Big data analysis method and system for expressing unit state

Abstract

The invention discloses a big data analysis method and system for expressing a unit state. The method comprises the steps of obtaining time sequence data of a historical first performance index and a historical second performance index; carrying out data screening on the time series data after the stepped processing, and screening out the most stable data segment; based on the most stable data segment, processing the screened data, and establishing a relation model; on the basis of the relation model, the correlation degree between the performance indexes is calculated by calculating the difference of bit sequence numbers; the key performance indexes of the health state of the specific part of the unit are identified by calculating the correlation degree and judging the correlation strength; and based on the relevance between the key performance index and the performance index, evaluating the health state of the specific part of the unit. According to the method, the change of the state of the unit is reflected in real time through a unique data processing and correlation degree calculation method, and the dynamic property and accuracy of evaluation are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of unit health status monitoring and evaluation, and particularly to a big data analysis method and system for expressing the state of a unit. Background Art

[0002] In the field of giant unit health status monitoring and evaluation, the existing technologies mainly rely on traditional fault monitoring methods and experience-based maintenance strategies. These methods generally include the following steps:

[0003] Data collection: Collect various performance index data during the operation of the unit, such as vibration, temperature, pressure, etc.

[0004] Fault identification: Identify possible faults or abnormal conditions by setting fixed thresholds or rule-based methods.

[0005] Status evaluation: Evaluate the overall or component health status of the unit according to the results of fault identification.

[0006] Maintenance decision-making: Based on the results of status evaluation, formulate a maintenance plan or take corresponding maintenance measures.

[0007] However, the existing technologies have the following limitations:

[0008] Lack of dynamics: Traditional methods are usually based on fixed thresholds or rules and cannot reflect the changes in the unit state in real time.

[0009] Insufficient correlation: The existing technologies often ignore the correlation between different performance indexes, resulting in incomplete and inaccurate evaluation results.

[0010] Reliance on experience: Many methods rely on the experience and intuition of maintenance personnel, lacking objectivity and accuracy. Summary of the Invention

[0011] In view of the above existing problems, the present invention is proposed.

[0012] Therefore, the present invention provides a big data analysis method and system for expressing the state of a unit to solve the limitations existing in the existing technologies in the health status monitoring and evaluation of giant units, especially the problems of lack of dynamics and insufficient correlation.

[0013] To solve the above technical problems, the present invention provides the following technical solutions:

[0014] In a first aspect, the present invention provides a big data analysis method for expressing the state of a unit, including:

[0015] Obtain the time series data of historical first performance indexes and second performance indexes;

[0016] Filter the time series data after stepped processing to select the most stable data segment;

[0017] Based on the most stable data segment, process the filtered data to establish a relationship model;

[0018] Based on the relationship model, calculate the correlation degree between performance indicators by calculating the difference in rank numbers;

[0019] Identify the key performance indicators of the health status of specific components of the unit by calculating the correlation degree and judging the correlation strength;

[0020] Evaluate the health status of specific components of the unit based on the key performance indicators and the correlation between performance indicators.

[0021] As a preferred solution of the big data analysis method for expressing the unit state described in the present invention, wherein:

[0022] The data after stepped processing includes the following steps:

[0023] Perform stepped processing on the respective time series data of the first performance indicator X and the second performance indicator Y to obtain the time series data to be processed;

[0024] Uniformly divide N continuous but non-overlapping steps between the maximum and minimum values of the first performance indicator X and the second performance indicator Y;

[0025] Replace all the time series data to be processed with the mean value within the step where it is located to form step values.

[0026] As a preferred solution of the big data analysis method for expressing the unit state described in the present invention, wherein:

[0027] The data filtering includes the following steps:

[0028] By setting the first threshold K, identify the time period T(s) during which the second performance indicator is in a continuous stable state;

[0029] If the change of the second performance indicator Y value within a time period is less than K, it is considered that this time period is stable;

[0030] Based on the identified stable time period T(s), extract all the sample values of the first performance indicator X therefrom to obtain a set of stable sample values of the first performance indicator X.

[0031] As a preferred solution of the big data analysis method for expressing the unit state described in the present invention, wherein:

[0032] The processing of the filtered data includes the following steps:

[0033] Pair the step values of each first performance metric X within the period T(s) with the corresponding step values of the second performance metric Y to form the original pairing array C(x, y).

