Industrial data optimization query method and system based on machine learning

The method addresses the challenge of identifying machine vibration anomalies by using machine learning to monitor temperature and vibration data, improving data query accuracy and reliability through outlier elimination.

CN118643453BActive Publication Date: 2024-12-06DALIAN SIX ELEMENTS TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411119999.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-15
Publication Date
2024-12-06
Estimated Expiration
2044-08-15

AI Technical Summary

Technical Problem

In complex industrial environments, minor anomalies caused by machine vibration, such as loose screws, are difficult to identify, leading to increased outliers in data and reduced accuracy of machine learning models, affecting their generalization ability and subsequent data query accuracy.

Method used

An industrial data optimization query method using machine learning that monitors temperature and vibration data to calculate dynamic temperature and vibration anomaly indexes, identifies abnormal data, and constructs a data query model to eliminate outliers, improving accuracy and reliability.

Benefits of technology

The method accurately diagnoses fault locations and minimizes the impact of outliers on model learning, enhancing query accuracy and reliability by systematically eliminating abnormal data.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118643453B_ABST
    Figure CN118643453B_ABST
Patent Text Reader

Abstract

The present invention relates to the field of data processing technology, and specifically to an industrial data optimization query method and system based on machine learning, including: collecting the instantaneous temperature, stress, amplitude and distance data of each connection part of the machine. Calculating the dynamic temperature difference contribution index and temperature difference index of each part, and then obtaining the abnormal evaluation index. Combining the amplitude, stress and abnormal index, the vibration abnormality detection index is calculated. Combining the vibration abnormality detection index to build a data query system. The present invention constructs a data query system by eliminating abnormal marking data, thereby improving the query accuracy and reliability.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to an industrial data optimization query method and system based on machine learning. Background Art

[0002] In modern industrial production, the widespread application of IoT technology has enabled a large number of industrial equipment to be interconnected and generate massive amounts of real-time data. The efficient processing and analysis of this data is of great significance for improving production efficiency and equipment maintenance. However, the operation of machines in complex production environments is affected, and minor anomalies caused by machine vibration, such as loose screws, are difficult to identify, which poses challenges to data analysis. Outliers in the data increase the chances of machine learning models learning abnormal behavior data, reduce the generalization ability of the model in practical applications, and thus reduce the accuracy of subsequent data queries. Summary of the invention

[0003] The present invention provides an industrial data optimization query method and system based on machine learning to solve existing problems.

[0004] The industrial data optimization query method and system based on machine learning of the present invention adopts the following technical solutions:

[0005] An embodiment of the present invention provides an industrial data optimization query method based on machine learning, the method comprising the following steps:

[0006] Obtain the temperature value, stress value and amplitude value of each connection part of the machine at each moment; obtain the distance between any two connection parts of the machine;

[0007] According to the temperature value of each connection part of the machine at each moment, the dynamic temperature difference contribution index of each connection part of the machine at each moment is obtained;

[0008] According to the temperature value of each connection part of the machine at the same time, the number of connection parts on the machine and the distance between any two connection parts, the temperature difference index of each connection part of the machine at each time is obtained;

[0009] According to the dynamic temperature difference contribution index and temperature difference index of each connection part of the machine at each moment, the abnormality assessment index of each connection part of the machine at each moment is obtained; according to the amplitude value, stress value and abnormality assessment index of each connection part of the machine at the same moment, the vibration abnormality detection index of each connection part of the machine at each moment is obtained; according to the vibration abnormality detection index of each connection part at each moment, a data query model is constructed.

[0010] Preferably, the step of obtaining the dynamic temperature difference contribution index of each connection part of the machine at each moment according to the temperature value of each connection part of the machine at each moment comprises the following specific steps:

[0011] The machine The first connection part Moment and The absolute value of the temperature difference at the moment is recorded as the machine Connection part The temperature change at each moment;

[0012] According to the temperature change and temperature value of each connection part of the machine at each moment, the temperature change of each connection part of the machine at each moment is obtained.

[0013] Preferably, the step of obtaining the temperature change amount of each connection part of the machine at each moment according to the temperature change amount and temperature value of each connection part of the machine at each moment comprises the following specific steps:

[0014]

[0015] Where: Indicates the machine Connection part The dynamic temperature difference contribution index at each moment, Indicates the machine Connection part The temperature change at a given moment, Indicates that the machine Connection part The temperature value at a moment, Indicates that the machine Connection part The temperature value at a moment, Indicates the machine The number of moments corresponding to each connection part.

