An industrial equipment operation and maintenance intelligent monitoring method based on multi-dimensional data fusion

Abstract

The application discloses an industrial equipment operation and maintenance intelligent monitoring method based on multi-dimensional data fusion, relates to the technical field of industrial equipment operation and maintenance, and comprises the following steps: obtaining real-time data acquisition time corresponding to industrial equipment; obtaining data missing time according to real-time operation and maintenance data corresponding to various data of the industrial equipment at different real-time data acquisition time; obtaining missing values corresponding to the missing time according to an interval reference diagram of a frequency interval corresponding to a real-time acquisition frequency; drawing a calibration line H with the value of the missing time on the X axis as an end point; selecting distribution area blocks belonging to the corresponding interval reference diagram by judging the positional relationship between the calibration line H and the area blocks; and calculating the mean value of the ordinate of the center points of the distribution area blocks as the missing values corresponding to the missing time. The missing values corresponding to the missing time can be accurately estimated, so that the integrity of the real-time operation and maintenance data is ensured.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of industrial equipment operation and maintenance, and specifically relates to an industrial equipment operation and maintenance intelligent monitoring method based on multi-dimensional data fusion. BACKGROUND

[0002] With the development of industrial intelligence, the dependence of industrial equipment operation and maintenance monitoring on data is increasing, the complexity and automation level of industrial equipment are continuously improving, and industrial equipment will acquire various data such as temperature, pressure, current, voltage and vibration in real time through various sensors and monitoring units during operation to support equipment health state evaluation, fault prediction and operation and maintenance decision-making, and provide more comprehensive and accurate visual basis for equipment health evaluation and fault prediction.

[0003] However, at present, the data acquisition frequency in industrial equipment operation and maintenance is mostly a fixed value, and due to network delay, sensor failure, data upload interruption and equipment aging and other factors, there are inevitably missing points or missing segments in industrial equipment operation and maintenance data, which leads to insufficient fusion analysis capability of multi-dimensional operation data, and when data is missing, it is difficult to accurately fill in the missing values, affecting the accuracy and timeliness of operation and maintenance monitoring. SUMMARY

[0004] The purpose of the application is to provide an industrial equipment operation and maintenance intelligent monitoring method based on multi-dimensional data fusion, which solves the technical problem that when industrial equipment has missing operation and maintenance data, it is difficult to accurately fill in the missing values, affecting the accuracy and timeliness of operation and maintenance monitoring.

[0005] An industrial equipment operation and maintenance intelligent monitoring method based on multi-dimensional data fusion, comprising the following steps:

[0006] Step 1: set multiple industrial equipment as an experimental equipment group, set multiple detection data acquisition frequencies within a preset detection time T, and obtain the multi-dimensional operation data corresponding to the experimental equipment group under different detection data acquisition frequencies;

[0007] Step 2: analyze the multi-dimensional operation data corresponding to the experimental equipment group under different detection data acquisition frequencies, and obtain the calculation values corresponding to the experimental equipment group at different data acquisition times within different detection data acquisition frequencies;

[0008] Step three: divide each detection data into multiple frequency intervals in the order from front to back, obtain the reference values corresponding to different data acquisition times in the frequency according to each detection data, obtain the detection value distribution diagram corresponding to each detection data acquisition frequency respectively, and obtain the interval reference diagram corresponding to different frequency intervals respectively according to the detection value distribution diagrams corresponding to two detection data acquisition frequencies constituting different frequency intervals;

[0009] Step four: obtain the real-time operation data and real-time acquisition frequency of various data of the industrial equipment according to the preset detection time T, and obtain the real-time data acquisition time corresponding to the industrial equipment, obtain the data missing time according to the real-time operation data corresponding to different real-time data acquisition times of various data of the industrial equipment, and obtain the missing value corresponding to the missing time according to the interval reference diagram of the frequency interval corresponding to the real-time acquisition frequency.

[0010] As a further scheme of the application: the specific way of obtaining the calculation value corresponding to different data acquisition times of the experimental equipment group in different detection data acquisition frequencies is:

[0011] S1: take the ratio between each detection data acquisition frequency Cc and the preset detection time T as different detection data acquisition frequencies Vc, wherein c represents different detection data acquisition frequencies, c=1, 2, …, r, r represents the total number of detection data acquisition frequencies, r is a positive integer, and r≧6, and randomly select one from the different detection data acquisition frequencies Cc as the target frequency C1; obtain the running data of various data of different industrial equipment in the experimental equipment group corresponding to the target frequency C1, take the ratio between the preset detection time T and the target frequency C1 as the time interval t1 corresponding to the target frequency, obtain different data acquisition times et1 under the detection data acquisition frequency corresponding to the target frequency, and obtain the multi-dimensional running data corresponding to each data acquisition time et1 of each industrial equipment in the experimental equipment respectively, wherein e represents different data acquisition times, e=1, 2, …, n1, n1 represents the total number of target frequencies, n1 is a positive integer, and n1≧6;

