a Mahalanobis distance of each datum in the main body flow data sequence; according to the judgment threshold value of the Mahalanobis distance of each data in the main body flow data sequence, noise data is obtained for smoothing, the denoising effect is improved, and the influence of abnormal flow data on noise data recognition is avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and more specifically, to a blowing flow rate regulating system in a metal powder processing process. Background Art

[0002] With the rising energy costs and the improvement of environmental protection awareness, the manufacturing industry is paying more and more attention to how to reduce production costs and resource consumption while ensuring quality. The use of a blowing flow regulation system can effectively control the amount of materials used and the energy consumption during the processing process. The application of this system responds to the needs of energy conservation, emission reduction and resource efficiency optimization. An existing algorithm for intelligently regulating the blowing flow during metal powder processing is the PID control algorithm. This method can effectively improve processing quality, production efficiency and system reliability through its accuracy, rapid response, stability and adaptability. It is one of the common control strategies in modern manufacturing. However, when collecting flow data during metal processing, the processing environment may have vibrations, temperature changes or other environmental factors, which may affect the accuracy of the flow sensor and introduce noise. The presence of noise data will reduce the accuracy of the PID control algorithm, making it impossible to achieve accurate flow control.

[0003] The patent application document currently published with the publication number CN114117334A discloses a method for anomaly detection in multidimensional sensor data, by receiving a time series of vectors from a sensor; determining a weighted moving mean of the vectors; determining an inverse covariance matrix of the vectors; receiving a current vector from the sensor; determining a squared Mahalanobis distance between the current vector and the weighted moving mean; and outputting an indicator of the sensor anomaly in response to the squared Mahalanobis distance exceeding a threshold; and determining the squared Mahalanobis distance by using the inverse covariance matrix.

[0004] It is known that when the Mahalanobis distance algorithm in the above method cleans the noise data in the flow data, it is necessary to obtain the Mahalanobis distance of each flow data, and finally the distance threshold is used to determine the Mahalanobis distance of each flow data to obtain the noise data. The Mahalanobis distance of each flow data is obtained based on its numerical performance, and the numerical value of the noise data in the flow data is similar to the numerical value of the abnormal flow data. Therefore, the abnormal flow data in the metal powder processing process will affect the detection of the noise data, resulting in inaccurate noise data and reducing the denoising effect. Summary of the invention

[0005] In order to solve the problem that the value of noise data in flow data is similar to the value of abnormal flow data, so the abnormal flow data in the metal powder processing process will affect the detection of noise points, resulting in inaccurate noise data and reducing the denoising effect, the present invention proposes a blowing flow regulation system in the metal powder processing process. The system includes the following modules: A data collection module, used to collect a main flow data sequence and a reference flow data sequence; The noise factor acquisition module is used to obtain the mutation factor of each data in the main flow data sequence, and the mutation factor represents the prominence of the data value; obtain the noise factor of each data in the main flow data sequence ; In the formula, Represents the noise factor of the nth data in the main flow data sequence; Represents the mutation factor of the nth data in the main flow data sequence; Represents the number of reference traffic data series; Represents the value of the nth data in the main flow data sequence; Represents the value of the corresponding data of the nth data in the subject flow data sequence in the rth reference flow data sequence; Represents the numerical reliability of the corresponding data of the nth data in the subject flow data sequence in the rth reference flow data sequence; A determination threshold acquisition module is used to obtain a determination threshold of the Mahalanobis distance of each data in the main flow data sequence, wherein the determination threshold is inversely proportional to the noise factor of each data; The noise data smoothing module is used to obtain noise data according to the determination threshold of the Mahalanobis distance of each data in the main flow data sequence; and to smooth the noise data.

[0006] The innovation of the present invention lies in analyzing the changing characteristics of flow data at multiple positions, obtaining the noise factor of each flow data at the position being analyzed, that is, the noise factor of each data in the main flow data sequence, and finally correcting the judgment threshold of the initial Mahalanobis distance of each data in the main flow data sequence based on the noise factor to obtain the judgment threshold of the Mahalanobis distance of each data in the main flow data sequence, thereby accurately screening out noise data points, avoiding the detection of noise data by abnormal flow data in the metal powder processing process, making the acquired noise data more accurate, improving the denoising effect, providing more accurate flow data for the subsequent application of the PID control algorithm, and making the PID algorithm more accurate in regulating the flow data.