[0034] Classify the data according to the step values of the second performance metric Y to form the first performance metric X sample sets C(X y ) corresponding to different second performance metric Y values.

[0035] For each C(X y ) use the weighted mean to calculate the expected value XP of the first performance metric X y , and establish the correspondence relationship model M(XP y , Y) between the second performance metric Y and the expected value of the first performance metric X.

[0036] As a preferred solution of the big data analysis method for expressing the unit state described in the present invention, wherein:

[0037] Calculating the correlation degree between the performance metrics includes the following steps:

[0038] Sort the expected values XP of the first performance metric X y from small to large, and assign the rank serial number Sxpy, starting from 1 and ending at N;

[0039] Sort the second performance metric Y values from small to large, and assign the rank serial number Sy, starting from 1 and ending at N;

[0040] Form a new rank serial number array M(Sxpy, Sy) by arranging the rank serial numbers of the expected value of the first performance metric X and the second performance metric Y according to the original corresponding relationship between the expected value of the first performance metric X and the second performance metric Y;

[0041] Calculate the correlation degree δ(X, Y) between the first performance metric X and the second performance metric Y.

[0042] As a preferred solution of the big data analysis method for expressing the unit state described in the present invention, wherein:

[0043] The correlation degree δ(X, Y) between the first performance metric X and the second performance metric Y is expressed as:

[0044]

[0045] wherein, S i =(Sxpy - Sy), and the range of i is from 1 to N.

[0046] As a preferred solution of the big data analysis method for expressing the unit state described in the present invention, wherein:

[0047] The determination of the association strength includes determining whether there is a strong association between the first performance index X and the second performance index Y based on the absolute value, mean value, and sign of δ(X, Y).

[0048] If the absolute value of δ(X, Y)>TH1, the mean value of the absolute value of δ(X, Y)>TH2, and the signs of δ(X, Y) are the same are satisfied simultaneously, where TH1 and TH2 are not less than 0.7, it is determined that there is a strong association between the first performance index X and the second performance index Y.

[0049] On the second aspect, the present invention provides a big data analysis system for expressing the state of a unit, including:

[0050] An acquisition module, configured to acquire the time series data of the historical first performance index and the second performance index;

[0051] A preprocessing module, configured to perform data screening on the time series data after stepped processing to screen out the most stable data segment;

[0052] A data screening module, configured to process the screened data based on the most stable data segment to establish a relationship model;

[0053] An association degree calculation module, configured to calculate the association degree between performance indexes based on the relationship model by calculating the difference in rank numbers;

[0054] An association strength judgment module, configured to identify the key performance indexes of the health state of specific components of the unit by calculating the association degree and judging the association strength;

[0055] A health state evaluation module, configured to evaluate the health state of specific components of the unit based on the key performance indexes and the association between performance indexes.

[0056] On the third aspect, the present invention provides a computing device, including:

[0057] A memory, configured to store a program;

[0058] A processor, configured to execute the computer-executable instructions, and when the computer-executable instructions are executed by the processor, the steps of the big data analysis method for expressing the state of the unit are implemented.

[0059] On the fourth aspect, the present invention provides a computer-readable storage medium, including: when the program is executed by the processor, the steps of the big data analysis method for expressing the state of the unit are implemented.

[0060] Advantages of the present invention: By means of methods such as time series analysis and calculation of rank correspondence relationships, the present invention can reflect the changes in the unit state in real time, improving the dynamics and accuracy of the evaluation. By calculating the correlation degrees between different performance indicators, the internal relationships between them are revealed, making the evaluation results more comprehensive and accurate. Based on the key performance indicators and their correlations, more precise and scientific maintenance plans or repair measures can be formulated, improving the maintenance efficiency and the reliability of the unit operation. The method proposed in this patent has clear implementation steps and parameter value-taking methods, making it easier to operate and implement in practical applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings. Among them:

[0062] Figure 1 FIG. is a schematic diagram of the basic process of a big data analysis method for expressing the unit state provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0063] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following will provide a detailed description of the specific embodiments of the present invention with reference to the accompanying drawings of the specification. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope of protection of the present invention.