[0012] S2: randomly obtain one from each data acquisition time et1 as the target time 1t1, and mark the running data corresponding to the target time 1t1 of each industrial equipment set as Aji, wherein i is the different types of data in the multi-dimensional running data, j represents different industrial equipment in the experimental equipment group, i=1, 2, …, a1, a1 represents the total number of data types in different types of data, a1 is a positive integer, and a1≧2;

[0013] According to the running data Aji of the different kinds of data of each industrial equipment at the target time 1t1, the reference values Di corresponding to the different kinds of data at the target time 1t1 are obtained, and the calculation values F11 corresponding to the industrial equipment at the target time 1t1 are obtained by calculating the reference values Di corresponding to the different kinds of data at the target time 1t1.

[0014] S3: repeating step S2, analyzing the running data of the different kinds of data of each industrial equipment at the remaining data acquisition times, so as to obtain the calculation values Fe1 corresponding to each data acquisition time et1 at the target frequency;

[0015] S4: repeating steps S1-S3, analyzing the running data of each kind of data of different industrial equipment in the experimental equipment group at the remaining detection data acquisition frequencies, so as to obtain the reference values corresponding to the different data acquisition times at each detection data acquisition frequency.

[0016] As a further scheme of the present application, the specific way of obtaining the interval reference diagram corresponding to each frequency interval is:

[0017] Obtaining the maximum value J in the calculation values corresponding to the different data acquisition times of the experimental equipment group at different detection data acquisition frequencies, and obtaining the value area according to the maximum value J and the preset detection time T, the value area is uniformly divided into a plurality of equal size area blocks, and then a two-dimensional coordinate model is generated;

[0018] S01: obtaining the frequency interval with the target frequency as the lower limit value of the interval from the plurality of frequency intervals as the analysis interval, obtaining the calculation values Fe1 corresponding to each data acquisition time et1 at the target frequency, marking the corresponding data points in the two-dimensional coordinate model according to the calculation values Fe1 corresponding to each data acquisition time et1 at the target frequency, marking the distribution area block corresponding to the target frequency according to the position of each data point in each area block in the two-dimensional coordinate model, and then obtaining the detection value distribution diagram corresponding to the target frequency; at the same time, obtaining the detection data acquisition frequency corresponding to the upper limit value of the analysis interval from the analysis interval, taking the ratio between the preset detection time T and the detection data acquisition frequency corresponding to the upper limit value as the corresponding time interval t2, and then obtaining the different data acquisition times ct2 at the detection data acquisition frequency corresponding to the upper limit value, and marking the calculation values at the different data acquisition times ct2 at the detection data acquisition frequency corresponding to the upper limit value as Dc, wherein c represents different data acquisition times, c=1, 2, …, n2, n2 represents the total number of different data acquisition times at the detection data acquisition frequency corresponding to the upper limit value, n2 is a positive integer, and n2≧6;

[0019] Mark the calculation value at different data acquisition time ct2 as Dc, mark the corresponding data points in the two-dimensional coordinate model, and obtain the detection data acquisition frequency corresponding to the upper limit value of the interval The detection value distribution diagram K2 is obtained in the same way as the detection value distribution diagram K1 corresponding to the target frequency, and the distribution area blocks corresponding to the detection value distribution diagrams K1 and K2 are merged, and then the interval reference diagram Q1 corresponding to the analysis interval is obtained.

[0020] S03: Repeat steps S01-S02 to perform the same analysis and processing on the remaining each frequency interval, and then obtain the interval reference diagram Qq corresponding to each frequency interval, wherein q represents different frequency intervals, q=1, 2, …, r-1.

[0021] As a further scheme of the present application: the specific way of marking the distribution area block corresponding to the target frequency is:

[0022] When the data point is located in the area block, the corresponding area block is taken as the distribution area block; when the data point is located on the horizontal edge line of the area block, the area blocks on the upper and lower sides of the corresponding horizontal edge line are taken as the distribution area block; when the data point is located on the vertical edge line of the area block, the area blocks on the left and right sides of the corresponding vertical edge line are taken as the distribution area block; When the data point is located at the boundary node of the area block, the area blocks on the upper, lower, left and right sides of the boundary node are taken as the distribution area block, that is, the distribution area block corresponding to the target frequency is obtained and marked, and then the detection value distribution diagram K1 corresponding to the target frequency is obtained.