[0007] Preferably, the step of obtaining the mutation factor of each data in the main flow data sequence includes: ; In the formula, Represents the mutation factor of the nth data in the main flow data sequence; Represents the value of the nth data in the main flow data sequence; represents the value of the kth data in the neighborhood data segment of the nth data in the main flow data sequence, excluding the nth data; K represents the number of data in the neighborhood data segment of the nth data in the main flow data sequence, excluding the nth data; Represents the absolute value of the difference between the nth data and the previous data in the main flow data sequence; Represents the mean of the absolute values ​​of the numerical differences of all adjacent data in the neighborhood data segment of the nth data in the main flow data sequence; || represents the absolute value symbol; norm() represents the normalization function.

[0008] The larger the mutation factor, the more likely the data is noise data, which makes it easier to obtain the noise factor of each data in the main traffic data sequence based on the mutation factor indicator.

[0009] Preferably, the acquisition of the neighborhood data segment includes: The number of data N is preset, and the first N data and the last N data of the nth data in the main flow data sequence are obtained to form a neighborhood data segment of the nth data in the main flow data sequence.

[0010] Preferably, the acquisition of the numerical reliability includes: ; In the formula, Represents the numerical reliability of the corresponding data of the nth data in the subject flow data sequence in the rth reference flow data sequence; Represents the number of reference traffic data series; The Pearson correlation coefficient between the neighborhood data segment of the nth data in the main flow data sequence and the neighborhood data segment of the corresponding data of the nth data in the main flow data sequence in the rth reference flow data sequence; Represents the standard deviation of the data in the neighborhood data segment of the corresponding data of the nth data in the main traffic data sequence in the rth reference traffic data sequence; exp() represents an exponential function with a natural constant as the base.

[0011] The acquisition of the numerical reliability of the corresponding data of each data in the main flow data sequence in each reference flow data sequence makes the noise factor of each data in the main flow data sequence more accurate.

[0012] Preferably, the determination threshold of the Mahalanobis distance of each data in the main flow data sequence is obtained, including: ; In the formula, represents the decision threshold of the Mahalanobis distance of the nth data in the main flow data sequence; represents the noise factor of the nth data in the main flow data sequence; B represents the judgment threshold of the initial Mahalanobis distance of each data in the main flow data sequence; exp() represents an exponential function with a natural constant as the base.

[0013] The initial Mahalanobis distance determination threshold of each data is modified according to the noise factor of each data, so that the Mahalanobis distance determination threshold of each data is more accurate, which is convenient for accurately detecting the noise data in the subsequent process.

[0014] Preferably, the step of obtaining noise data according to a determination threshold of the Mahalanobis distance of each data in the main flow data sequence includes: According to the Mahalanobis distance algorithm, the Mahalanobis distance of each data in the main flow data sequence is obtained; if the Mahalanobis distance of the nth data in the main flow data sequence is greater than or equal to the judgment threshold of the Mahalanobis distance of the nth data in the main flow data sequence, the nth data in the main flow data sequence is noise data, and all noise data in the main flow data sequence are obtained.

[0015] It can accurately filter out noise data points and avoid the detection of noise data by abnormal flow data during metal powder processing.

[0016] Preferably, the smoothing of the noise data comprises: Obtain the average value of the 10 consecutive non-noise data before the f-th noise data in the main flow data sequence as the updated value of the f-th noise data in the main flow data sequence, and obtain the updated value of each noise data in the main flow data sequence.

[0017] Preferably, the collected main flow data sequence and reference flow data sequence include: A flow sensor needs to be placed in the blowing duct, which is recorded as the main flow sensor, and two flow sensors are placed in front and behind the main flow sensor, which are reference flow sensors. The distance between two adjacent sensors is three centimeters. Every 2 seconds is a sampling moment, and the flow data of the main flow sensor and each reference flow sensor are collected each time. The collection is done for two hours in total. The flow data collected by the main flow sensor are sorted in ascending order according to the sampling time to obtain the main flow data sequence; the flow data collected by each reference flow sensor are sorted in ascending order according to the sampling time to obtain each reference flow data sequence.

