A method and computer program product for determining abnormal wear of an end face seal

Abstract

The application discloses a kind of methods and computer program products for judging abnormal wear of end face seal, judges abnormal wear of end face seal based on multi-parameter linear programming, belongs to the field of seal anomaly monitoring, by the multiple monitoring values such as end face seal compression amount, temperature monitoring point, leakage monitoring point and pressure monitoring point in end face seal test system as characteristic value, with the temperature parameter rising rate collected as the basis, by considering the fluctuation of end face seal compression amount, leakage monitoring point and pressure monitoring point signal, the current temperature rising condition is corrected, finally compared with preset temperature rising speed threshold value, multiple analysis and determination are carried out, the accuracy of temperature anomaly determination is improved, at the same time, this method is based on an infrared end face seal measurement test device, improves the accuracy of seal surface temperature rise measurement.

Description

Technical Field

[0001] This invention belongs to the field of sealing fault diagnosis, and in particular relates to a method for judging abnormal wear of end face seals based on multi-parameter linear programming. Background Technology

[0002] In the field of liquid rocket engine turbopumps, various forms of end face seals are widely used, such as contact and non-contact seals. To monitor the stability of the seal's operating status, temperature, flow rate, and power measurement points are often set up in the seal test device, and the data is collected and stored through an automated system. Once the seal wears abnormally, it will cause serious damage to the product's sealing surface, resulting in seal failure, and in severe cases, it will lead to the destruction of the operating test system.

[0003] Current technologies often rely on manual or equipment-based determination of maximum temperature thresholds to assess whether a seal is operating normally. However, this approach suffers from several drawbacks. First, setting a single threshold fails to account for temperature rises caused by ambient temperature and excessive product pressure. Second, it doesn't adequately consider the additional impact of the end-face seal's own compression on temperature rise. Furthermore, the temperature rise from end-face seal wear primarily occurs on the sealing contact surfaces. During operation, abnormal wear in the end-face seal would result in a rapid cycle of contact-separation and re-contact friction between the stationary and rotating rings, causing fluctuations in power and leakage parameters. Therefore, a monitoring scheme based solely on a single intracavity air temperature threshold is unsuitable for identifying abnormal wear due to its slow monitoring and warning mechanisms and inaccurate parameter assessments.

[0004] To address these issues, we propose a method based on multi-parameter linear programming for judging abnormal wear of end face seals. Summary of the Invention

[0005] The technical problem solved by this application is to overcome the shortcomings of the prior art and provide a method for judging abnormal wear of end face seals based on multi-parameter linear programming. Based on the end face seal surface temperature obtained by infrared thermometry, by considering the fluctuation of product compression parameters, operating power parameters, and leakage parameters, the method effectively reduces the influence of product specific pressure factors, product compression factors, ambient temperature, and untimely cavity temperature response. This helps to improve the accuracy of end face seal temperature rise judgment, detect abnormal wear of end face seals in advance during operation, and protect the product and test system to a certain extent.

[0006] The technical solution provided in this application is as follows:

[0007] A method for determining abnormal wear of an end face seal includes:

[0008] S1: Obtain data on the sealing cavity during x normal operation of the end face seal;

[0009] In the data of the sealing cavity during each normal operation of the end face seal, the values ​​of multiple measurement parameters are extracted from the temperature rise signal segments corresponding to multiple different time points. The multiple measurement parameters are the target parameter and n decision parameters. The target parameter is the temperature parameter. x sets of samples are obtained. Each set of samples contains n+1 measurement parameters, and each measurement parameter contains m data.

[0010] S2: Based on the temperature parameter data in S1, calculate the temperature rise slope data for x temperature parameters; based on the x temperature rise slope data, calculate the mean temperature rise slope and the root mean square value of the temperature rise; and calculate the temperature rise slope judgment threshold [K] based on the mean temperature rise slope and the root mean square value of the temperature rise.

[0011] S3: Based on the decision parameter data in S1, obtain the normalized decision parameter data matrix AN and the normalized objective parameter matrix TN, and obtain the prediction weight coefficient matrix [n1 n2...n] based on the decision parameter data matrix AN and the normalized objective parameter matrix TN. n The normalized objective parameter matrix TN' and the weight coefficient matrix should satisfy the constraints.

[0012] S4: Process the data to be judged. The data to be judged includes data of multiple measurement parameters. The multiple measurement parameters include the target parameter and n decision parameters. The target parameter is the temperature parameter. Each measurement parameter contains m data within the time window length Δt.

[0013] Based on step S3 and the data to be judged, the normalized temperature rise slope adjustment empirical value K2' is obtained. The normalized temperature rise slope adjustment empirical value K2' is then reverse-normalized within the sample data range to obtain the temperature rise slope adjustment empirical value K2.

[0014] Based on the m data points included in the temperature parameters, the current temperature rise slope K1 is calculated.

[0015] S5: Adjust the empirical value K2 and the current temperature rise slope K1 according to the temperature rise slope to obtain the temperature rise slope judgment value K; when K is greater than the temperature rise slope judgment threshold [K], it is determined that the end face seal is abnormally worn, otherwise it is normal.