[0023] As a further scheme of the present application: the specific way of obtaining the value area is:

[0024] A two-dimensional coordinate system is established by taking the calculation value as the Y axis and the data acquisition time as the X axis, and the maximum calculation value J and its preset detection time T are marked at the scale corresponding to the Y axis and the X axis in the two-dimensional coordinate system. Draw a horizontal line L1 parallel to the X axis with the vertical coordinate as the maximum value J, and draw a vertical line L2 parallel to the Y axis with the horizontal coordinate as the preset detection time T. The closed area between the horizontal line L1, the vertical line L2 and the Y axis and the X axis is marked as the value area.

[0025] As a further scheme of the present application: the specific way of obtaining the data missing time is:

[0026] In the real-time operation data corresponding to each type of data of the industrial equipment at different real-time data acquisition time, if one type of data or multiple types of data are missing at the real-time data acquisition time, mark it as the data missing time, and mark it as the data missing time.

[0027] As a further scheme of the present application: the specific way of obtaining the missing value corresponding to the missing time is:

[0028] Obtain the interval reference diagram corresponding to the frequency interval corresponding to the real-time acquisition frequency, draw a calibration line H perpendicular to the X axis with the missing time corresponding to the numerical value on the X axis as the end point;

[0029] When the calibration line H passes through the area block, the distribution area block belonging to the corresponding interval reference diagram is obtained from each area block passed through by the calibration line H, and is marked as a calibration area block, the longitudinal coordinate of the center point corresponding to each calibration area block is obtained, and the average of each longitudinal coordinate is taken as the missing value corresponding to the missing time.

[0030] As a further scheme of the application: when the calibration line H coincides with the vertical edge line of the area block, the distribution area block belonging to the corresponding interval reference diagram located on both sides of the calibration line H is marked as a calibration area block, and the longitudinal coordinate of the center point corresponding to each calibration area block is obtained, and the average of each longitudinal coordinate is taken as the missing value corresponding to the missing time.

[0031] As a further scheme of the application: the specific way of obtaining the reference value corresponding to the target time of different kinds of data is:

[0032] The average of the maximum value and the minimum value of each kind of data in the running data Aji corresponding to the target time 1t1 of the different kinds of data set by each industrial equipment is taken as the reference value Di corresponding to the target time 1t1 of the different kinds of data.

[0033] As a further scheme of the application: the specific way of obtaining the calculation value corresponding to the target time of the industrial equipment is:

[0034] The sum of the products between the reference value Di corresponding to the target time 1t1 of the different kinds of data and the corresponding preset coefficient βi is taken as the calculation value F11 corresponding to the target time 1t1 of the industrial equipment, where 1=β1+β2+……+βa1.

[0035] Compared with the prior art, the application has the following beneficial effects:

[0036] (1) The application can more accurately reflect the running state of the equipment at each time by obtaining the calculation value corresponding to each data acquisition time of the experimental equipment group in different detection data acquisition frequency ranges;

[0037] (2) The application directly displays the distribution characteristics of the equipment running data in different frequency intervals by obtaining the interval reference diagram corresponding to each frequency interval, and by comparing different interval reference diagrams, the influence of different data acquisition frequency ranges on the distribution of the equipment running data can be understood, thereby providing a reference basis for subsequent real-time monitoring of the data missing problem;

[0038] (3) The present application, by monitoring the real-time operation data of industrial equipment, marking the data acquisition time point with data missing, then finding the corresponding frequency interval interval reference map according to the real-time acquisition frequency, drawing the calibration line H with the value of the missing time point on the X axis as the end point, selecting the distribution area block belonging to the corresponding interval reference map by judging the position relationship between the calibration line H and the area block, and calculating the mean value of the longitudinal coordinates of the center points of these distribution area blocks as the missing value corresponding to the missing time point, can accurately estimate the missing value corresponding to the missing time point, so as to ensure the integrity of the real-time operation data, and ensure the integrity of the real-time operation data. BRIEF DESCRIPTION OF DRAWINGS

[0039] Figure 1 The method framework structure of the present application is shown in the figure;

[0040] Figure 2 The two-dimensional coordinate model structure of the present application is shown in the figure;

[0041] Figure 3 The framework structure of the detection value distribution map K1 of the present application is shown in the figure;

[0042] Figure 4 The framework structure of the detection value distribution map K2 of the present application is shown in the figure;

[0043] Figure 5 The framework structure of the interval reference map Q1 corresponding to the analysis interval of the present application is shown in the figure. DETAILED DESCRIPTION

[0044] The technical solutions of the present application will be described in detail below with reference to the embodiments, obviously, the described embodiments are only a part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0045] Example one: please refer to Figure 1 The present application provides an industrial equipment operation and maintenance intelligent monitoring method based on multi-dimensional data fusion, comprising the following steps:

[0046] Step one: set multiple industrial equipment as experimental equipment group, set multiple detection data acquisition frequencies Cc within the preset detection time T, take the ratio of each detection data acquisition frequency Cc to the preset detection time T as different detection data acquisition frequency Vc, place the experimental equipment group in the same preset detection time T multiple times and acquire the running data of each type of data corresponding to different industrial equipment in the experimental equipment group at different detection data acquisition frequencies each time, and then obtain the multi-dimensional running data corresponding to different detection data acquisition frequencies of the experimental equipment group respectively.