[0016] Preferably, the step of obtaining the temperature difference index of each connection part of the machine at each moment according to the temperature value of each connection part of the machine at the same moment, the number of connection parts on the machine and the distance between any two connection parts comprises the following specific steps:

[0017]

[0018] Where: Indicates the machine Connection part The temperature difference index at each moment, Indicates the machine Connection part The temperature value at a moment, Indicates the machine The connection point is The temperature value at a moment, Indicates the machine The connection part and The distance between the connection parts, Indicates the number of connection parts on the machine. Indicates taking the absolute value, Represents an exponential function with a natural constant as its base.

[0019] Preferably, the method of obtaining the abnormality evaluation index of each connection part of the machine at each moment according to the dynamic temperature difference contribution index and the temperature difference index of each connection part of the machine at each moment comprises the following specific steps:

[0020] The product of the dynamic temperature difference contribution index and the temperature difference index of each connection part of the machine at each moment is recorded as the abnormal evaluation index of each connection part of the machine at each moment.

[0021] Preferably, the step of obtaining the vibration anomaly detection index of each connection part of the machine at each moment according to the amplitude value, stress value and anomaly assessment index of each connection part of the machine at the same moment comprises the following specific steps:

[0022] According to the amplitude value and stress value of each connection part of the machine at the same time, the difference cumulative index between each connection part of the machine and each connection part at each time is obtained;

[0023] According to the difference accumulation index and the abnormality evaluation index between each connection part of the machine and each connection part at each moment, the vibration abnormality detection index of each connection part of the machine at each moment is obtained.

[0024] Preferably, the step of obtaining the vibration anomaly detection index of each connection part of the machine at each moment according to the difference cumulative index and the anomaly evaluation index between each connection part of the machine at each moment comprises the following specific steps:

[0025]

[0026] Where: Indicates the machine Connection part The vibration anomaly detection index at each moment, Indicates the machine Connection part The abnormal evaluation index at the moment, Indicates the machine Connection part The stress value at the moment, Indicates the machine Connection part The amplitude value at a moment, Indicates the machine Connection part The amplitude value at a moment, Indicates taking the absolute value, Indicates the number of connected parts of the machine. represents the normalization function, Indicates the machine The connection part and The connection point is The difference accumulation index at each moment.

[0027] Preferably, the data query model is constructed according to the vibration anomaly detection index of each connection part at each moment, and the specific steps include the following:

[0028] when When it is greater than or equal to the preset threshold, the machine Connection part All data at each moment are marked with abnormal labels; among them, Indicates the machine Connection part The vibration anomaly detection index at each moment;

[0029] when When the value is less than the preset threshold, the machine Connection part All data at each moment are labeled normally;

[0030] A data query model is constructed based on all data corresponding to the abnormal label markings and the normal label markings.

[0031] Preferably, the data query model is constructed according to all the data corresponding to the abnormal label marks and the normal label marks, and the specific steps include the following:

[0032] All data at the time marked with abnormal labels are eliminated, and all data at the time corresponding to the normal label marks are input into the algorithm model for training to build a data optimization query model.

[0033] The present invention also proposes an industrial data optimization query system based on machine learning, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program stored in the memory to implement the steps of the aforementioned industrial data optimization query method based on machine learning.

[0034] The beneficial effects of the technical solution of the present invention are as follows: the present invention monitors the dynamic changes in the temperature of each connection part of the machine and calculates the temperature difference index to warn of faults. At the same time, the temperature differences between components are compared to accurately diagnose the fault location. Combining these two indexes, an abnormality assessment index is generated to quantify the degree of abnormality in the operating status of the equipment. Further combined with data such as amplitude and stress, the vibration anomaly detection index is calculated to accurately identify problem data. Finally, by eliminating problem data in data processing to build a model, the impact of outliers on the model learning process is minimized, thereby improving query accuracy and reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0036] Figure 1 The present invention is a flowchart of the steps of the industrial data optimization query method based on machine learning. DETAILED DESCRIPTION

[0037] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation, structure, features and effects of the industrial data optimization query method and system based on machine learning proposed by the present invention in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0038] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0039] The specific scheme of the industrial data optimization query method and system based on machine learning provided by the present invention is described in detail below with reference to the accompanying drawings.

[0040] See also Figure 1 , which shows a flowchart of a method for optimizing industrial data query based on machine learning provided by an embodiment of the present invention, the method comprising the following steps:

[0041] Step S001: Obtain the temperature value, stress value and amplitude value of each connection part of the machine at each moment; obtain the distance between any two connection parts of the machine.