[0018] The present invention has the following technical effects: the present invention analyzes the change characteristics of the flow data in the reference flow data sequence and the main flow data sequence, obtains the noise factor of each flow data in the main flow data sequence, and finally corrects the judgment threshold of the initial Mahalanobis distance of each data in the main flow data sequence based on the noise factor to obtain the judgment threshold of the Mahalanobis distance of each data in the main flow data sequence, thereby accurately screening out noise data points, avoiding the detection of noise data by abnormal flow data in the metal powder processing process, making the acquired noise data more accurate, improving the denoising effect, providing more accurate flow data for the subsequent application of the PID control algorithm, and making the PID algorithm more accurate in regulating the flow data. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, 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-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein: Figure 1 The system block diagram of a blowing flow rate regulating system in a metal powder processing process according to an embodiment of the present invention. DETAILED DESCRIPTION

[0020] 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.

[0021] 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.

[0022] The present invention provides a blowing flow rate regulating system in a metal powder processing process. Figure 1 As shown, a blowing flow rate regulating system in a metal powder processing process includes modules 101 to 104, which are described in detail below.

[0023] The data collection module 101 collects flow data.

[0024] In an embodiment of the present invention, a flow sensor needs to be placed in the blowing duct, which is recorded as the main flow sensor, and two flow sensors are placed in front and behind the main flow sensor, each with an area of ​​​​one reference flow sensor, and the distance between two adjacent sensors is three centimeters. Every 2 seconds is a sampling moment, and the flow data of the main flow sensor and each reference flow sensor are collected each time, and the collection is done for a total of two hours. The flow data collected by the main flow sensor is sorted in ascending order according to the sampling time to obtain a main flow data sequence; the flow data collected by each reference flow sensor is sorted in ascending order according to the sampling time to obtain each reference flow data sequence. It should be noted that the data timing in the main flow data sequence and each reference flow data sequence is aligned.

[0025] The noise factor acquisition module 102 acquires the mutation factor of each data in the main flow data sequence, and acquires the noise factor of each data in the main flow data sequence according to the mutation factor of each data in the main flow data sequence.

[0026] It should be noted that an existing algorithm for intelligently regulating flow data in the metal powder processing process is a PID control algorithm. This method can effectively improve processing quality, production efficiency and system reliability through its accuracy, rapid response, stability and adaptability. It is one of the common control strategies in modern manufacturing. However, when collecting flow data in the metal processing process, the processing environment may have vibrations, temperature changes or other environmental factors, which may affect the accuracy of the flow sensor, thereby introducing noise. The presence of noise data will reduce the accuracy of the PID control algorithm, making it impossible to achieve accurate flow control. Therefore, before using the PID control algorithm to regulate the flow data, the present invention will use the Mahalanobis distance algorithm to clean the flow data to obtain smoother flow data that conforms to the actual situation. It is known that the Mahalanobis distance of each flow data is obtained based on its numerical performance, and the numerical value of the noise data in the flow data is similar to the numerical value of the abnormal flow data. Therefore, when using the Mahalanobis distance algorithm for data cleaning, the abnormal flow data in the metal powder processing process will affect the detection of the noise points. Therefore, it is necessary to distinguish the abnormal flow data and the noise data in the main flow data sequence.

[0027] It should be noted that it is necessary to analyze the numerical characteristics of the data in the main flow data sequence. The more prominent the numerical performance of any data in the main flow data sequence is, the larger the mutation factor of the data is. Therefore, in the present invention, the neighborhood data segment of any data in the main flow data sequence is obtained, and the mutation factor of the data in the main flow data sequence is obtained according to the numerical difference between the data in the main flow data sequence and other data in its neighborhood data segment. When the numerical difference between the data in the main flow data sequence and other data in its neighborhood data segment is large, it means that the mutation factor of the data in the main flow data sequence is large.

[0028] In the embodiment of the present invention, the neighborhood data segment of each data in the main flow data sequence is obtained: the first N data and the last N data of the nth data in the main flow data sequence are obtained to form the neighborhood data segment of the nth data in the main flow data sequence; in the embodiment of the present invention, the preset number of data N=40, in other embodiments, the implementer can preset the value of N according to the specific implementation situation. It should be noted that if there is a data in the front side or the back side of the main flow data sequence that does not meet N, a 0-filling operation is performed.

