Shot blasting tank safety monitoring method and system based on machine learning

Abstract

The present application relates to the field of safety monitoring technology, and in particular to a shot peening tank safety monitoring method and system based on machine learning. The method includes real-time acquisition of pressure data and spray gun flow data, constructing a real-time pressure sequence and a real-time spray gun flow sequence, determining the cluster to which the real-time pressure sequence belongs based on the difference between the real-time pressure sequence and the historical pressure sequence corresponding to the cluster center of the pressure cluster, adjusting the real-time spray gun flow sequence according to the error distribution state corresponding to the cluster to obtain a real-time shot peening sequence; calculating the similarity between the real-time shot peening sequence and the real-time pressure sequence, and performing a weighted summation of the similarities based on the error distribution state; using the result of the weighted summation of the similarities as the change consistency, and judging the operating status of the tank based on the change consistency. The present application has the effect of improving the timeliness and comprehensiveness of shot peening tank safety monitoring.

Description

Technical Field

[0001] The present application relates to the field of safety monitoring technology, and in particular to a shot blasting tank safety monitoring method and system based on machine learning. Background Art

[0002] Shot peening is a surface strengthening process that bombards the workpiece with high-speed pellets, introducing compressive stress and increasing the overall strength of the material. During shot peening operation, the shot is stored in a shot tank. Compressed air transports the shot to the spray gun, where it is ejected, completing the shot peening of the workpiece surface. Over time, the shot tank may develop safety hazards such as metal fatigue and corrosion damage. Without timely detection and monitoring, these hazards can seriously impact the safe operation of the equipment. Traditional inspection methods include manual visual inspection, which can detect cracks, corrosion, and other issues on shot peening tanks. However, this method suffers from a certain lag in actual monitoring and is prone to missed detections. Therefore, with the advancement of sensing technology, ultrasonic testing is currently the primary method used in the industry. The principle is that when a crack or leak develops in a tank, the leak generates eddy currents, which produce ultrasonic waves. The waveform received by the ultrasonic probe determines whether the tank is leaking.

[0003] Compared to manual visual inspection, ultrasonic flaw detection improves detection accuracy to a certain extent and reduces missed detections. However, ultrasonic flaw detection can only detect cracks after they are present in the shot peening tank, and cannot promptly detect subtle changes in the early stages of crack development. Furthermore, traditional ultrasonic flaw detection still relies on manual monitoring of the shot peening tank using ultrasonic detection devices, which results in a certain lag in the detection process. Therefore, traditional shot peening tank inspection methods are unable to accurately and in real time monitor potential risks in some shot peening tanks, resulting in certain safety risks during the use of shot peening equipment. Summary of the Invention

[0004] In order to solve the problem in related technologies that potential risks of shot blasting tanks cannot be accurately detected in real time, this application provides a shot blasting tank safety monitoring method and system based on machine learning.

[0005] In the first aspect, this application provides a shot blasting tank safety monitoring method based on machine learning, which adopts the following technical solutions:

[0006] A machine learning-based shot blasting tank safety monitoring method collects historical data and divides the historical data into multiple data sequences, including historical pressure sequences, historical spray gun flow sequences, and historical shot blasting sequences. The historical pressure sequences are clustered to construct multiple pressure clusters.

[0007] Based on the historical spray gun flow sequence and the historical shot peening sequence, the flow error sequence corresponding to the historical pressure sequence in the pressure cluster is obtained; based on the number and type of flow error sequences in the pressure cluster, the error distribution state of the pressure cluster is determined;

[0008] Pressure data and spray gun flow data are collected in real time, and real-time pressure sequences and real-time spray gun flow sequences are constructed. The cluster to which the real-time pressure sequence belongs is determined based on the difference between the real-time pressure sequence and the historical pressure sequence corresponding to the cluster center of the pressure cluster. The real-time spray gun flow sequence is adjusted according to the error distribution state corresponding to the cluster to obtain the real-time shot peening sequence. The similarity between the real-time shot peening sequence and the real-time pressure sequence is calculated, and the similarity is weighted and summed based on the error distribution state. The result of the weighted sum of the similarities is used as the change consistency, and the operating status of the tank is judged based on the change consistency.