[0042] Obtain the temperature, stress and amplitude values ​​of each connection part of the machine at every moment.

[0043] Get the distance between any two connected parts of the machine.

[0044] It should be noted that, in this embodiment, the temperature time series data is obtained by temperature sensing, and the stress value and vibration amplitude are obtained by placing vibration and stress sensors at each connection part. The acquisition frequency is once a minute and the acquisition time is 8 hours.

[0045] Step S002: According to the temperature value of each connection part of the machine at each moment, obtain the dynamic temperature difference contribution index of each connection part of the machine at each moment.

[0046] During the operation of the machine, due to its working nature, it will produce continuous vibration. This vibration often causes the screws in key parts to loosen or fall off, making the connection parts loose. Although this problem is not caused by component damage, it is very difficult to determine this looseness through vibration or abnormal noise. In addition, as the machine continues to operate, this abnormal vibration occurs synchronously, affecting the accuracy of the collected data. Therefore, by analyzing the specificity of the vibration at the connection part after the screw is loosened, the connection condition of the machine part can be judged, and then the accuracy of the relevant data can be determined.

[0047] When the machine is in operation, if the screws of the connection parts are loose, it will cause relative friction between the parts and the mating parts. This friction generates heat, causing the temperature of the loose parts to rise rapidly, exceeding the temperature of other normal parts. On the contrary, under normal circumstances, since the screws of the connection parts are tightened, the various parts of the machine are tightly connected into one, and no relative movement or friction occurs during operation, so the temperature of the connection parts remains stable and does not change significantly.

[0048] The machine The first connection part Moment and The absolute value of the temperature difference at the moment is recorded as the machine Connection part The temperature change at a given moment.

[0049] It should be noted that: in this embodiment, the temperature change amount at the last moment is set to be the temperature change amount at the second to last moment, and this is used as an example for description.

[0050] This embodiment provides a method for calculating the dynamic temperature difference contribution index as follows:

[0051]

[0052] Where: Indicates the machine Connection part The dynamic temperature difference contribution index at each moment, Indicates the machine Connection part The temperature change at a given moment, Indicates that the machine Connection part The temperature value at a moment, Indicates that the machine Connection part The temperature value at a moment, Indicates the machine The number of moments corresponding to each connection part.

[0053] The above formula indicates the degree of temperature change of the connection part as time changes in the time series. The larger the value, the higher the temperature of the connection part of the machine in operation. In the time series The more drastic the temperature change at a certain time, the In the time series The greater the non-robustness to temperature changes at each moment.

[0054] Step S003: According to the temperature value of each connection part of the machine at the same time, the number of connection parts on the machine and the distance between any two connection parts, the temperature difference index of each connection part of the machine at each time is obtained.

[0055] When analyzing temperature data during machine operation, the focus is on the temperature differences between the various connections. This analysis must take into account that the operating machine is affected by the heat generated internally and the changes in the external ambient temperature. To accurately evaluate these effects, it is important to compare the temperature differences between different connections at the same time point.

[0056] This embodiment provides a method for calculating the temperature difference index of the connection part:

[0057]

[0058] Where: Indicates the machine Connection part The temperature difference index at each moment, Indicates the machine Connection part The temperature value at a moment, Indicates the machine The connection point is The temperature value at a moment, Indicates the machine The connection part and The distance between the connection parts, Indicates the number of connection parts on the machine. Indicates taking the absolute value, represents an exponential function with a natural constant as the base. To present the inverse proportional relationship and normalization processing, the implementer can set the inverse proportional function and normalization function according to the actual situation.

[0059] It should be noted that when analyzing the heat conduction characteristics of a machine, the key is to understand that the machine's connection parts are usually made of metal materials with good heat conductivity, such as iron, steel, nickel, and copper. These materials can quickly transfer heat, and the machine design often considers aesthetics or reducing the volume, so the connection parts are spaced closer together. Therefore, if the screws of a certain connection part are loose, as the machine runs, the high temperature generated there can quickly affect the temperature data of the surrounding connection parts. Therefore, in order to accurately reflect the characteristics of this temperature effect changing with distance, the inverse function is used in the formula Multiply by the absolute value of the temperature difference This method ensures that the farther away the target connection is, the smaller its impact on the temperature difference calculation, thus more accurately capturing the long-range temperature changes caused by heat conduction.