[0029] Get the mutation factor of each data in the main flow data sequence: ; In the formula, Represents the mutation factor of the nth data in the main flow data sequence; Represents the value of the nth data in the main flow data sequence; represents the value of the kth data in the neighborhood data segment of the nth data in the main flow data sequence, excluding the nth data; K represents the number of data in the neighborhood data segment of the nth data in the main flow data sequence, excluding the nth data; Represents the absolute value of the difference between the nth data and the previous data in the main flow data sequence; Represents the mean of the absolute values ​​of the numerical differences of all adjacent data in the neighborhood data segment of the nth data in the main flow data sequence; It represents the sum of the differences between the value of the nth data in the main flow data sequence and the corresponding values ​​of all data in the neighboring data segment of the data except itself. The larger the value, the more outstanding the value of the nth data in the main flow data sequence in its neighboring data segment, and the larger its mutation factor will be. It represents the difference between the absolute value of the numerical difference between the nth data in the main flow data sequence and its previous data and the average of the absolute values ​​of the numerical differences of all adjacent data in the neighborhood data segment of the data. The larger the value is, the more credible it is that the numerical performance of the nth data in the main flow data sequence in its neighborhood data segment is more outstanding, which also means that the mutation factor of the data is larger.

[0030] It should be noted that the mutation factor of each data in the known main flow data sequence is obtained based on the analysis of the numerical difference between each data in the main flow data sequence and each data in its neighborhood data segment. However, since the abnormal flow data and noise data generated in the metal powder processing process have similar numerical performances, the two types of data cannot be well distinguished based on the mutation factor alone. Therefore, the present invention analyzes the consistency of data changes between each reference flow data sequence and the main flow data sequence, and optimizes the mutation factor of each data in the main flow data sequence to obtain the noise factor of each data in the main flow data sequence. If the consistency (numerical difference) of changes between any data in the main flow data sequence and the corresponding data of the data in multiple reference flow data sequences is lower, it means that the data is more likely to be affected by noise, and its noise factor is larger. In an embodiment of the present invention, the sampling time of the nth data in the main flow data sequence is recorded as the nth representative sampling time, and the data at the nth representative sampling time in the rth reference flow data sequence is recorded as the corresponding data of the nth data in the main flow data sequence in the rth reference flow data sequence.

[0031] According to the method for obtaining the neighborhood data segment of each data in the main flow data sequence, the neighborhood data segment of each data in each reference flow data sequence is obtained.

[0032] Get the noise factor for each data in the main flow data series: ; ; In the formula, Represents the noise factor of the nth data in the main flow data sequence; Represents the mutation factor of the nth data in the main flow data sequence; Represents the number of reference traffic data series; Represents the value of the nth data in the main flow data sequence; Represents the value of the corresponding data of the nth data in the subject flow data sequence in the rth reference flow data sequence; The Pearson correlation coefficient between the neighborhood data segment of the nth data in the main flow data sequence and the neighborhood data segment of the corresponding data of the nth data in the main flow data sequence in the rth reference flow data sequence; Represents the standard deviation of the data in the neighborhood data segment of the corresponding data of the nth data in the main flow data sequence in the rth reference flow data sequence; || represents the absolute value symbol; exp() represents the exponential function with a natural constant as the base; The larger the value of is, the more drastic the change in the nth data in the main flow data sequence is, and the greater the possibility that it is noise, so the noise factor is also larger; The sum of the absolute values ​​of the differences between the value of the nth data in the main flow data sequence and the value of the corresponding data of the data in each reference flow data sequence. The larger the value, the lower the consistency of the change between the nth data in the main flow data sequence and the corresponding data of the data in multiple reference flow data sequences, indicating that the nth data is more likely to be affected by noise and its noise factor is larger; It represents the numerical reliability of the corresponding data of the nth data in the main flow data sequence in the rth reference flow data sequence. The larger the value, the better. The greater the credibility of the result; The larger the value of is, the larger the value is, indicating that the consistency of the data change in the neighborhood data segment between the nth data of the main flow data sequence and the corresponding data of the data in multiple reference flow data sequences is higher, which can explain that the reference of the numerical value of the corresponding data of the nth data in the main flow data sequence in multiple reference flow data sequences is higher; The smaller the value is, the more stable the data in the neighborhood data segment of the corresponding data of the nth data in the main traffic data sequence in each reference traffic data sequence is, which means that the reference value of the corresponding data of the nth data in the main traffic data sequence in multiple reference traffic data sequences is higher.