[0009] The beneficial effect is that during the shot peening process, there is a certain error between the shot peening flow rate and the gun flow rate detected by the sensor at the spray gun. Therefore, in this application, pressure data, gun flow rate data, and shot peening data from the historical operation of the shot peening equipment are collected to form a historical pressure sequence, a historical gun flow rate sequence, and a historical shot peening sequence. The historical data is analyzed to calculate the error between the gun flow rate and the shot peening flow rate under different pressure conditions, forming a flow error sequence. Subsequently, during the real-time shot peening process, pressure data is collected in real time to form a real-time pressure sequence. The relationship between the real-time pressure sequence and the historical pressure sequence is compared to determine the state to which the real-time pressure sequence belongs. The real-time gun flow sequence is adjusted based on the flow error sequence under different pressure conditions calculated from the historical data to obtain a real-time shot peening sequence. The operating state of the shot peening tank is determined based on the consistency of changes between the real-time shot peening sequence and the real-time pressure sequence, completing the inspection of the shot peening tank. Compared to traditional methods in which workers use a detection device to inspect the shot peening tank, the method of this application can detect potential defects that affect the pressure and shot peening flow rate of the shot peening tank, providing a more comprehensive inspection. At the same time, with the increase of historical data and the deepening of machine learning, the detection of shot peening tanks is more accurate; moreover, the real-time data collection and detection in this application improves the timeliness of shot peening tank detection and reduces the safety risks during the shot peening process.

[0010] Optionally, the step of clustering the historical pressure sequences to construct multiple pressure clusters includes: calculating a sequence distance between any two historical pressure sequences, and clustering the historical pressure sequences based on the sequence distance between the two historical pressure sequences.

[0011] The beneficial effect is that the sequence distance represents the similarity between two historical pressure sequences. The more similar the data in the two historical pressure sequences are, the smaller the distance between them is. The historical pressure sequences are clustered according to the sequence distance, so that the historical pressure sequences with similar data in the historical pressure sequences are clustered into a cluster.

[0012] Optionally, use kmeans clustering method to cluster the historical pressure series.

[0013] Optionally, the step of calculating the sequence distance between any two historical pressure sequences includes: taking the square of the difference between the same-order data in the two historical pressure sequences as the local distance, and taking the sum of the local distances corresponding to each data in the historical pressure sequence as the sequence distance between the two historical pressure sequences.

[0014] The beneficial effect is: calculating the difference of each data in the historical pressure series, and accumulating the difference corresponding to each data in the historical pressure data, thereby reflecting the sequence distance between the two historical pressure series.

[0015] Optionally, the step of calculating the sequence distance between any two historical pressure sequences includes: obtaining the mean and variance of the pressure data in the historical pressure sequences; and determining the sequence distance between the two historical pressure sequences by the difference between the means and the difference between the variances of the two historical pressure sequences.

[0016] The beneficial effect is that the difference in the mean and variance of two historical pressure series is compared to reflect the sequence distance between the two historical pressure series. This method focuses more on the overall difference between the two historical pressure series, reduces the impact of individual data on the sequence distance, and improves the robustness of the calculation.

[0017] Optionally, the step of obtaining the flow error sequence includes calculating the flow difference between the data of the same sequence in the historical spray gun flow sequence and the historical shot peening sequence, and a plurality of flow differences constitute the flow error sequence.

[0018] The beneficial effect is that the historical gun flow rate sequence is detected by the sensor, and the historical shot peening sequence is reflected by the number or type of shots actually peened per unit time. The flow rate difference between the historical gun flow rate sequence and the shot peening sequence reflects the error between the sensor-detected flow rate (i.e., the gun flow rate) and the actual shot peening flow rate. Multiple flow rate difference data then constitute a flow rate error sequence.

[0019] Optionally, the step of determining the error distribution state of the pressure clustering cluster based on the number and types of flow error sequences in the pressure clustering cluster includes: clustering the corresponding flow error sequences in the pressure clustering cluster to obtain multiple error clustering clusters; obtaining the mean of the same-order data of multiple flow error sequences in the error clustering cluster to form a representative error sequence; taking the ratio of the number of elements in the error clustering cluster to the total number of corresponding flow error sequences in the pressure clustering group as the probability of the representative error sequence; obtaining the probabilities of multiple representative error sequences to obtain the error distribution state corresponding to the pressure clustering cluster.

[0020] The beneficial effect is that within the same pressure cluster, multiple historical pressure sequences have similar pressure fluctuations, so a pressure cluster represents a specific pressure state. Under a given pressure condition, there are multiple flow error sequences. Therefore, by clustering flow error sequences, the probability of different errors existing under that pressure state can be determined.