[0064] Embodiment 1

[0065] Referring to Figure 1 , an embodiment of the present invention provides a big data analysis method for expressing the unit state, including:

[0066] S1: Obtain the time series data of the historical first performance indicator and the second performance indicator;

[0067] In the embodiments of the present application, collect the time series data of the historical first performance indicator X (hereinafter referred to as the secondary variable) and the second performance indicator Y (hereinafter referred to as the main variable). The sample size of these data should be large enough;

[0068] In the embodiments of the present application, the determination of a large enough sample is: the number of samples in each main variable step in S2 should be not less than the first threshold L, and L is an integer not less than 10;

[0069] S2: Screen the time series data after stepwise processing to select the most stable data segment;

[0070] In the embodiment of the present application, stepwise processing is performed on the respective time series data of the variable X and the main variable Y: N continuous but non-overlapping steps are evenly divided between the maximum and minimum values of X (or Y), and all sample values are replaced with the mean value within the step where they are located. Hereinafter, this step mean value is simply referred to as the step value of the variable;

[0071] S3: Based on the most stable data segment, process the screened data to establish a relationship model;

[0072] In the embodiment of the present application, establishing a relationship model includes the following steps:

[0073] Find out the time period T(s) during which the main variable Y is in a continuous stable state;

[0074] Extract all samples of the slave variable X within the time period T(S);

[0075] Within the range of T(S), according to the principle of time consistency, each step value of X is paired with a step value of Y to form an original pairing array C(x, y);

[0076] In C(x, y), classify according to the value of the main variable Y, and group all samples of X that are equally associated with the same Y step value into one category, thus forming a sample set C(X y ) of X corresponding to different Y step values, which is a one-dimensional array;

[0077] For each C(X y ) use the weighted mean to find the expected value XP y of the sample, y and establish a correspondence relationship model M(XP y , Y) between the second performance index Y and the expected value of the first performance index X.

[0078] In the embodiment of the present application, the condition for finding that the main variable Y is in a continuous stable stage is: when the ratio of the absolute value of the difference between the Y value at a moment and the Y values at all any moments within the previous time period M to the Y value at this moment is less than the second threshold K; and L * K = 1.

[0079] S4: Based on the relationship model, calculate the correlation degree between performance indicators by calculating the difference in rank numbers;

[0080] In the embodiment of the present application, calculate the correlation degree δ(X, Y) between X and Y according to the following steps:

[0081] (1) Sort the Xpy values in ascending order and assign them the rank serial numbers Sxpy, starting from 1 and ending at N;

[0082] (2) Sort the Y values in ascending order and assign them the rank serial numbers Sy, starting from 1 and ending at N;

[0083] (3) Combine the rank serial numbers of XPy and Y according to the original corresponding relationship between XPy and Y to form a new rank serial number array, denoted as M(Sxpy, Sy). Then, the calculation method of the correlation degree δ(X, Y) between X and Y is as follows:

[0084]

[0085] where S i =(Sxpy - Sy), and the range of i is from 1 to N.

[0086] S5: Identify the key performance indicators of the health status of specific components of the unit by calculating the correlation degree and judging the correlation strength;

[0087] In the embodiment of the present application, based on the steps of data screening, data processing, and calculating the correlation degree, multiple different ladder numbers N (such as N1, N2, N3...) are set, and the correlation degree calculation process is repeatedly executed under each ladder number condition to obtain a series of correlation degree values δ i (X, Y).

[0088] If the following three criteria are simultaneously met, it is determined that there is a significant and stable strong correlation relationship between the variable X and the main variable Y:

[0089] ① The absolute value of δ i is greater than TH1, indicating that the correlation degree calculated once reaches the standard of strong correlation.

[0090] ② The average value of the absolute value of δ i is greater than TH2, ensuring a relatively high overall correlation degree level for multiple calculations.

[0091] ③ The signs of δ i are the same, that is, all δ i (X, Y) are either positively correlated or negatively correlated, ensuring the consistency of the correlation direction.

[0092] where TH1 and TH2 are not less than 0.7.

[0093] In the embodiment of the present application, if the correlation degree is high and stable, it indicates that there is a significant interaction between these two variables, which may indicate certain physical or operational changes within the system.

[0094] S6: Evaluate the health status of specific components of the unit based on the correlation between key performance indicators and performance indicators.

[0095] In the embodiment of the present application, based on the above analysis results and domain knowledge, a comprehensive assessment of the health status of specific components or the entire unit is carried out. For example, if an abnormal change is found in the correlation between certain key performance indicators, it may mean that the component is about to fail and maintenance measures need to be taken in a timely manner.