[0033] The determination threshold acquisition module 103 adaptively acquires the determination threshold of the Mahalanobis distance of each data in the main flow data sequence according to the noise factor of each data in the main flow data sequence.

[0034] It should be noted that it is known that when the Mahalanobis distance is used for noise analysis, when the Mahalanobis distance of any data is greater than or equal to the judgment threshold of the Mahalanobis distance of the data, the data is considered to be noise data, and when the noise factor corresponding to any data in the main flow data sequence is larger, it means that the possibility that it belongs to noise data is greater, and therefore a smaller Mahalanobis distance judgment threshold should be required to judge the Mahalanobis distance corresponding to the data as noise data; when the noise factor corresponding to any data in the main flow data sequence is smaller, it means that the possibility that it belongs to abnormal flow data or normal data is greater, and then a larger Mahalanobis distance judgment threshold should be required to judge the Mahalanobis distance corresponding to the data as non-noise data, and therefore the present invention needs to adaptively obtain the Mahalanobis distance judgment threshold of each data in the main flow data sequence according to the noise factor corresponding to any data in the main flow data sequence.

[0035] In the embodiment of the present invention, the determination threshold of the Mahalanobis distance of each data in the main flow data sequence is obtained: ; In the formula, represents the decision threshold of the Mahalanobis distance of the nth data in the main flow data sequence; represents the noise factor of the nth data in the main flow data sequence; B represents the judgment threshold of the initial Mahalanobis distance of each data in the main flow data sequence; exp() represents an exponential function with a natural constant as the base; in the embodiment of the present invention, B=0.9 is preset, and in other embodiments, the implementer can preset the value of B according to the specific implementation situation.

[0036] The noise data smoothing module 104 obtains noise data for smoothing according to a determination threshold of the Mahalanobis distance of each data in the main flow data sequence.

[0037] In an embodiment of the present invention, the Mahalanobis distance of each data in the main flow data sequence is obtained according to the Mahalanobis distance algorithm; if the Mahalanobis distance of the nth data in the main flow data sequence is greater than or equal to the determination threshold of the Mahalanobis distance of the nth data in the main flow data sequence, it indicates that the nth data in the main flow data sequence is noise data, and all the noise data in the main flow data sequence are obtained. The average value of the 10 consecutive non-noise data before the f-th noise data in the main flow data sequence is obtained as the updated value of the f-th noise data in the main flow data sequence, and the updated value of each noise data in the main flow data sequence is obtained to obtain the updated main flow data sequence. Finally, the PID control algorithm is used to intelligently control the data in the main flow data sequence. It should be noted that the PID control algorithm is a well-known technology and will not be described in detail in the embodiments of the present invention.

[0038] The present invention analyzes the change characteristics of the flow data in the reference flow data sequence and the main flow data sequence, obtains the noise factor of each flow data in the main flow data sequence, and finally corrects the judgment threshold of the initial Mahalanobis distance of each data in the main flow data sequence based on the noise factor to obtain the judgment threshold of the Mahalanobis distance of each data in the main flow data sequence, thereby accurately screening out the noise data points, avoiding the detection of the noise data by the abnormal flow data in the metal powder processing process, making the acquired noise data more accurate, improving the denoising effect, providing more accurate flow data for the subsequent application of the PID control algorithm, and making the PID algorithm more accurate in regulating the flow data.

[0039] The above description is only a preferred embodiment of the present invention, and is not intended to limit the protection scope of the present invention. Therefore, any equivalent changes made according to the structure, shape, and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A blowing flow rate regulating system in a metal powder processing process, characterized in that: Includes the following modules: A data collection module, used to collect a main flow data sequence and a reference flow data sequence; A noise factor acquisition module is used to obtain a mutation factor of each data in the main flow data sequence, wherein the mutation factor represents the prominence of the data value; Get the noise factor of each data in the main flow data series ; In the formula, Represents the noise factor of the nth data in the main flow data sequence; Represents the mutation factor of the nth data in the main flow data sequence; Represents the number of reference traffic data series; Represents the value of the nth data in the main flow data sequence; Represents the value of the corresponding data of the nth data in the subject flow data sequence in the rth reference flow data sequence; Represents the numerical reliability of the corresponding data of the nth data in the subject flow data sequence in the rth reference flow data sequence; A determination threshold acquisition module is used to obtain a determination threshold of the Mahalanobis distance of each data in the main flow data sequence, wherein the determination threshold is inversely proportional to the noise factor of each data; A smooth noise data module is used to obtain noise data according to a determination threshold of the Mahalanobis distance of each data in the main flow data sequence; Smooth the noisy data.