[0021] Optionally, the step of adjusting the real-time spray gun flow sequence according to the flow error sequence in the pressure cluster to obtain the real-time shot peening sequence includes: adding the real-time spray gun flow sequence to the data representing the same sequence in the error sequence to obtain the real-time shot peening sequence.

[0022] The beneficial effect is: after determining which pressure state the real-time pressure sequence belongs to according to the real-time pressure sequence, the real-time shot peening flow rate is calculated based on the flow error obtained by historical data analysis to obtain the real-time shot peening sequence.

[0023] Optionally, the step of determining the operating status of the storage tank based on the change consistency includes setting an alarm threshold, and issuing an alarm in response to the change consistency being less than the alarm threshold.

[0024] Secondly, this application provides a shot blasting tank safety monitoring system based on machine learning, which adopts the following technical solutions:

[0025] A shot blasting tank safety monitoring system based on machine learning includes a processor and a memory, wherein the memory stores computer program instructions. When the computer program instructions are executed by the processor, the shot blasting tank safety monitoring method based on machine learning is implemented.

[0026] The beneficial effect is: the above-mentioned shot blasting tank safety monitoring method based on machine learning is generated into a computer program and stored in a memory so as to be loaded and executed by a processor, thereby making a system based on the memory and the processor for easy use.

[0027] This application has the following technical effects:

[0028] In this application, historical data is analyzed based on machine learning, the difference between the detected gun flow and the actual shot peening flow is calculated, and then the possible error between the currently collected gun flow and the actual shot peening flow is determined based on the real-time collected pressure data. The real-time shot peening sequence is obtained by adjusting the real-time collected gun flow sequence. The operating status of the current equipment is judged in real time based on the consistency of changes between the real-time shot peening sequence and the real-time pressure sequence. On the one hand, the timeliness of the inspection of the shot peening tank is improved, and on the other hand, it is convenient to discover potential defects of the shot peening tank other than crack defects, thereby reducing the safety risks during the shot peening process. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 This is a method flow chart of the shot blasting tank safety monitoring method based on machine learning in an embodiment of the present application. DETAILED DESCRIPTION

[0030] The embodiment of the present application discloses a shot peening tank safety monitoring method based on machine learning, which collects historical operating data of the shot peening tank, analyzes the historical data, and obtains a flow error sequence between the historical spray gun flow sequence and the specific historical shot peening sequence. The flow error sequence represents the difference between the actual shot peening flow and the spray gun flow. The real-time collected spray gun flow data is adjusted by the difference, so that the real-time shot peening flow data can be obtained. The shot peening tank is detected in real time by the similarity between the shot peening flow and the pressure data. During the operation of the shot peening tank, the defects and potential defects of the shot peening tank cause certain fluctuations in the pressure data and the actual shot peening data. The defects of the shot peening tank are analyzed based on the fluctuations. On the one hand, as the amount of historical data continues to increase and the depth of machine learning continues to deepen, it is easy to discover the potential defects of the shot peening tank. On the other hand, the shot peening tank is detected in real time to improve the timeliness of the shot peening tank detection.

[0031] Reference Figure 1 The shot blasting tank safety monitoring method based on machine learning includes steps S1 to S3.

[0032] S1: Collect historical data and divide the historical data into multiple data sequences, where the data sequences include historical pressure sequences, historical spray gun flow sequences, and historical shot peening sequences; cluster the historical pressure sequences to construct multiple pressure clusters.

[0033] Collect multiple operational data from the shot blasting tank's historical operation. In this embodiment, the collected operational data includes the tank's operating pressure, gun flow rate, and shot flow rate. The acquisition frequency in this embodiment is 1 Hz. The operating pressure data is collected by a pressure sensor, and the gun flow rate is collected by a flow sensor mounted on the gun. The shot flow rate refers to the actual flow rate of the ejected shots, which can be reflected by the number or type of shots ejected per unit time.

[0034] After historical data is collected, it forms a data sequence. A split window is set, and each time series is divided into multiple data sequences based on the split window size. Since multiple data items are collected in this application, the data sequences formed after historical data segmentation include a historical pressure sequence, a historical spray gun flow sequence, and a historical shot peening sequence. In this example, the split window size is 200, so the corresponding historical pressure sequence length is 200.

[0035] Calculate the distance between any two historical pressure series;

[0036] In one embodiment, the step of calculating two historical pressure sequences includes: taking the square of the difference between the same-order data in the two historical pressure sequences as the local distance, and taking the sum of the local distances corresponding to each data in the historical pressure sequences as the sequence distance between the two historical pressure sequences.