[0096] This embodiment also provides a big data analysis system for expressing the status of the unit, including:

[0097] An acquisition module for acquiring the time series data of historical first performance indicators and second performance indicators;

[0098] A preprocessing module for screening the time series data after stepped processing to screen out the most stable data segment;

[0099] A data screening module for processing the screened data based on the most stable data segment to establish a relationship model;

[0100] A correlation calculation module for calculating the correlation between performance indicators based on the relationship model by calculating the difference in rank numbers;

[0101] A correlation strength judgment module for identifying the key performance indicators of the health status of specific components of the unit by calculating the correlation and judging the correlation strength;

[0102] A health status assessment module for assessing the health status of specific components of the unit based on the key performance indicators and the correlation between performance indicators.

[0103] Furthermore, it further includes:

[0104] A memory for storing programs;

[0105] A processor for loading the program to execute the big data analysis method for expressing the status of the unit.

[0106] This embodiment also provides a computer-readable storage medium storing a program, which when executed by a processor, implements the big data analysis method for expressing the status of the unit.

[0107] The storage medium proposed in this embodiment and the big data analysis method for expressing the status of the unit proposed in the above embodiment belong to the same inventive concept. Technical details not described in detail in this embodiment can be referred to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0108] Through the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented by means of software and necessary general hardware. Of course, it can also be implemented by hardware, but in many cases the former is a better implementation manner. Based on such an understanding, the technical solution of the present invention, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a floppy disk, read-only memory (ROM), random access memory (RAM), flash memory (FLASH), hard disk or optical disc of a computer, etc., including several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods of the various embodiments of the present invention.

[0109] Embodiment 2

[0110] This is an embodiment of the present invention, which provides a big data analysis method for expressing the state of a unit set. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through specific implementation manners and implementation effects.

[0111] The specific embodiments are as follows:

[0112] There is a giant unit set, and there are two key performance indicators: X (for example, the vibration amplitude of the unit) and Y (for example, the load condition of the unit). It is necessary to evaluate the magnitude of the influence of X by Y, that is, the correlation degree between them.

[0113] Step 1: Data collection

[0114] Collect the time series data of X and Y in history. These data should be large enough to ensure the accuracy of subsequent analysis. Collect one year's data, and each indicator has thousands of data points.

[0115] Step 2: Data laddering processing

[0116] Evenly divide the continuous but non-overlapping N ladders between the maximum and minimum values of X and Y. Divide both X and Y into 10 ladders (N = 10).

[0117] Replace all sample values with the mean value within their respective ladders to form ladder values.

[0118] Step 3: Determine the stable state period of the main variable Y

[0119] Find out the period T(s) when Y is in a continuous stable state. This can be achieved by comparing the changes in Y values at consecutive time points. Set a threshold K. If the change in Y value within a period is less than K, then this period is considered stable.

[0120] Step Four: Extract samples of variable X:

[0121] During the period T(s), extract all sample values of X.

[0122] Step Five: Form a two-dimensional array C(x,y)

[0123] Correspond each step value of X during the period T(s) with a step value of Y to form an original paired array C(x,y).

[0124] Step Six: Classify the sample set C(Xy)

[0125] In C(x,y), classify according to the value of Y, and group all samples of X that are equally associated with the same step value of Y into one category, forming a sample set C(Xy) corresponding to different step values of Y.

[0126] Step Seven: Calculate the expected value XPy

[0127] For each C(Xy), use the weighted mean to calculate the expected value XPy of the samples, obtaining an expected value relationship array M(XPy, Y) that has a one-to-one correspondence between the main variable Y and the expected value of X.

[0128] Step Eight: Calculate the correlation degree δ(X, Y)

[0129] Sort the rank numbers of XPy and Y, and calculate the correlation degree δ(X, Y) between them. This can be achieved by calculating the sum of the squares of the differences in rank numbers.

[0130] Step Nine: Judge the correlation strength

[0131] According to the absolute value, mean, and sign of δ(X, Y), judge whether there is a strong correlation between X and Y. If the absolute value of δ(X, Y) is greater than 0.7 and the signs are the same, it is considered that there is a strong correlation between X and Y.

[0132] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered by the scope of the claims of the present invention.