[0060] Step S004: According to the dynamic temperature difference contribution index and temperature difference index of each connection part of the machine at each moment, obtain the abnormality evaluation index of each connection part of the machine at each moment; according to the amplitude value, stress value and abnormality evaluation index of each connection part of the machine at the same moment, obtain the vibration abnormality detection index of each connection part of the machine at each moment; construct a data query model according to the vibration abnormality detection index of each connection part at each moment.

[0061] Based on the temperature data recorded during the operation of each connection part of the machine, the abnormal temperature change values ​​of each part can be systematically calculated. This process involves comparing the temperature data at each time point with the historical average or expected range to identify temperature changes that exceed the normal fluctuation range. The calculated abnormal values ​​help to detect potential failures or performance degradation problems in a timely manner, so that corresponding maintenance measures can be taken to ensure the efficient and reliable operation of the machine.

[0062] This embodiment provides a method for calculating the abnormal evaluation index as follows:

[0063]

[0064] Where: Indicates the machine Connection part The abnormal evaluation index at the moment, Indicates the machine Connection part The dynamic temperature difference contribution index at each moment, Indicates the machine Connection part The temperature difference index at each moment.

[0065] The vibration generated by the machine during operation makes it difficult to determine the vibration source of the connection. However, when the screws of the connection are loose, the machine that originally operated in coordination will produce additional relative vibration due to the local looseness, resulting in an increase in the vibration amplitude of the part. In addition, the loosening of the screws will break the force balance of the original part of the machine, reducing the stress of the connection part. By analyzing the vibration data of the machine during operation, the degree of increase in the vibration amplitude of the connection part caused by the loosening of the screws can be calculated, thus providing an important basis for the maintenance of the machine.

[0066] This embodiment provides a method for calculating the connection part vibration abnormality detection index:

[0067]

[0068] Where: Indicates the machine Connection part The vibration anomaly detection index at each moment, Indicates the machine Connection part The abnormal evaluation index at the moment, Indicates the machine Connection part The stress value at the moment, Indicates the machine Connection part The amplitude value at a moment, Indicates the machine Connection part The amplitude value at a moment, Indicates taking the absolute value, Indicates the number of connected parts of the machine. represents the normalization function, Indicates the machine The connection part and The connection point is The difference accumulation index at each moment.

[0069] It should be noted that in the above formula It is a moment in the collection process and its value is not 1.

[0070] The above formula first calculates the different connection parts at a given time and The difference in vibration amplitude between the two parts is extended to the connection part. The difference between the first and the second connection points is calculated based on the difference between the first and the second connection points. As time goes by, the stress in the connection point gradually decreases due to the loosening of the screws. Therefore, different weight values ​​are assigned to different time points according to the change in stress. By accumulating these weighted differences, the first The degree of increase in vibration amplitude at each moment.

[0071] In order to improve the monitoring efficiency and accuracy of the machine, this embodiment sets a preset threshold value for analyzing the amplitude data of the connecting part of the machine.

[0072] when When it is greater than or equal to the preset threshold, the machine Connection part All data at a certain moment are marked with abnormal labels.

[0073] when When the value is less than the preset threshold, the machine Connection part All data at each moment are labeled normally.

[0074] According to the above method, the label type corresponding to all the data at each connection part of the machine at each moment is obtained.

[0075] In this embodiment, the preset threshold is 0.7, which is used as an example for description. In other implementations, it can be set to other values, which are not limited in this embodiment.

[0076] All data at the moment marked with abnormal labels are eliminated, and all data at the moment corresponding to the normal label marking are input into the algorithm model for training, so as to build an efficient and reliable data optimization query model, which can realize the subsequent early warning of the input data marked with abnormal labels.

[0077] It should be noted that by removing outliers from the training data, the possibility of the model learning these abnormal behaviors can be reduced, thereby avoiding the model's misjudgment of abnormal data when data is marked with abnormal labels during the prediction stage. When data marked with abnormal labels is subsequently input, the system will be able to effectively issue an early warning.

[0078] The algorithm model may be machine learning, convolutional neural network (CNN), Python, etc. This is described by way of example and is not specifically limited in this embodiment.

[0079] The present invention also provides an industrial data optimization query system based on machine learning, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program stored in the memory to implement the steps of the aforementioned industrial data optimization query method based on machine learning.

[0080] So far, the present invention is completed.