[0047] wherein c represents different detection data acquisition frequencies, c = 1, 2, …, r, r represents the total number of detection data acquisition frequencies, r is a positive integer, and r > 6, the specific value of the detection data acquisition frequency Cc is determined by relevant personnel according to actual needs, each detection data acquisition frequency Cc is greater than 10, the detection data acquisition frequency Vc = detection data acquisition frequency Cc / preset detection time T, the specific value of the preset detection time T is determined by relevant personnel according to actual needs, T is greater than or equal to 120 minutes;

[0048] By setting multiple different detection data acquisition frequencies and corresponding detection data acquisition frequencies, multiple data acquisition is performed at different frequencies, which is to comprehensively explore the characteristics of equipment operation data under different frequencies, so as to provide a rich data basis for subsequent analysis, and to ensure that the data required for subsequent analysis is comprehensive, not only containing data under a single frequency, but also covering the operation characteristics of the equipment under multiple sampling conditions. Through subsequent analysis of these data, the performance of the equipment under different sampling conditions can be more comprehensively understood, which provides a basis for determining the appropriate data acquisition frequency and avoids missing or redundant data information due to improper acquisition frequency.

[0049] Step two: analyzing the multi-dimensional operation data of the experimental equipment group under different detection data acquisition frequencies to obtain the calculation values of the experimental equipment group under different data acquisition times at different detection data acquisition frequencies, the specific method being:

[0050] S1: randomly selecting one of the different detection data acquisition frequencies Cc as the target frequency C1; obtaining the operation data of each type of data of each industrial equipment in the experimental equipment group corresponding to the target frequency C1, taking the ratio between the preset detection time T and the target frequency C1 as the time interval t1 corresponding to the target frequency, and then obtaining the different data acquisition times et1 corresponding to the detection data acquisition frequency of the target frequency, and simultaneously obtaining the multi-dimensional operation data of each industrial equipment in the experimental equipment corresponding to each data acquisition time et1, wherein e represents different data acquisition times, e = 1, 2, …, n1, n1 represents the total number of target frequencies, n1 is a positive integer, and n1 > 6;

[0051] S2: randomly obtaining one of the data acquisition times et1 as the target time 1t1, and marking the operation data of each type of data of each industrial equipment corresponding to the target time 1t1 as Aji, wherein i is the total number of different types of data in the multi-dimensional operation data, j represents different industrial equipment in the experimental equipment group, i = 1, 2, …, a1, a1 represents the total number of different types of data, a1 is a positive integer, and a1 > 2;

[0052] The average of the maximum value and the minimum value of each type of data in the running data Aji corresponding to the different types of data set by each industrial equipment at the target time 1t1 is taken as the reference value Di corresponding to each type of data at the target time 1t1. The average of the maximum value and the minimum value of a type of data at a time is taken to eliminate the interference of extreme values and reflect the concentration trend of the data. The sum of the products of the reference value Di corresponding to each type of data at the target time 1t1 and the corresponding preset coefficient βi is taken as the calculation value F11 of the industrial equipment at the target time 1t1. The specific value of the preset coefficient βi is determined by relevant personnel according to actual needs, and satisfies 1 = β1 + β2 + … + βa1. The preset coefficient βi corresponds to each type of data one by one.

[0053] S3: Repeat step S2 to analyze the running data corresponding to the different types of data set by each industrial equipment at the remaining data acquisition times, and obtain the calculation value Fe1 corresponding to each data acquisition time et1 at the target frequency.

[0054] S4: Repeat steps S1-S3 to analyze the running data corresponding to each type of data of different industrial equipment in the experimental equipment group at the remaining detection data acquisition frequencies, and obtain the reference value corresponding to each data acquisition time at different data acquisition frequencies within each detection data acquisition frequency.

[0055] The calculation value corresponding to each data acquisition time at different detection data acquisition frequencies of the experimental equipment group is obtained. The importance of different types of data and the overall level of data at this time are considered by the calculation value, which can more accurately reflect the running state of the equipment at each time, and can understand the influence of different data acquisition frequencies on the description of the running state of the equipment, providing support for subsequent determination of appropriate data acquisition strategy.