[0037] Specifically, the calculation formula of the sequence distance between two historical pressure series can be expressed as:

[0038] Where, Indicates the The historical pressure series and The sequence distance between the historical pressure series; Indicates the The first in the historical pressure series pressure values; Indicates the The first in the historical pressure series pressure values; Indicates the number of data in the historical pressure series.

[0039] This represents the difference between two pressure values of the same order in two historical pressure series, i.e., the local distance between the two historical pressure series. The sequence distance between the two historical pressure series is obtained by traversing the differences of all data in the historical pressure series. The smaller the sequence distance between the two historical pressure series, the more similar the data fluctuations of the two historical pressure series are.

[0040] In another embodiment, the step of calculating the two historical pressure sequences includes: obtaining the mean and variance of the pressure data in the historical pressure sequences; and determining the sequence distance between the two historical pressure sequences by the difference between the means and the variances of the two historical pressure sequences.

[0041] Specifically, the calculation formula for the sequence distance between two historical pressure series is:

[0042] Where, Indicates the The historical pressure series and The sequence distance between the historical pressure series; Indicates the The mean of the historical pressure series; Indicates the The mean of the historical pressure series; Indicates the The variance of the historical pressure series; Indicates the The variance of the historical pressure series.

[0043] In this method, the sequence distance between two historical pressure series is reflected from the differences in the overall mean and variance of the historical pressure series, so that the final calculated sequence distance is less susceptible to the influence of individual abnormal data, thereby improving the accuracy and robustness of the final sequence distance calculation.

[0044] The historical pressure sequences are clustered based on the sequence distances between the historical pressure sequences to form multiple pressure clusters. The clustering method used in this embodiment is k-means clustering, and other clustering methods may also be used in other embodiments.

[0045] S2: Based on the historical spray gun flow sequence and the historical shot peening sequence, the flow error sequence corresponding to the historical pressure sequence in the pressure cluster is obtained; based on the number and type of flow error sequences in the pressure cluster, the error distribution state of the pressure cluster is determined.

[0046] During the shot peening process, since shot ejection is based on compressed air pressure, changes in pressure data will lead to increases in both the gun and shot flow rates. There is a correlation between pressure and shot flow rates. However, during monitoring, fluctuations in pressure can lead to errors in shot flow rate collection (i.e., the difference between gun and shot flow rates). Furthermore, the error in shot flow rate collection varies under varying pressure conditions. Therefore, it is necessary to analyze the impact of pressure changes on the difference between gun and shot flow rates.

[0047] In the pressure cluster, each historical pressure sequence corresponds to a historical spray gun flow sequence and a shot peening sequence. The difference between the historical spray gun flow sequence and the shot peening sequence represents the difference between the shot peening flow data collected by the sensor at the spray gun and the actual shot peening flow data.

[0048] The historical spray gun flow sequence and the historical shot peening sequence have the same length. The flow difference is calculated by subtracting the flow rate of the same position in the historical shot peening sequence from the flow rate in the historical spray gun flow sequence. The flow difference between each data point in the historical spray gun sequence and the corresponding flow data point in the historical shot peening sequence together constitutes the flow error sequence. The flow error sequence corresponds to the historical pressure sequence, and the data in the flow error sequence represents the error between the measured spray gun flow rate and the actual shot peening flow rate under certain pressure conditions.

[0049] A pressure cluster group includes multiple historical pressure sequences, each of which corresponds to a flow error sequence. The multiple flow error sequences corresponding to the pressure clusters are clustered. The method for clustering flow error sequences is the same as the principle for clustering historical pressure sequences, so it will not be repeated here.

[0050] After clustering the flow error sequences, multiple error clusters are formed. That is, each pressure cluster corresponds to multiple error clusters. The mean of the errors of the same order in the multiple flow error sequences in the error cluster is calculated to obtain the representative error sequence corresponding to the error cluster. The number of flow error sequences in the error cluster represents the frequency of occurrence of the representative error sequence. The ratio between this frequency and the total number of flow error sequences corresponding to the pressure cluster is used as the probability of the representative error sequence. One pressure cluster corresponds to multiple error clusters, that is, one pressure cluster corresponds to multiple representative error sequences. The probability of each representative error sequence is calculated to obtain the error distribution data corresponding to the pressure cluster.