2. The air flow regulating system in the metal powder processing process according to claim 1, characterized in that: The step of obtaining the mutation factor of each data in the main flow data sequence includes: ； In the formula, Represents the mutation factor of the nth data in the main flow data sequence; Represents the value of the nth data in the main flow data sequence; represents the value of the kth data in the neighborhood data segment of the nth data in the main flow data sequence, excluding the nth data; K represents the number of data in the neighborhood data segment of the nth data in the main flow data sequence, excluding the nth data; Represents the absolute value of the difference between the nth data and the previous data in the main flow data sequence; Represents the mean of the absolute values ​​of the numerical differences of all adjacent data in the neighborhood data segment of the nth data in the main flow data sequence; || represents the absolute value symbol; norm() represents the normalization function.

3. The air flow regulating system in the metal powder processing process according to claim 2, characterized in that: The acquisition of the neighborhood data segment includes: The number of data N is preset, and the first N data and the last N data of the nth data in the main flow data sequence are obtained to form a neighborhood data segment of the nth data in the main flow data sequence.

4. The air flow regulating system in the metal powder processing process according to claim 1, characterized in that: The acquisition of the numerical reliability includes: ； In the formula, Represents the numerical reliability of the corresponding data of the nth data in the subject flow data sequence in the rth reference flow data sequence; Represents the number of reference traffic data series; The Pearson correlation coefficient between the neighborhood data segment of the nth data in the main flow data sequence and the neighborhood data segment of the corresponding data of the nth data in the main flow data sequence in the rth reference flow data sequence; Represents the standard deviation of the data in the neighborhood data segment of the corresponding data of the nth data in the main traffic data sequence in the rth reference traffic data sequence; exp() represents an exponential function with a natural constant as the base.

5. The air flow regulating system in the metal powder processing process according to claim 1, characterized in that: The determination threshold of the Mahalanobis distance of each data in the main flow data sequence is obtained, including: ； In the formula, represents the decision threshold of the Mahalanobis distance of the nth data in the main flow data sequence; represents the noise factor of the nth data in the main flow data sequence; B represents the judgment threshold of the initial Mahalanobis distance of each data in the main flow data sequence; exp() represents an exponential function with a natural constant as the base.

6. The air flow regulating system in the metal powder processing process according to claim 1, characterized in that: The step of obtaining noise data according to the determination threshold of the Mahalanobis distance of each data in the main flow data sequence includes: According to the Mahalanobis distance algorithm, the Mahalanobis distance of each data in the main flow data sequence is obtained; if the Mahalanobis distance of the nth data in the main flow data sequence is greater than or equal to the judgment threshold of the Mahalanobis distance of the nth data in the main flow data sequence, the nth data in the main flow data sequence is noise data, and all noise data in the main flow data sequence are obtained.

7. The air flow regulating system in the metal powder processing process according to claim 1, characterized in that: The smoothing of the noise data comprises: Obtain the average value of the 10 consecutive non-noise data before the f-th noise data in the main flow data sequence as the updated value of the f-th noise data in the main flow data sequence, and obtain the updated value of each noise data in the main flow data sequence.

8. The air flow regulating system in the metal powder processing process according to claim 1, characterized in that: The collection subject flow data sequence and the reference flow data sequence include: A flow sensor needs to be placed in the blowing duct, which is recorded as the main flow sensor, and two flow sensors are placed in front and behind the main flow sensor, which are reference flow sensors. The distance between two adjacent sensors is three centimeters. Every 2 seconds is a sampling moment, and the flow data of the main flow sensor and each reference flow sensor are collected each time. The collection is done for two hours in total. The flow data collected by the main flow sensor are sorted in ascending order according to the sampling time to obtain the main flow data sequence; the flow data collected by each reference flow sensor are sorted in ascending order according to the sampling time to obtain each reference flow data sequence.

Images