[0016] Preferably, the decision parameters are operating power parameters, leakage parameters of sealing products, or assembly compression of sealing products; the temperature parameters and m data points of the decision parameters are collected sequentially at different time points within the sampling time.

[0017] Preferably, in S1, the sample size x ≥ 10 × (number of measurement parameters ÷ end-face sealing product qualification rate).

[0018] Preferably, in step S2, ±2.5s of the time point with the maximum slope of the temperature rise data is used as the time window length Δt, and the data of the measured parameters are taken within the time window length Δt.

[0019] Preferably, in step S2, based on the temperature parameter data in S1, the temperature rise slope data for x temperature parameters is calculated; based on the x temperature rise slope data, the mean slope and root mean square value are calculated; and the temperature rise slope judgment threshold [K] is calculated based on the mean slope and root mean square value, including:

[0020] Based on the temperature parameter data in S1, calculate the temperature slope value of each sample's temperature parameter data within the time window Δt, and obtain the temperature slope value matrix K = [K1 K2 ... K...]. x ]';

[0021] Then, based on the temperature slope matrix K = [K1 K2 ... K... x The average temperature rise slope of the temperature parameter data of x groups of samples is obtained. The threshold for judging the temperature rise slope is obtained by combining the root mean square value σ.

[0022] Preferably, in step S3, based on the decision parameter data in S1, a normalized decision parameter data matrix AN and a normalized target parameter matrix TN are obtained, and a prediction weight coefficient matrix [n1 n2...n] is obtained based on the decision parameter data matrix AN and the normalized target parameter matrix TN. n The normalized objective parameter matrix TN' and the weight coefficient matrix should satisfy the following constraints:

[0023] The n decision parameters are divided into l predetermined parameters and k operational fluctuation parameters. The predetermined parameters are those that remain unchanged after the sealing assembly is assembled, while the operational fluctuation parameters are those that fluctuate with operation. The predetermined parameters and operational fluctuation parameters are represented in matrix format, and the predetermined parameter matrix of the y-th sample is obtained. Coefficient of variation matrix of operating fluctuation parameters

[0024] The coefficient of variation matrix of the operational fluctuation parameters of the y-th sample group With the predetermined parameter matrix The decision parameter data matrix obtained by combining the data is as follows: For AN y Normalize the data within the range of 0 to 1 to obtain the normalized decision parameter data matrix AN;

[0025] Based on the temperature parameter data, the temperature slope matrix K = [K1 K2 ... K] is calculated. y ]', K yLet K represent the temperature slope of the y-th sample group within the time window length Δt. After normalizing the temperature slope value matrix K to the range of 0 to 1, the normalized objective parameter matrix TN = [TN1 TN2 ... TN2] is obtained. y ]';

[0026] The weight coefficient matrix is ​​[n1 n2 ... n n The predicted normalized objective parameter matrix TN' = AN*[n1 n2 ... n n The weight coefficient matrix should satisfy the constraint condition to minimize the difference between TN' and TN.

[0027] Preferably, the predetermined parameters and operational fluctuation parameters are represented in matrix format to obtain the predetermined parameter matrix. Coefficient of variation matrix of operating fluctuation parameters include:

[0028] For l predetermined parameters, for each of x sets of data, within m data points of time window length Δt, the predetermined matrix... Predefined parameter matrix

[0029] For k operational fluctuation parameters, for each of x sets of data, within m time series of a time window length Δt, calculate the k operational fluctuation parameters a2 = [a1 a2 ... a2 ... a3]. k The standard deviation matrix of the operational fluctuation parameters. With the mean matrix of operating fluctuation parameters The m sequence matrices of the g-th operational fluctuation parameter in the y-th group are: The standard deviation matrix of the operational fluctuation parameters is as follows The mean matrix of operational fluctuation parameters is Based on the standard deviation matrix of operational fluctuation parameters With the mean matrix of operating fluctuation parameters The quotient is used to obtain the coefficient of variation matrix of the operational fluctuation parameters. The coefficient of variation matrix of the operational fluctuation parameters is as follows:

[0030] Preferably, the weighting coefficient matrix satisfies the following constraints:

[0031] The minimum error value is achieved, i.e., o = min(TN' - TN);

[0032] n1 n2...n n They are all distributed between 0 and 1;

[0033] The l predetermined parameters have the same weight, i.e., n1, n2...n l All are equal;

[0034] The k operational fluctuation parameters have the same weight, i.e., n l+1 ...n l+k All are equal, n = l + k;

[0035] The sum of the weights corresponding to the predetermined parameters is equal to the sum of the weights corresponding to the operational fluctuation parameters.

[0036] Preferably, in step S4, the normalized temperature rise slope adjustment empirical value K2' is obtained based on step S3 and the data to be judged, including:

[0037] Based on step S3 and the decision parameter data of the data to be judged, the decision parameter data matrix is ​​obtained. The decision parameter data matrix is ​​then normalized within the sample data range to obtain the predicted normalized decision parameter matrix TN'.

[0038] Based on the predicted normalized target parameter matrix TN' in S3, and combined with the temperature parameter data of the data to be judged, the empirical value of the normalized temperature rise slope adjustment K2' is obtained, K2' = AN. i *[n1 n2...n n ]'.