Claims

1. A big data analysis method for expressing the state of a unit, characterized in that, Including: Obtain the time series data of the historical first performance index and the second performance index; Perform data screening on the time series data after stepped processing to screen out the most stable data segment; Based on the most stable data segment, process the screened data to establish a relationship model; Based on the relationship model, calculate the correlation degree between performance indicators by calculating the difference in ranking numbers; Identify the key performance indicators of the health status of specific components of the unit by calculating the correlation degree and judging the correlation strength; Evaluate the health status of specific components of the unit based on the correlation between the key performance indicators and the performance indicators.

2. The big data analysis method for expressing the unit status according to claim 1, characterized in that: The data after the stepped processing includes the following steps: Perform stepped processing on the respective time series data of the first performance index X and the second performance index Y to obtain the time series data to be processed; Evenly divide N continuous but non-overlapping steps between the maximum and minimum values of the first performance index X and the second performance index Y; Replace all the time series data to be processed with the mean value within the step where it is located to form step values.

3. The big data analysis method for expressing the state of the unit set according to claim 1 or 2, characterized in that: The data screening includes the following steps: By setting the first threshold K, identify the time period T(s) during which the second performance index is in a continuous stable state; If the change of the second performance index Y value within a time period is less than K, then this time period is considered stable; Based on the identified stable period T(s), extract all the sample values of the first performance index X from it to obtain a set of stable first performance index X sample values.

4. The big data analysis method for expressing the unit state according to claim 3, wherein: The processing of the screened data includes the following steps: Pair the step value of each first performance index X within the T(s) period with the corresponding step value of the second performance index Y to form the original paired array C(x,y); Classify the data according to the step values of the second performance index Y to form a set C(X y ) of first performance index X samples corresponding to different second performance index Y values; For each C(X y ) calculate the expected value XP of the first performance metric X using the weighted mean y , and establish a correspondence relationship model M(XP y , Y) between the second performance metric Y and the expected value of the first performance metric X 5. The big data analysis method for expressing the unit status according to claim 4, characterized in that: The calculation of the correlation degree between performance indicators includes the following steps: Sort the expected value XP of the first performance metric X y in ascending order, and assign a rank number Sxpy, starting from 1 and ending at N; Sort the second performance index Y values from small to large and assign ranking numbers Sy, starting from 1 and ending at N; Form a new ranking number array M(Sxpy, Sy) by combining the expected value of the first performance index X and the ranking number of the second performance index Y according to the original corresponding relationship between the expected value of the first performance index X and the second performance index Y; Calculate the correlation degree δ(X, Y) between the first performance index X and the second performance index Y.

6. The big data analysis method for expressing the unit state according to claim 5, characterized in that: The correlation degree δ(X, Y) between the first performance index X and the second performance index Y is expressed as: where S i = (Sxpy - Sy), and i ranges from 1 to N.

7. The big data analysis method for expressing the unit status according to claim 6, characterized in that: The judgment of the correlation strength includes judging whether there is a strong correlation between the first performance index X and the second performance index Y according to the first judgment criterion, the second judgment criterion and the third judgment criterion; The first judgment criterion is that the absolute values of all δ(X, Y) are greater than TH1; The second judgment criterion is that the average value of the absolute values of all δ(X, Y) is greater than TH2; The third judgment criterion is that the signs of all δ(X, Y) must be the same; When the first judgment criterion, the second judgment criterion and the third judgment criterion are all satisfied at the same time, it is determined that there is a strong correlation between the first performance index X and the second performance index Y.

8. A big data analysis system for expressing the state of a unit, which applies the method according to any one of claims 1-7, characterized in that, Including: An acquisition module for acquiring the time series data of the historical first performance index and the second performance index; A preprocessing module for performing data screening on the time series data after stepped processing to screen out the most stable data segment; A data screening module, which is used to process the screened data based on the most stable data segment and establish a relationship model; A correlation degree calculation module, which is used to calculate the correlation degree between performance indicators based on the relationship model by calculating the difference in rank numbers; A correlation strength judgment module, which is used to identify the key performance indicators of the health status of specific components of the unit by calculating the correlation degree and judging the correlation strength; A health status evaluation module, which is used to evaluate the health status of specific components of the unit based on the key performance indicators and the correlation between performance indicators.

9. A computing device, characterized in that, It includes: A memory, which is used to store programs; A processor, which is used to load the program to execute the steps of the big data analysis method for expressing the unit state as described in any one of claims 1-7.

10. A computer-readable storage medium stores a program, characterized in that, When the program is executed by the processor, the steps of the big data analysis method for expressing the unit state as described in any one of claims 1-7 are realized.

Images