[0051] Specifically, the calculation formula for the probability of representing the error sequence is: ; Indicates the The representative error sequence in the pressure cluster is probability; Indicates the The first pressure cluster represents the number of elements in the error cluster corresponding to the error sequence; Indicates the The total number of flow error sequences corresponding to the pressure sequences.

[0052] For example, in one pressure cluster, there are 8 flow error sequences, and the 8 flow error sequences correspond to three error clusters. The three error clusters are defined as the first cluster, the second cluster, and the third cluster. The representative sequences corresponding to the first cluster, the second cluster, and the third cluster are distributed as the first representative sequence, the second representative sequence, and the third representative sequence. The first cluster includes 3 flow error sequences; the second cluster includes two flow error sequences; and the third cluster includes three flow error sequences. When the pressure sequence is in this pressure cluster, the flow error sequence between the spray gun flow and the shot peening flow under this pressure condition is The probability of the first representative sequence is The probability of is the second representative sequence; The probability of being the third representative sequence.

[0053] S3: Collect pressure data and spray gun flow data in real time, construct real-time pressure series and real-time spray gun flow series, and determine the cluster to which the real-time pressure series belongs based on the difference between the real-time pressure series and the historical pressure series corresponding to the cluster center of the pressure cluster.

[0054] During the shot peening process, real-time pressure data of the shot peening tank and spray gun flow data are collected. A real-time pressure sequence and a real-time spray gun flow sequence are obtained. The pressure sequence corresponding to the cluster center of the pressure cluster is obtained, and the pressure sequence is determined to be the center pressure sequence, and the attribution distance between the real-time pressure sequence and the different center pressure sequences is calculated. The calculation of the attribution distance is the same as the calculation method of the sequence distance between the pressure sequences in step S2, and will not be repeated here. The pressure cluster corresponding to the center pressure sequence with the smallest attribution distance is used as the attribution cluster of the real-time pressure sequence.

[0055] S4: According to the error distribution state corresponding to the cluster to which it belongs, the real-time spray gun flow sequence is adjusted to obtain the real-time shot peening sequence; the similarity between the real-time shot peening sequence and the real-time pressure sequence is calculated, and the similarity is weighted and summed based on the error distribution state; the result of the weighted sum of the similarities is used as the change consistency.

[0056] The belonging clusters correspond to multiple representative error sequences, and the real-time spray gun flow sequence is adjusted by the representative error sequences. Specifically, the real-time spray gun flow sequence is added to the flow data of the same sequence of different representative error sequences to obtain multiple real-time shot peening sequences.

[0057] The similarity of fluctuations between the real-time shot peening sequence and the real-time pressure sequence is calculated, and the similarities are weighted and summed based on the error distribution state. The result of the weighted summation of the similarities is used as the change consistency, and the operating status of the tank is judged based on the change consistency.

[0058] In this embodiment, the similarity is the Pearson correlation coefficient between the real-time shot peening sequence and the real-time pressure sequence.

[0059] Specifically, the calculation formula for change consistency can be expressed as: ; Indicates the consistency of changes between the real-time shot peening sequence and the real-time pressure sequence; Indicates the The similarity between the real-time shot peening sequence and the real-time pressure sequence; Indicates the The probability of representing the error sequence corresponding to a real-time shot peening sequence.

[0060] After determining the cluster to which the real-time pressure sequence belongs, it is shown that under this pressure condition, there are multiple error sequences (i.e., multiple representative error sequences) formed by the error between the real-time spray gun flow and the real-time shot peening flow, and different representative sequences have different corresponding probabilities. Based on this probability, the similarity between the real-time shot peening sequence and the real-time pressure sequence is weighted and summed to obtain the change consistency between the real-time shot peening flow and the real-time pressure sequence, and then it can be judged whether the changes in the real-time shot peening flow with the real-time pressure are related, and then the current working status of the shot peening tank can be judged.

[0061] S5: Determine the operating status of the storage tank based on the consistency of the changes.

[0062] An alarm threshold is set, and an alarm is issued in response to when the change consistency is less than the alarm threshold, judging that the current working state of the shot peening tank is abnormal.

[0063] An embodiment of the present application also discloses a shot blasting tank safety monitoring system based on machine learning, including a processor and a memory, wherein the memory stores computer program instructions. When the computer program instructions are executed by the processor, the shot blasting tank safety monitoring method based on machine learning according to the present application is implemented.