[0039] A computer program product comprising a computer program / instructions that, when executed by a processor, implement the steps of any of the methods described above.

[0040] In summary, this application includes at least the following beneficial technical effects:

[0041] By considering multiple parameters, including pre-defined parameters such as product assembly parameters in addition to temperature parameters, as well as operational fluctuation parameters such as product power and leakage, a comprehensive judgment is made on the current temperature rise of the end face seal. This results in a temperature rise caused by the operation of the end face seal under different actual working conditions, and it is used to determine whether wear has occurred in the current state. This reduces misjudgments caused by a single temperature parameter on the wear status of the product and improves the accuracy of judging abnormal wear of the end face seal product. Attached Figure Description

[0042] Figure 1 This document provides a flowchart of a method for determining abnormal wear of end face seals based on multi-parameter linear programming, mainly including the process from fitting historical data to processing and judging the data to be judged.

[0043] Figure 2 This describes the processing flow of judgment data in a method for judging abnormal wear of end face seals based on multi-parameter linear programming.

[0044] Figure 3 This is a measuring device for obtaining end-face sealing temperature parameters based on infrared thermometry. It obtains the temperature of the end-face sealing product by infrared thermometry, and simultaneously obtains the leakage and power of the end-face sealing product during operation.

[0045] Explanation of reference numerals: 1. Infrared temperature sensor; 2. Temperature measuring base; 3. Visualizing lens; 4. Temperature measuring adapter; 5. Shaft head locking; 6. Moving ring; 7. Stationary ring; 8. Spindle; 9. Leakage measurement connector; 10. End cap; 11. Housing. Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments disclosed in the present invention will be described in further detail below with reference to the accompanying drawings.

[0047] This application discloses a method for determining abnormal wear of end face seals based on multi-parameter linear programming, such as... Figure 1 and Figure 2 As shown, the judgment method includes the following steps:

[0048] S1: Obtain data on the sealing cavity during x normal operation of the end face seal;

[0049] In the data of the sealing cavity during normal operation of the end face seal, the values ​​of multiple measurement parameters are extracted from the temperature rise signal segments corresponding to multiple different time points. The multiple measurement parameters are the target parameter and n decision parameters. The decision parameters are the operating power parameter, the leakage parameter of the sealing product, or the assembly compression of the sealing product, and other data that can be collected or are needed. The target parameter is the temperature parameter. Each set of temperature parameters and decision parameters contains m data points, which are collected sequentially at different time points within the sampling time.

[0050] Data from the sealing cavity during each normal operation of the end face seal is extracted to obtain x sets of samples, each set of samples containing (n+1) measurement parameters;

[0051] S2 calculates the temperature rise slope data of the temperature parameters x times mentioned in S1 based on the temperature parameter samples in S1; calculates the mean temperature rise slope and root mean square temperature rise value based on the x temperature rise slope data, and calculates the temperature rise slope judgment threshold [K] based on the mean temperature rise slope and root mean square temperature rise value.

[0052] Based on the decision parameter data in S1, S3 obtains the normalized decision parameter data matrix AN and the normalized objective parameter matrix TN, and then obtains the prediction weight coefficient matrix [n1n2...n] based on the decision parameter data matrix AN and the normalized objective parameter matrix TN. n The normalized objective parameter matrix TN' and the weight coefficient matrix should satisfy the constraints.

[0053] S4 processes the data to be judged, which includes data from multiple measurement parameters, such as the compression amount δ and leakage amount Q of the worn product. qThe system collects or requires data on n parameters, such as operating power P; each measured parameter contains m data points with a time window length Δt; based on step S3 and the data to be judged, the normalized temperature rise slope adjustment empirical value K2' is obtained, and the normalized temperature rise slope adjustment empirical value K2' is inversely normalized within the sample data range to obtain the temperature rise slope adjustment empirical value K2;

[0054] For the newly acquired temperature rise parameters of the sealed cavity, calculate the corresponding current temperature rise slope K1;

[0055] S5 calculates the temperature rise slope adjustment empirical value K2 and the current temperature rise slope K1, and the temperature rise slope judgment value K; when K is greater than the temperature rise slope judgment threshold [K], it is determined that the end face seal is abnormally worn, otherwise it is normal.

[0056] Preferably, S1 specifically includes: the collected data may include temperature parameters, operating power parameters, leakage parameters of sealing products, and other collectable or required data such as the compression amount of sealing products during assembly.

[0057] Preferably, S1 specifically includes: the parameter data involved must be collected simultaneously, that is, the temperature parameter and decision parameter data have the same time point.

[0058] Preferably, S1 specifically includes: the number of selected measurement parameter samples should conform to the EVP (events pervariable) principle, that is, the number of selected end-face sealing test normal operation data samples multiplied by not less than ten times the number of selected measurement parameters and the pass rate of the end-face sealing product; the selected measurement parameters include, but are not limited to, temperature parameters, operating power parameters, leakage parameters of sealing products and assembly compression of sealing products.

[0059] Preferably, S1 specifically includes: the time range of the selected measurement parameter data should be ±2.5s of the time point with the maximum slope of the temperature rise data as the time window length Δt of the corresponding sample.