[0056] Step three: Sort each detection data acquisition frequency according to the corresponding value from small to large, then sort each adjacent two detection data acquisition frequencies according to the order from front to back to form a frequency interval, and then obtain a plurality of frequency intervals. According to the reference value corresponding to each data acquisition time at different data acquisition frequencies within each detection data acquisition frequency, obtain the detection value distribution diagram corresponding to each detection data acquisition frequency, and according to the detection value distribution diagram corresponding to each detection data acquisition frequency, analyze the detection value distribution diagrams corresponding to the two detection data acquisition frequencies forming different frequency intervals to obtain the interval reference diagram corresponding to different frequency intervals. The specific method is:

[0057] Obtaining the maximum value J in the calculation value corresponding to each data acquisition time of the experimental equipment group in different data acquisition frequencies, taking the calculation value as the ordinate, i.e. the Y axis, and the data acquisition time as the abscissa, i.e. the X axis, to establish a two-dimensional coordinate system, and marking the maximum calculation value J and its preset detection time T on the scale corresponding to the Y axis and the X axis in the two-dimensional coordinate system, drawing a horizontal line L1 parallel to the X axis with the ordinate as the maximum value J, and drawing a vertical line L2 parallel to the Y axis with the abscissa as the preset detection time T, marking the closed area between the horizontal line L1, the vertical line L2 and the Y axis and the X axis as a value area, and uniformly dividing the value area into a plurality of equal area blocks to generate a two-dimensional coordinate model;

[0058] S01: Obtain the frequency interval with the lower limit value of the target frequency as the analysis interval from the plurality of frequency intervals, obtain the calculation value Fe1 corresponding to each data acquisition time et1 of the target frequency, mark the corresponding data points in the two-dimensional coordinate model according to the calculation value Fe1 corresponding to each data acquisition time et1 of the target frequency, and mark the distribution area block corresponding to the target frequency according to the position of each data point in each area block in the two-dimensional coordinate model, and obtain the detection value distribution diagram corresponding to the target frequency.

[0059] The specific way of marking the distribution area block corresponding to the target frequency according to the position of each data point in each area block in the two-dimensional coordinate model is:

[0060] For the case that the data point is located in the area block, the corresponding area block is taken as the distribution area block, for the case that the data point is located on the edge line of the area block, when the data point is located on the horizontal edge line of the area block, the area blocks on the upper and lower sides of the corresponding horizontal edge line are taken as the distribution area blocks; when the data point is located on the vertical edge line of the area block, the area blocks on the left and right sides of the corresponding vertical edge line are taken as the distribution area blocks; for the case that the data point is located at the boundary node of the area block, the area blocks corresponding to the upper, lower, left and right of the boundary node of the area block are taken as the distribution area blocks, and then the distribution area block corresponding to the target frequency is obtained and marked, and the detection value distribution diagram K1 corresponding to the target frequency is obtained.

[0061] Simultaneously, the detection data acquisition frequency corresponding to the upper limit of the analysis interval is obtained from the analysis interval. The ratio between the preset detection duration T and the detection data acquisition frequency corresponding to the upper limit of the interval is taken as the corresponding time interval t2. Then, the different data acquisition times ct2 under the detection data acquisition frequency corresponding to the upper limit of the interval are obtained. The calculated values ​​at the different data acquisition times ct2 under the detection data acquisition frequency corresponding to the upper limit of the interval are marked as Dc, where c represents different data acquisition times, c=1, 2, ..., n2, and n2 represents the total number of different data acquisition times under the detection data acquisition frequency corresponding to the upper limit of the interval. n2 is a positive integer and n2≧6;

[0062] Based on the calculated values ​​at ct2 at different data acquisition times, Dc is used to mark the corresponding data points in the two-dimensional coordinate model. In the same way as obtaining the detection value distribution map K1 corresponding to the target frequency, the detection value distribution map K2 corresponding to the upper limit of the interval is obtained. The distribution area blocks corresponding to the detection value distribution maps K1 and K2 are merged to obtain the interval reference Q1 corresponding to the analysis interval.