[0081] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An industrial data optimization query method based on machine learning, characterized in that: The method comprises the following steps: Obtain the temperature value, stress value and amplitude value of each connection part of the machine at each moment; obtain the distance between any two connection parts of the machine; According to the temperature value of each connection part of the machine at each moment, the dynamic temperature difference contribution index of each connection part of the machine at each moment is obtained; According to the temperature value of each connection part of the machine at the same time, the number of connection parts on the machine and the distance between any two connection parts, the temperature difference index of each connection part of the machine at each time is obtained; According to the dynamic temperature difference contribution index and temperature difference index of each connection part of the machine at each moment, the abnormal evaluation index of each connection part of the machine at each moment is obtained; according to the amplitude value, stress value and abnormal evaluation index of each connection part of the machine at the same moment, the vibration abnormal detection index of each connection part of the machine at each moment is obtained; according to the vibration abnormal detection index of each connection part at each moment, a data query model is constructed; The specific steps of obtaining the dynamic temperature difference contribution index of each connection part of the machine at each moment according to the temperature value of each connection part of the machine at each moment are as follows: The machine The first connection part The moment and The absolute value of the temperature difference at the moment is recorded as the machine Connection part The temperature change at each moment; According to the temperature change and temperature value of each connection part of the machine at each moment, the temperature change of each connection part of the machine at each moment is obtained; The specific steps of obtaining the temperature change amount of each connection part of the machine at each moment according to the temperature change amount and temperature value of each connection part of the machine at each moment are as follows: Where: Indicates the machine Connection part The dynamic temperature difference contribution index at each moment, Indicates the machine Connection part The temperature change at a given moment, Indicates that the machine Connection part The temperature value at a moment, Indicates that the machine Connection part The temperature value at a moment, Indicates the machine The number of moments corresponding to each connection part; The specific steps of obtaining the temperature difference index of each connection part of the machine at each moment according to the temperature value of each connection part of the machine at the same moment, the number of connection parts on the machine and the distance between any two connection parts are as follows: Where: Indicates the machine Connection part The temperature difference index at each moment, Indicates the machine Connection part The temperature value at a moment, Indicates the machine The connection point is The temperature value at a moment, Indicates the machine The connection part and The distance between the connection parts, Indicates the number of connection parts on the machine. Indicates taking the absolute value, represents an exponential function with a natural constant as base; The specific steps of obtaining the abnormal evaluation index of each connection part of the machine at each moment according to the dynamic temperature difference contribution index and the temperature difference index of each connection part of the machine at each moment are as follows: The product of the dynamic temperature difference contribution index and the temperature difference index of each connection part of the machine at each moment is recorded as the abnormal evaluation index of each connection part of the machine at each moment; The specific steps of obtaining the vibration anomaly detection index of each connection part of the machine at each moment according to the amplitude value, stress value and anomaly evaluation index of each connection part of the machine at the same moment are as follows: According to the amplitude value and stress value of each connection part of the machine at the same time, the difference cumulative index between each connection part of the machine and each connection part at each time is obtained; According to the difference accumulation index and abnormal evaluation index of each connection part of the machine at each moment, the vibration abnormality detection index of each connection part of the machine at each moment is obtained; The specific steps of obtaining the vibration anomaly detection index of each connection part of the machine at each moment according to the difference accumulation index and the anomaly evaluation index of each connection part of the machine at each moment are as follows: Where: Indicates the machine Connection part The vibration anomaly detection index at each moment, Indicates the machine Connection part The abnormal evaluation index at the moment, Indicates the machine Connection part The stress value at the moment, Indicates the machine Connection part The amplitude value at a moment, Indicates the machine Connection part The amplitude value at a moment, Indicates taking the absolute value, Indicates the number of connected parts of the machine. represents the normalization function, Indicates the machine The connection part and The connection point is The difference accumulation index at each moment; The data query model is constructed according to the vibration anomaly detection index of each connection part at each moment, and the specific steps include the following: when When it is greater than or equal to the preset threshold, the machine Connection part All data at each moment are marked with abnormal labels; among them, Indicates the machine Connection part The vibration anomaly detection index at each moment; when When the value is less than the preset threshold, the machine Connection part All data at each moment are labeled normally; A data query model is constructed based on all data corresponding to the abnormal label markings and the normal label markings.

2. The industrial data optimization query method based on machine learning according to claim 1 is characterized in that: The data query model is constructed according to all the data corresponding to the abnormal label mark and the normal label mark, and the specific steps include the following: All data at the time marked with abnormal labels are eliminated, and all data at the time corresponding to the normal label marks are input into the algorithm model for training to build a data optimization query model.

3. An industrial data optimization query system based on machine learning, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the computer program is executed by a processor, the steps of the industrial data optimization query method based on machine learning are implemented as described in any one of claims 1-2.