[0060] Preferably, S2 specifically includes: the temperature rise slope judgment threshold [K] is determined by the 3σ principle, firstly, the temperature slope value of the temperature parameter data of each group of samples in the time window Δt is obtained, and the temperature slope value matrix K = [K1 K2 ... K x Then, calculate the average temperature rise slope of the temperature parameter data for x groups of samples. The threshold for judging the temperature rise slope is obtained by combining the root mean square value σ.

[0061] Preferably, S3 specifically includes decision parameters, target parameters, and weighting coefficients:

[0062] Among the n decision parameters, there are l predetermined parameters and k operational fluctuation parameters. Predetermined parameters are parameters that do not change after the sealing component is assembled, such as the product compression parameter; operational fluctuation parameters are parameters that fluctuate with operation, such as the operating power parameter and the leakage parameter of the sealing product.

[0063] For l predetermined parameters, for each of x sets of data, within m data points of time window length Δt, the predetermined matrix... Each predetermined parameter value remains consistent across m data points. Therefore, the first set of data is used as the predetermined matrix value, and no preprocessing is performed. The resulting predetermined parameter matrix... For k operational fluctuation parameters, for each of the x sets of data, preprocessing is required within the m time series of a time window length Δt. The preprocessing method is to calculate the k operational fluctuation parameters. Standard deviation matrix of operational fluctuation parameters With the mean matrix of operating fluctuation parameters The m-time series matrices of the g-th operational fluctuation parameter in the y-th group are as follows: Calculating the standard deviation and mean of each, we obtain the standard deviation matrix of the operational fluctuation parameters as follows: The mean matrix of operational fluctuation parameters is Based on the standard deviation matrix of operational fluctuation parameters With the mean matrix of operating fluctuation parameters The quotient is used to obtain the coefficient of variation matrix of the operational fluctuation parameters; as shown in the formula in the table below; Coefficient of variation matrix of operational fluctuation parameters

[0064] The coefficient of variation matrix of the operational fluctuation parameters is as follows:

[0065] The decision parameter data matrix is ​​obtained by combining the coefficient of variation matrix of the operational fluctuation parameters of the y-th sample with the predetermined parameter matrix.

[0066] For n decision parameters (AN) y The decision parameter data matrix is ​​AN, which is normalized to the range of 0 to 1.

[0067] The target parameter is used as the evaluation parameter for prediction, and the temperature slope value matrix K = [K1 K2 ... K] is obtained from all sample data. y ]', where K y Let K represent the temperature slope of the y-th sample group within the time window length Δt. After normalizing the temperature slope value matrix K to the range of 0 to 1, the normalized objective parameter matrix TN = [TN1 TN2 ... TN2] is obtained. y ]'.

[0068] Wherein, the weight coefficients are the optimal weights for obtaining the linear programming objective parameters under all sample weight constraints, [n1 n2...n n [n1 n2 ... n] is the weight coefficient matrix, which assigns weights to each normalized decision parameter. The product of the weight coefficients and the normalized parameter matrix yields the predicted normalized target parameter matrix TN' = AN*[n1 n2 ... n]. n ]'.

[0069] Preferably, the weighting coefficient matrix should meet the following constraints, including: the linear programming function o = TN' - TN satisfies the minimum error value, i.e., o = min(TN' - TN); the weighting coefficient matrix n1 n2...n n The values ​​are distributed between 0 and 1; the weights corresponding to the l predetermined parameters are the same (n1, n2...n). l All are equal), and the k operational fluctuation parameters have the same weight (n l+1 ...n l+k All parameters are equal (n = l + k), and the sum of the weights corresponding to the predetermined parameters is equal to the sum of the weights corresponding to the operational fluctuation parameters. Under these constraints, a weight coefficient matrix of dimension 1×n is obtained.

[0070] Preferably, S4 specifically includes: processing the data to be judged, wherein the collection parameter range of the data to be judged is Δt, and the number of data to be judged is m;

[0071] Preferably, step S4 specifically includes: within the time window Δt of the data to be judged, obtaining a predetermined parameter matrix according to step S3; and simultaneously solving the coefficient of variation matrix of the operational fluctuation parameters for k of the n decision parameters according to step S3, to obtain the coefficient of variation matrix of the operational fluctuation parameters of the data to be judged.

[0072] Preferably, S4 specifically includes: within the time window Δt of the data to be judged, obtaining the decision parameter data matrix according to S3, and normalizing the n sets of decision parameter matrices within the corresponding sample data range of S3 (the corresponding normalization range of x sets of data) to obtain the predicted normalized decision parameter matrix AN. i = Normalization

[0073] Preferably, step S4 specifically includes obtaining the normalized temperature rise slope adjustment empirical value K2' by substituting the empirical formula for temperature rise slope adjustment from step S3: K2' = AN i *[n1 n2...n n The normalized temperature rise slope adjustment empirical value K2' is calculated by inverse normalization within the range of the corresponding sample data of S3 (the corresponding normalization range of x groups of data).

[0074] Preferably, S4 specifically includes: within the time window Δt of the data to be judged, based on m data points, calculating the current temperature rise slope K1, where K1 is the slope value of the temperature rise of the sealed cavity T within the unit time window length during the data collection.