[0063] S02: Repeat step S01 to perform the same analysis and processing on the remaining frequency intervals, and then obtain the interval reference Qq corresponding to each frequency interval, where q refers to different frequency intervals, q=1, 2, ..., r-1;

[0064] The detection data acquisition frequency is sorted and formed into frequency intervals. Based on the baseline values ​​corresponding to different data acquisition times within each detection data acquisition frequency, a detection value distribution map is generated. By forming frequency intervals from two adjacent detection data acquisition frequencies, different data acquisition frequencies can be grouped, facilitating the analysis of data characteristics within different frequency ranges. By establishing a two-dimensional coordinate model, with the calculated maximum value J as the vertical axis and the data acquisition time as the horizontal axis, data points are marked in the coordinate system. Based on the situation of the region block where the data point is located, the distribution region block and interval reference are determined. The distribution of data under different frequencies is displayed in a visual way. The region blocks are used to classify and summarize the data, thereby more intuitively analyzing the characteristic distribution patterns of data within different frequency intervals.

[0065] By obtaining interval references corresponding to each frequency interval, the distribution characteristics of equipment operation data within different frequency intervals can be intuitively displayed. By comparing different interval references, the impact of different data acquisition frequency ranges on the distribution of equipment operation data can be understood, providing a reference for handling data missing issues in subsequent real-time monitoring and enabling more accurate estimation of the missing values ​​at the time of data loss.

[0066] Step four: according to the preset detection duration T, the real-time operation and maintenance data of the industrial equipment is obtained for z times, the ratio between z and T is taken as the real-time acquisition frequency of the industrial equipment operation and maintenance data, the ratio between the preset detection duration T and the target frequency z is taken as the time interval t3 corresponding to the target frequency, and then the real-time data acquisition time ut3 of the industrial equipment is obtained, u represents different real-time data acquisition times, the real-time operation and maintenance data of the industrial equipment at different real-time data acquisition times is analyzed, the data missing time is obtained, the missing value corresponding to the missing time is obtained according to the interval reference diagram of the frequency interval corresponding to the real-time acquisition frequency, and the specific mode is as follows:

[0067] The specific mode of obtaining the data missing time is as follows:

[0068] Among the real-time operation and maintenance data of the industrial equipment at different real-time data acquisition times, for the real-time data acquisition time at which one or more types of data are missing, it is marked as a data missing time;

[0069] The specific mode of obtaining the missing value corresponding to the missing time is as follows:

[0070] The interval reference diagram of the frequency interval corresponding to the real-time acquisition frequency is obtained, and the calibration line H perpendicular to the X-axis is drawn with the value corresponding to the missing time on the X-axis as the end point;

[0071] When the calibration line H passes through the region block, the distribution region block belonging to the corresponding interval reference diagram is obtained from each region block passed by the calibration line H, and it is marked as a calibration region block, the vertical coordinates of the center points of each calibration region block are obtained, and the average of each vertical coordinate is taken as the missing value corresponding to the missing time;

[0072] When the calibration line H coincides with the vertical edge line of the region block, the distribution region block belonging to the corresponding interval reference diagram located on both sides of the calibration line H is marked as a calibration region block, and the vertical coordinates of the center points of each calibration region block are obtained, and the average of each vertical coordinate is taken as the missing value corresponding to the missing time;

[0073] In the same way, the missing value corresponding to each missing time is calculated and obtained in turn;

[0074] The real-time operation data of the industrial equipment is monitored, and the data acquisition time points with data missing are marked. Then, according to the real-time acquisition frequency, the interval reference graph of the corresponding frequency interval is found, the calibration line H is drawn with the numerical value of the missing time point on the X-axis as the end point, the distribution area block belonging to the corresponding interval reference graph is selected by judging the positional relationship between the calibration line H and the area block, and the vertical coordinate mean value of the center point of the distribution area block is calculated as the missing numerical value corresponding to the missing time point. The data missing time point in the real-time operation data of the industrial equipment can be accurately identified, and the missing numerical value corresponding to the missing time point can be estimated, so as to ensure the integrity of the real-time operation data. This is helpful for continuously and accurately monitoring and analyzing the running state of the industrial equipment, discovering possible problems of the equipment in time, improving the reliability and effectiveness of the operation and maintenance of the industrial equipment, and reasonably estimating the missing data by using the data distribution characteristics represented by different area blocks in the existing interval reference graph, so as to ensure the integrity of the real-time operation data.

[0075] The above formulas are dimensionless values calculated, and the formulas are obtained by software simulation of a large amount of data to obtain a formula of the nearest real situation. The preset parameters and threshold values in the formula are set by a person skilled in the art according to the actual situation.