[0064] The above system also includes other components well known to those skilled in the art, such as a communication bus and a communication interface. The configuration and functions of these components are known in the art and will not be described in detail here.

[0065] The above are all preferred embodiments of the present application, and are not intended to limit the scope of protection of the present application. Therefore, any equivalent changes made based on the structure, shape, and principle of the present application should be included in the scope of protection of the present application.

Claims

1. A shot peening tank safety monitoring method based on machine learning, characterized in that: Collect historical data and divide the historical data into multiple data sequences, wherein the data sequences include historical pressure sequences, historical spray gun flow sequences, and historical shot peening sequences; cluster the historical pressure sequences to construct multiple pressure clusters; Based on the historical spray gun flow sequence and the historical shot peening sequence, the flow error sequence corresponding to the historical pressure sequence in the pressure cluster is obtained; Determining the error distribution state of the pressure cluster based on the number and type of flow error sequences in the pressure cluster; Collect pressure data and spray gun flow data in real time, construct real-time pressure series and real-time spray gun flow series, and determine the cluster to which the real-time pressure series belongs based on the difference between the real-time pressure series and the historical pressure series corresponding to the cluster center of the pressure cluster; The real-time shot peening sequence is obtained by adjusting the real-time gun flow rate sequence according to the error distribution state corresponding to the cluster to which it belongs; the similarity between the real-time shot peening sequence and the real-time pressure sequence is calculated, and the similarity is weighted and summed based on the error distribution state; the result of the weighted sum of the similarities is used as the change consistency; The operating status of the tank is determined based on the consistency of changes.

2. The shot peening tank safety monitoring method based on machine learning according to claim 1 is characterized in that: The steps of clustering the historical pressure sequences and constructing a plurality of pressure clusters include: calculating the sequence distance between any two historical pressure sequences, and clustering the historical pressure sequences based on the sequence distance between the two historical pressure sequences.

3. The shot blasting tank safety monitoring method based on machine learning according to claim 2 is characterized in that: The kmeans clustering method is used to cluster the historical pressure series.

4. The shot peening tank safety monitoring method based on machine learning according to claim 2 is characterized in that: The step of calculating the sequence distance between any two historical pressure sequences includes: taking the square of the difference between the same-order data in the two historical pressure sequences as the local distance, and taking the sum of the local distances corresponding to each data in the historical pressure sequences as the sequence distance between the two historical pressure sequences.

5. The shot blasting tank safety monitoring method based on machine learning according to claim 2 is characterized in that: The steps of calculating the sequence distance between any two historical pressure sequences include: obtaining the mean and variance of the pressure data in the historical pressure sequences; and determining the sequence distance between the two historical pressure sequences by the difference between the means and the variances of the two historical pressure sequences.

6. The shot blasting tank safety monitoring method based on machine learning according to claim 1 is characterized in that: The step of obtaining the flow error sequence includes calculating the flow difference between the historical spray gun flow sequence and the data of the same sequence in the shot peening sequence, and a plurality of flow differences constitute the flow error sequence.

7. The shot blasting tank safety monitoring method based on machine learning according to claim 1 is characterized in that: The steps of determining the error distribution state of a pressure cluster based on the number and types of flow error sequences in the pressure cluster include: clustering the corresponding flow error sequences in the pressure cluster to obtain multiple error clusters; obtaining the mean of the same-order data of multiple flow error sequences in the error cluster to form a representative error sequence; taking the ratio of the number of elements in the error cluster to the total number of corresponding flow error sequences in the pressure cluster group as the probability of the representative error sequence; obtaining the probabilities of multiple representative error sequences to obtain the error distribution state corresponding to the pressure cluster.

8. The shot blasting tank safety monitoring method based on machine learning according to claim 1 is characterized in that: The step of adjusting the real-time spray gun flow sequence according to the flow error sequence in the pressure cluster to obtain the real-time shot peening sequence includes: adding the real-time spray gun flow sequence to the data representing the same sequence in the error sequence to obtain the real-time shot peening sequence.

9. The shot blasting tank safety monitoring method based on machine learning according to claim 1 is characterized in that: The step of determining the operating state of the storage tank based on the change consistency includes setting an alarm threshold, and issuing an alarm in response to the change consistency being less than the alarm threshold.

10. The shot blasting tank safety monitoring system based on machine learning is characterized by: 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 shot blasting tank safety monitoring method based on machine learning according to any one of claims 1 to 9 is implemented.

Images