[0075] Preferably, S5 specifically includes the following: the temperature rise slope judgment value K is obtained by averaging the temperature rise slope adjustment value K2 and the current temperature rise slope K1.

[0076] Preferably, step S5 specifically includes comparing the temperature rise slope judgment value K with the temperature rise slope judgment threshold [K]; if K is greater than the temperature rise slope judgment threshold [K], it is determined that the end face seal is abnormally worn, otherwise it is normal.

[0077]

[0078]

[0079] K1 Current temperature rise slope (15)

[0080] K = (K1 + K2) / 2 Temperature rise slope judgment value (16)

[0081] like Figure 3 As shown, this embodiment of the invention also provides an infrared end-face sealing temperature measurement test device. The system is used to monitor the end-face sealing structure, which includes a housing 11, a main shaft 8, a rotating ring 6, a stationary ring 7, and an end cap 10. The main shaft 8 is rotatably connected inside the housing 11, the stationary ring 7 is connected to the housing 11, the rotating ring 6 is connected to the main shaft 8, and the rotating ring 6 and the stationary ring 7 are in contact. The end cap 10 is fixedly connected to the end of the housing 11, achieving end-face sealing. A fixed assembly compression amount for the sealing product is generated during the installation of the main shaft 8 and the housing 11.

[0082] The test apparatus includes: a sealed operation test apparatus, an end cap for visual temperature measurement, and a measurement system.

[0083] Preferably, the sealing operation test device is a test fixture that provides rated speed and pressure for end face sealing, specifically including a main shaft, a rotating ring, a stationary ring assembly, and a housing. The main shaft and the rotating ring are fixedly connected by a locking method and operate together with the shaft system. The stationary ring is fixed to the housing, and the sealing surface of the stationary ring and the sealing surface of the rotating ring cooperate and move relative to each other. The stationary ring, the housing, and the shaft system form a sealed space, and a medium of corresponding pressure is introduced through the pipe nozzle. The shaft system is rotated by power input.

[0084] Preferably, the visualized temperature measuring end cap provides a sealing environment for the end face sealing while ensuring the visualization function of infrared temperature measurement, specifically including an end cap, a leakage measurement connector, a temperature measuring adapter, a visualization lens, and a temperature measuring base.

[0085] The end cap 10 is equipped with a temperature measuring adapter 4, which is connected to a temperature measuring base 2. A viewing lens 3 is provided between the temperature measuring base 2 and the temperature measuring adapter 4. The temperature measuring base 2 is used to connect to an infrared temperature sensor 1. Temperature parameters are obtained through the infrared temperature sensor 1.

[0086] Preferably, the measurement system specifically includes infrared temperature measurement, leakage measurement, compression measurement and power measurement equipment, wherein the infrared temperature sensor can measure the temperature of the end sealing surface inside the pressure vessel cavity through a visual lens, and the measurement system can collect and record data.

[0087] The end cap 10 is equipped with a leakage measurement connector 9, which is used to connect to a leakage measurement device to obtain leakage parameters of the sealing product.

[0088] The spindle 8 is connected to a power measuring device, through which the operating power parameters are obtained.

[0089] The end of the spindle 8 is connected to a shaft head locking structure 5.

[0090] Example 1

[0091] The working principle and characteristics of this invention will be further explained below using a certain end-face sealing product as an example. In this embodiment, a method for judging abnormal wear of end-face seals based on multi-parameter linear programming includes a sealing test device, which is equipped with an infrared temperature sensor, a leakage measurement flow sensor, and a power sensor. The judgment method includes the following steps:

[0092] S1: Obtain data on the sealing cavity during x normal operation of the end face seal;

[0093] In the data of the sealing cavity during normal operation of the end face seal, the values ​​of multiple measurement parameters are extracted from the temperature rise signal segments corresponding to multiple different time points. The multiple measurement parameters are the target parameter and n decision parameters. The decision parameters are the operating power parameter, the leakage parameter of the sealing product, or the assembly compression of the sealing product, and other data that can be collected or are needed. The target parameter is the temperature parameter. Each set of temperature parameters and decision parameters contains m data points, which are collected sequentially at different time points within the sampling time.

[0094] Data from the sealing cavity during each normal operation of the end face seal is extracted to obtain x sets of samples, each set of samples containing (n+1) measurement parameters;

[0095] S2 calculates the temperature rise slope data of the temperature parameters x times mentioned in S1 based on the temperature parameter samples in S1; calculates the mean temperature rise slope and root mean square temperature rise value based on the x temperature rise slope data, and calculates the temperature rise slope judgment threshold [K] based on the mean temperature rise slope and root mean square temperature rise value.

[0096] Based on the decision parameter data in S1, S3 obtains the normalized decision parameter data matrix and the normalized target parameter matrix, and obtains the prediction weight coefficient matrix and the normalized target parameter matrix based on the decision parameter data matrix and the normalized target parameter matrix. The weight coefficient matrix should satisfy the constraint conditions.