[0076] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for intelligent monitoring of industrial equipment operation and maintenance based on multi-dimensional data fusion, characterized in that, Includes the following steps: Step 1: Set up multiple industrial devices as an experimental equipment group, set multiple detection data acquisition frequencies within a preset detection duration T, and obtain multi-dimensional operating data of the experimental equipment group under different detection data acquisition frequencies; Step Two: Analyze the multidimensional operational data of the experimental equipment group at different data acquisition frequencies to obtain the calculated values ​​corresponding to different data acquisition times within different data acquisition frequencies. The specific method is as follows: S1: The ratio between each detection data acquisition frequency Cc and the preset detection duration T is used as the different detection data acquisition frequencies Vc, where c represents different detection data acquisition frequencies, c=1, 2, ..., r, r represents the total number of detection data acquisition frequencies, r is a positive integer and r≧6. One of the different detection data acquisition frequencies Cc is randomly selected as the target frequency C1. The operating data corresponding to various types of data from different industrial equipment in the experimental equipment group at the target frequency C1 is obtained. The ratio between the preset detection duration T and the target frequency C1 is used as the time interval t1 corresponding to the target frequency. Different data acquisition times et1 are obtained at the detection data acquisition frequency corresponding to the target frequency. Simultaneously, the multidimensional operating data corresponding to each industrial equipment in the experimental equipment at each data acquisition time et1 is obtained, where e represents different data acquisition times, e=1, 2, ..., n1, n1 represents the total number of target frequencies, n1 is a positive integer and n1≧6. S2: Randomly select one of the data acquisition times et1 as the target time 1t1, and mark the operation data corresponding to the different types of data set by each industrial equipment at the target time 1t1 as Aji, where i is the different types of data in the multidimensional operation data, j refers to the different industrial equipment in the experimental equipment group, i=1, 2, ..., a1, a1 refers to the total number of data types in the different types of data, a1 is a positive integer, and a1≧2; Based on the different types of data set for each industrial equipment, the corresponding operating data Aji at the target time 1t1 is obtained. The reference values ​​Di corresponding to the different types of data at the target time 1t1 are calculated to obtain the calculated value F11 corresponding to the industrial equipment at the target time 1t1. S3: Repeat step S2 to analyze the operating data corresponding to the different types of data set for each industrial equipment at the remaining data acquisition time, so as to obtain the calculated value Fe1 corresponding to each data acquisition time et1 under the target frequency. S4: Repeat steps S1-S3 to analyze the operating data corresponding to various types of data of different industrial equipment in the experimental equipment group at the remaining detection data acquisition frequency. This will give you the benchmark values ​​corresponding to different data acquisition times within each detection data acquisition frequency. Step 3: Divide each detection data into multiple frequency intervals in order from front to back. Based on the baseline values ​​corresponding to different data acquisition times within each detection data acquisition frequency, obtain the detection value distribution map corresponding to each detection data acquisition frequency. Based on the detection value distribution map corresponding to the acquisition frequencies of two detection data that make up different frequency intervals, obtain the interval reference map corresponding to each frequency interval. Step 4: Acquire real-time operation and maintenance data and real-time acquisition frequency of various types of industrial equipment data according to the preset detection duration T. At the same time, obtain the real-time data acquisition time corresponding to the industrial equipment. Obtain the missing data time according to the real-time operation and maintenance data corresponding to different real-time data acquisition times of various types of industrial equipment data. Obtain the missing value corresponding to the missing time according to the interval reference of the frequency interval corresponding to the real-time acquisition frequency.

2. The intelligent monitoring method for industrial equipment operation and maintenance based on multi-dimensional data fusion according to claim 1, characterized in that, The specific method for obtaining the interval references corresponding to different frequency intervals is as follows: The maximum value J of the calculated values ​​corresponding to different data acquisition times within different detection data acquisition frequencies of the experimental equipment group is obtained. The numerical region is obtained based on the maximum value J and the preset detection duration T. The numerical region is evenly divided into multiple regions of equal size, and then a two-dimensional coordinate model is generated. S01: From multiple frequency intervals, obtain the frequency interval where the target frequency is the lower limit of the interval as the analysis interval. Obtain the calculated value Fe1 corresponding to each data acquisition time et1 at the target frequency. Based on the calculated value Fe1 corresponding to each data acquisition time et1 at the target frequency, mark the corresponding data points in the two-dimensional coordinate model. According to the position of each data point in each region block of the two-dimensional coordinate model, mark the distribution region block corresponding to the target frequency, thereby obtaining the detection value distribution map corresponding to the target frequency. Simultaneously, obtain the detection value corresponding to the upper limit of the analysis interval from the analysis interval. The data acquisition frequency is measured by taking the ratio between the preset detection duration T and the data acquisition frequency corresponding to the upper limit of the interval as the corresponding time interval t2. This yields different data acquisition times ct2 at the data acquisition frequency corresponding to the upper limit of the interval. The calculated values ​​at these different data acquisition times ct2 are labeled Dc, where c represents different data acquisition times, c = 1, 2, ..., n2, and n2 represents the total number of different data acquisition times at the data acquisition frequency corresponding to the upper limit of the interval. n2 is a positive integer and n2 ≥ 6. Based on the calculated values ​​at ct2 at different data acquisition times, Dc is used to mark the corresponding data points in the two-dimensional coordinate model. In the same way as obtaining the detection value distribution map K1 corresponding to the target frequency, the detection value distribution map K2 corresponding to the upper limit of the interval is obtained. The distribution area blocks corresponding to the detection value distribution maps K1 and K2 are merged to obtain the interval reference Q1 corresponding to the analysis interval. S03: Repeat steps S01-S02 to perform the same analysis and processing on the remaining frequency intervals, thereby obtaining the interval reference Qq corresponding to each frequency interval, where q represents different frequency intervals, q=1, 2, ..., r-1.