[0097] S4 processes the data to be judged, which includes data from multiple measurement parameters, such as the compression amount δ and leakage amount Q of the worn product. q The system collects or requires data on n parameters, such as operating power P; each measured parameter contains m data points with a time window length Δt; based on step S3 and the data to be judged, the normalized temperature rise slope adjustment empirical value K2' is obtained, and the normalized temperature rise slope adjustment empirical value K2' is inversely normalized within the sample data range to obtain the temperature rise slope adjustment empirical value K2;

[0098] For the newly acquired temperature rise parameters of the sealed cavity, calculate the corresponding current temperature rise slope K1;

[0099] S5 calculates the temperature rise slope adjustment empirical value K2 and the current temperature rise slope K1, and the temperature rise slope judgment value K; when K is greater than the temperature rise slope judgment threshold [K], it is determined that the end face seal is abnormally worn, otherwise it is normal.

[0100] Preferably, S1 specifically includes: the collected data may include temperature parameters, operating power parameters, leakage parameters of sealing products, and assembly compression of sealing products.

[0101] Preferably, S1 specifically includes: the parameter data involved must be collected simultaneously, that is, the temperature parameter and decision parameter data have the same time point.

[0102] Preferably, S1 specifically includes: the number of selected measurement parameter samples should conform to the EVP (events pervariable) principle, that is, the number of selected end-face sealing test normal operation data samples x is not less than ten times the quotient of the number of selected measurement parameters and the pass rate of end-face sealing products. Based on the number of selected parameters being 3, and based on the product pass rate being 90%, the number of selected samples should not be less than 34 pieces, and the sample number x is 34.

[0103] Preferably, S1 specifically includes: the time range of the selected measurement parameter data should be ±2.5s of the time point with the maximum slope of the temperature rise data as the time window length Δt of the corresponding sample, that is, Δt is 5s.

[0104] Preferably, S2 specifically includes: the temperature rise slope judgment threshold [K] is determined by the 3σ principle, firstly, the temperature slope value of the temperature parameter data of each group of samples in the time window Δt is obtained, and the temperature slope value matrix K = [K1 K2 ... K x Then, calculate the average temperature rise slope of the temperature parameter data for x groups of samples. The threshold for judging the temperature rise slope is obtained by combining the root mean square value σ. Based on the selected 34 sets of sample data, the average temperature rise slope was obtained. The root mean square value σ is 0.0511, and the final threshold [K] for judging the temperature rise slope is 0.2017.

[0105] Preferably, S3 specifically includes decision parameters, target parameters, and weighting coefficients:

[0106] Among the n decision parameters, there are l predetermined parameters and k operational fluctuation parameters. The predetermined parameters are those that do not change after the sealing assembly is assembled, and the selected parameter is the product compression amount δ, i.e., l = 1. The operational fluctuation parameters are those that fluctuate with operation, and the selected parameter is the leakage amount Q. q Operating power P, i.e., k = 2.

[0107] For one predetermined parameter, for each of the 34 data sets, within the time window length Δt, select m data points (m = 500), and predetermine the matrix. Since the predetermined parameter values ​​remain consistent across m data points, the first set of data is used as the predetermined matrix value, and no preprocessing is performed. The resulting predetermined parameter matrix... For each of the 34 data sets, given two operational fluctuation parameters, preprocessing is required within the m time series of a time window length Δt. The preprocessing method involves calculating the two operational fluctuation parameters. Standard deviation matrix of operational fluctuation parameters With the mean matrix of operating fluctuation parameters The quotient of the y-th data set, and the coefficient of variation matrix of the preprocessed operational fluctuation parameters.

[0108] The coefficient of variation matrix of the operational fluctuation parameters is as follows: The decision parameter data matrix is ​​obtained by combining it with the predetermined parameter matrix. For 3 decision parameters (AN) y The decision parameter data matrix is ​​AN, which is normalized to the range of 0 to 1.

[0109] The target parameter is used as the evaluation parameter for prediction, and the temperature slope value matrix K = [K1 K2 ... K] is obtained from all sample data. 34After normalizing the temperature slope matrix K to the range of 0 to 1, the normalized objective parameter matrix TN = [TN1 TN2 ... TN] is obtained. 34 ]'.

[0110] Wherein, the weight coefficient is the optimal weight for obtaining the linear programming objective parameter under all sample weight constraints, [n1 n2 n3] is the weight coefficient matrix, which assigns weights to each normalized decision parameter, and the product of the weight coefficient and the normalized processing parameter matrix is ​​the predicted normalized objective parameter matrix TN'=AN*[n1 n2 n3]'.

[0111] Preferably, the weight coefficient matrix should meet the following constraints: the linear programming function o = TN' - TN satisfies the minimum error value, i.e., o = min(TN' - TN); the weight coefficient matrix [n1 n2 n3] is distributed between 0 and 1; the weights corresponding to one predetermined parameter are the same, the weights corresponding to two operational fluctuation parameters are the same (n2 = n3), and the sum of the weights corresponding to the predetermined parameters is equal to the sum of the weights corresponding to the operational fluctuation parameters, n1 = n2 + n3. The final weight parameter matrix is ​​[n1 n2 n3] = [0.2104 0.1052 0.1052].