3. The intelligent monitoring method for industrial equipment operation and maintenance based on multi-dimensional data fusion according to claim 2, characterized in that, The specific method for marking the distribution area blocks corresponding to the target frequency is as follows: When a data point is located within a region block, the corresponding region block is used as a distribution region block. When a data point is located on the horizontal edge of a region block, the region blocks located above and below the corresponding horizontal edge are used as distribution region blocks. When a data point is located on the vertical edge of a region block, the region blocks located to the left and right of the corresponding vertical edge are used as distribution region blocks. When a data point is located at a boundary node of a region block, the region blocks located above, below, to the left and right of the boundary node are used as distribution region blocks. This allows us to obtain the distribution region blocks corresponding to the target frequency and mark them, thereby obtaining the detection value distribution map K1 corresponding to the target frequency.

4. The intelligent monitoring method for industrial equipment operation and maintenance based on multi-dimensional data fusion according to claim 2, characterized in that, The specific method for obtaining the numerical range is as follows: A two-dimensional coordinate system is established with the calculated value as the Y-axis and the data acquisition time as the X-axis. The maximum calculated value J and its preset detection duration T are marked at the corresponding scales of the Y-axis and X-axis in the two-dimensional coordinate system. A horizontal line L1 parallel to the X-axis is drawn with the vertical coordinate as the maximum value J, and a vertical line L2 parallel to the Y-axis is drawn with the horizontal coordinate as the preset detection duration T. The closed area between the horizontal line L1, the vertical line L2 and the Y-axis and X-axis is marked as the numerical region.

5. The intelligent monitoring method for industrial equipment operation and maintenance based on multi-dimensional data fusion according to claim 3, characterized in that, The specific method for obtaining the moment when data is missing is as follows: If any type of data or multiple types of data are missing in the real-time operation and maintenance data corresponding to different real-time data acquisition times of industrial equipment, then that real-time data acquisition time is marked as a data missing time.

6. The intelligent monitoring method for industrial equipment operation and maintenance based on multi-dimensional data fusion according to claim 5, characterized in that, The specific method for obtaining the missing values ​​corresponding to the missing time points is as follows: Obtain a reference image of the frequency interval corresponding to the real-time acquisition frequency, and draw a calibration line H perpendicular to the X-axis with the value corresponding to the missing time on the X-axis as the endpoint; When the calibration line H passes through a region block, the distribution region block belonging to the corresponding interval reference is obtained from each region block passed by the calibration line H, and it is marked as the calibration region block. The ordinate of the center point corresponding to each calibration region block is obtained, and the mean of each ordinate is used as the missing value corresponding to the missing time.

7. The intelligent monitoring method for industrial equipment operation and maintenance based on multi-dimensional data fusion according to claim 6, characterized in that, When the calibration line H coincides with the vertical edge of the region block, the distribution region blocks located on both sides of the calibration line H that belong to the corresponding interval reference are marked as calibration region blocks. Similarly, the ordinate of the center point corresponding to each calibration region block is obtained, and the mean of each ordinate is used as the missing value corresponding to the missing time.

8. The intelligent monitoring method for industrial equipment operation and maintenance based on multi-dimensional data fusion according to claim 1, characterized in that, The specific method for obtaining the baseline values ​​corresponding to different types of data at the target time is as follows: The average of the maximum and minimum values ​​of each type of data in the operating data Aji corresponding to the different types of data set for each industrial equipment at the target time 1t1 is used as the baseline value Di corresponding to each type of data at the target time 1t1.

9. The intelligent monitoring method for industrial equipment operation and maintenance based on multi-dimensional data fusion according to claim 8, characterized in that, The specific method for obtaining the calculated values ​​of industrial equipment at the target time is as follows: The sum of the products of the reference values ​​Di corresponding to different types of data at the target time 1t1 and their corresponding preset coefficients βi is taken as the calculated value F11 of the industrial equipment at the target time 1t1, where 1=β1+β2+……+βa1.

Images