[0112] Preferably, S4 specifically includes: processing the data to be judged, wherein the range of the acquisition parameters of the data to be judged is Δt, and the number of data to be judged is m (m = 500);

[0113] Preferably, S4 specifically includes: within the time window Δt of the data to be judged, obtaining a predetermined parameter matrix according to S3; simultaneously solving for one predetermined parameter matrix, i.e., the current collected compression amount δ, for k operational fluctuation parameters among the n decision parameters according to S3; and simultaneously comparing the current power P with the current leakage amount Q. q Perform mutation processing, that is, modify the current collected power P and the current leakage amount Q respectively. q After calculating the standard deviation and mean, following the procedure described in S3, the coefficient of variation E is obtained by calculating the quotient of the standard deviation and mean of the two parameters. p E Qq .

[0114] Preferably, step S4 specifically includes: within the time window Δt of the data to be judged, obtaining the decision parameter data matrix according to step S3, and considering the current collected compression amount δ and the variation power E. p Variable leakage amount E Qq Normalization is performed within the sample data range to obtain the normalized compression δ0 and normalized power E. P0 Normalized leakage E Qq0 The corresponding prediction normalized decision parameter matrix AN is obtained by assembling it.i = [0.8400 1.9364 0.8481].

[0115] Preferably, step S4 specifically includes obtaining the normalized temperature rise slope adjustment empirical value K2' by substituting the empirical formula for temperature rise slope adjustment from step S3: K2' = AN i *[n1 n2 n3]', the final calculated K2' is 0.4697; the normalized temperature rise slope adjustment empirical value K2' is inversely normalized within the range of the sample data corresponding to S3 (the corresponding normalization range of x groups of data) to obtain the temperature rise slope adjustment empirical value K2 as 0.0786.

[0116] Preferably, S4 specifically includes: within the time window Δt of the data to be judged, based on m data (m=500), calculating the current temperature rise slope K1, which is 0.4025.

[0117] Preferably, step S5 specifically includes obtaining the temperature rise slope judgment value K by averaging the temperature rise slope adjustment value K2 and the current temperature rise slope K1, and obtaining K as 0.2799.

[0118] Preferably, step S5 specifically includes comparing the temperature rise slope judgment value K = 0.2799 with the temperature rise slope judgment threshold [K] = 0.2017; if the temperature rise slope judgment value K is greater than the temperature rise slope judgment threshold [K], it is determined that the end face seal is abnormally worn.

[0119] The contents not described in detail in this application specification are common knowledge to those skilled in the art.

[0120] The present application has been described in detail above with reference to specific embodiments and exemplary examples; however, these descriptions should not be construed as limiting the present application. Those skilled in the art will understand that various equivalent substitutions, modifications, or improvements can be made to the technical solutions and implementation methods of the present application without departing from the spirit and scope of the present application, and all such modifications and improvements fall within the scope of the present application. The scope of protection of the present application is determined by the appended claims.

Claims

1. A method for determining abnormal wear of an end face seal, characterized in that, include: S1: Obtain data on the sealing cavity during x normal operation of the end face seal; In the data of the sealing cavity during each normal operation of the end face seal, the values ​​of multiple measurement parameters are extracted from the temperature rise signal segments corresponding to multiple different time points. The multiple measurement parameters are the target parameter and n decision parameters. The target parameter is the temperature parameter. x sets of samples are obtained. Each set of samples contains n+1 measurement parameters, and each measurement parameter contains m data. S2: Based on the temperature parameter data in S1, calculate the temperature rise slope data for x temperature parameters; based on the x temperature rise slope data, calculate the mean slope and root mean square value, and calculate the temperature rise slope judgment threshold [K] based on the mean slope and root mean square value. S3: Based on the decision parameter data and temperature parameter data in S1, obtain the normalized decision parameter data matrix AN and the normalized objective parameter matrix TN, and obtain the prediction weight coefficient matrix [n1 n2...n] based on the decision parameter data matrix AN and the normalized objective parameter matrix TN. n The normalized objective parameter matrix TN' and the weight coefficient matrix should satisfy the constraints. S4: Process the data to be judged. The data to be judged includes data of multiple measurement parameters. The multiple measurement parameters include the target parameter and n decision parameters. The target parameter is the temperature parameter. Each measurement parameter contains m data within the time window length Δt. Based on step S3 and the data to be judged, the normalized temperature rise slope adjustment empirical value K2' is obtained. The normalized temperature rise slope adjustment empirical value K2' is then reverse-normalized within the sample data range to obtain the temperature rise slope adjustment empirical value K2. Based on the m data points included in the temperature parameters, the current temperature rise slope K1 is calculated. S5: Adjust the empirical value K2 and the current temperature rise slope K1 according to the temperature rise slope to obtain the temperature rise slope judgment value K; when K is greater than the temperature rise slope judgment threshold [K], it is determined that the end face seal is abnormally worn, otherwise it is normal.

2. The method for determining abnormal wear of an end face seal according to claim 1, characterized in that: The decision parameters are operating power parameters, leakage parameters of sealing products, or assembly compression of sealing products. The m data points for temperature and decision parameters were collected sequentially at different time points within the sampling period.

3. The method for determining abnormal wear of an end face seal according to claim 1, characterized in that: In S1, the sample size x ≥ 10 × (the quotient of the number of measurement parameters and the pass rate of the end-face sealing product).

4. The method for determining abnormal wear of an end face seal according to claim 1, characterized in that: In S2, ±2.5s of the time point with the maximum slope of the temperature rise data is used as the time window length Δt, and the data of the measured parameters are taken within the time window length Δt.

5. The method for determining abnormal wear of an end face seal according to claim 1, characterized in that, In step S2, based on the temperature parameter data in S1, the temperature rise slope data for x temperature parameters is calculated; based on the x temperature rise slope data, the mean slope and root mean square value are calculated; and the temperature rise slope judgment threshold [K] is calculated based on the mean slope and root mean square value, including: Based on the temperature parameter data in S1, calculate the temperature slope value of each sample's temperature parameter data within the time window Δt, and obtain the temperature slope value matrix K = [K1 K2 ... K...]. x ]'; Then, based on the temperature slope matrix K = [K1 K2 ... K... x The average temperature rise slope of the temperature parameter data of x groups of samples is obtained. The threshold for judging the temperature rise slope is obtained by combining the root mean square value σ.

6. The method for determining abnormal wear of an end face seal according to claim 1, characterized in that, In step S3, based on the decision parameter data and temperature parameter data from S1, a linear programming is performed to construct the decision parameter data matrix AN, the normalized objective parameter matrix TN, and the weight coefficient matrix [n1 n2...n]. n Based on the weight coefficient matrix and the decision parameter data matrix AN, the prediction normalized objective parameter matrix TN' is obtained, where TN' = AN * weight coefficient matrix, including: The n decision parameters are divided into l predetermined parameters and k operational fluctuation parameters. The predetermined parameters are those that remain unchanged after the sealing assembly is assembled, while the operational fluctuation parameters are those that fluctuate with operation. The predetermined parameters and operational fluctuation parameters are represented in matrix format to obtain the coefficient of variation matrix of the operational fluctuation parameters. With the predetermined parameter matrix Operating fluctuation parameter coefficient of variation matrix With the predetermined parameter matrix The resulting decision parameter data matrix is ​​as follows For AN y Normalize the data to the range of 0 to 1 to obtain the decision parameter data matrix AN; Based on the temperature parameter data, the temperature slope matrix K = [K1 K2 ... K] is calculated. y ]', K y Let K represent the temperature slope of the y-th sample group within the time window length Δt. After normalizing the temperature slope value matrix K to the range of 0 to 1, the normalized objective parameter matrix TN = [TN1 TN2 ... TN2] is obtained. y ]'; The weight coefficient matrix is ​​[n1 n2 ... n n ],TN'=AN*[n1 n2...n n ]'.

7. The method for determining abnormal wear of an end face seal according to claim 6, characterized in that, The predetermined parameters and operational fluctuation parameters are represented in matrix format to obtain the coefficient of variation matrix of the operational fluctuation parameters. With the predetermined parameter matrix include: For l predetermined parameters, for each of x sets of data, within m data points of time window length Δt, the predetermined matrix... y∈[1, x]; Predefined parameter matrix For k operational fluctuation parameters, for each of the y sets of data, within the m time series of the time window length Δt, calculate the k operational fluctuation parameters a2 = [a1 a2 ... ... k The standard deviation matrix of the operational fluctuation parameters. With the mean matrix of operating fluctuation parameters Where a k =[a1 a2...a m ] T The standard deviation matrix of the operational fluctuation parameters is as follows: The mean matrix of operational fluctuation parameters is Based on the standard deviation matrix of operational fluctuation parameters With the mean matrix of operating fluctuation parameters The quotient is used to obtain the coefficient of variation matrix of the operational fluctuation parameters. The coefficient of variation matrix of the operational fluctuation parameters is as follows:

8. The method for determining abnormal wear of an end face seal according to claim 6, characterized in that, The weighting coefficient matrix satisfies the following constraints: The minimum error value is achieved, i.e., o = min(TN' - TN); n1 n2...n n They are all distributed between 0 and 1; The l predetermined parameters have the same weight, i.e., n1, n2...n l All are equal; The k operational fluctuation parameters have the same weight, i.e., n l+1 ...n l+k All are equal, n = l + k; The sum of the weights corresponding to the predetermined parameters is equal to the sum of the weights corresponding to the operational fluctuation parameters.

9. The method for determining abnormal wear of an end face seal according to claim 1, characterized in that, In step S4, based on step S3 and the data to be judged, the normalized temperature rise slope adjustment empirical value K2' is obtained, including: Based on the decision parameter data of step S3 and the data to be judged, the decision parameter data matrix is ​​obtained. The decision parameter data matrix is ​​then normalized within the sample data range to obtain the predicted normalized decision parameter matrix. Based on the predicted normalized target parameter matrix TN' in S3, and combined with the temperature parameter data of the data to be judged, the empirical value of the normalized temperature rise slope adjustment K2' is obtained, K2' = AN. i *[n1 n2...n n ]'.

10. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or instructions are executed by a processor, they implement the steps of the method according to any one of claims 1-9.

Images