Method for predicting service life of sealing ring of ethylene-propylene-diene rubber for cooling system of converter valve

Abstract

The application discloses a life prediction method for a three-element ethylene propylene rubber sealing ring of a converter valve cooling system, and comprises the following steps: firstly, medium accelerated aging experiments of the three-element ethylene propylene rubber sealing ring are carried out at n temperatures; a nonlinear fitting method is used to establish a relationship between a compression permanent deformation retention rate 1-s and a logarithm lnt of an aging time; the relationship is used to calculate a time when the compression permanent deformation retention rate is P% under each operating environment temperature; the calculated time is brought into a life prediction model to obtain the life prediction model of the three-element ethylene propylene rubber sealing ring; the life prediction model is trained and optimized, and finally, the life prediction model can be used. The accelerated aging test adopted in the method is a medium resistance aging, and the medium is deionized water used in an actual operating environment, so that a real aging environment of the rubber sealing ring in the converter valve cooling system is simulated, and aging data is more reliable.

Description

Technical Field

[0001] This invention relates to a method for predicting the lifespan of rubber seals, and particularly to a method for predicting the lifespan of EPDM rubber seals used in converter valve cooling systems. Background Technology

[0002] The converter valve is one of the core components of a converter station. During operation, it generates various losses, leading to increased junction temperatures in the thyristor elements. Excessive temperatures can cause thyristor failure, necessitating a valve cooling system to maintain the thyristor junction temperature within the normal range and ensure reliable valve operation. The rubber ring in the valve's internal cooling system is a key sealing component for the converter valve's cooling system; its performance and service life directly affect the cooling performance of the valve cooling system.

[0003] Extensive research has been conducted both domestically and internationally on the aging performance and life prediction of rubber seals. However, this research primarily focuses on the life prediction of rubber seals used in missiles, gas pipelines, high-speed trains, the petroleum industry, rocket engines, and pressure vessels. There is a lack of research on the life prediction of rubber seals in the internal cooling systems of converter valves in high-voltage direct current transmission systems. This lack of research prevents the provision of reference for the operation, management, and maintenance of converter valve cooling systems. The main reason for this is the complexity of the aging process of rubber seals in internal cooling systems. The aging mechanism under the combined effects of temperature, pressure, moisture, oxygen, and vibration is still unclear. Furthermore, the composition and performance requirements of rubber seals for different applications vary significantly, making it impossible to simply apply research findings from other applications to rubber seals in internal cooling systems. Summary of the Invention

[0004] In view of the above-mentioned problems in the existing technology, the technical problem to be solved by the present invention is: how to predict the life of the EPDM rubber seal used in the converter valve cooling system as accurately as possible.

[0005] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: a method for predicting the lifespan of EPDM rubber sealing rings used in converter valve cooling systems, comprising the following steps:

[0006] S1: Accelerated aging tests were conducted on EPDM rubber seals at n temperatures. The rubber seals were compressed to 25% using a fixture, and samples were taken daily at room temperature for t min to record the compression set of the rubber seals.

[0007] S2: A nonlinear fitting method is used to establish the relationship between the compression set retention rate 1-s and the logarithm of aging time lnt for the data in S1;

[0008] S3: Calculate the time when the compression permanent deformation retention rate is P% at each operating environment temperature using the relationship fitted by S2;

[0009] S4: Substitute the calculation time in S3 into the life prediction model after deformation processing based on the Arrhenius equation to obtain the life prediction model of EPDM rubber seal.

[0010] S5: Obtain failure data of EPDM rubber seals in existing converter valve cooling systems, use the BP (Backpropagation) neural network model to continuously optimize the life prediction model of EPDM rubber seals, and adjust the parameters of the life prediction model of EPDM rubber seals to obtain the optimized life prediction model of EPDM rubber seals.

[0011] S6: Collect the operating ambient temperature of the EPDM rubber seal in the cooling system of the converter valve to be predicted, and input the operating ambient temperature into the optimized EPDM rubber seal life prediction model obtained in S5 to calculate the life of the EPDM rubber seal in the cooling system of the converter valve to be predicted.

[0012] Preferably, the relationship between 1-s and lnt established in S2 is as follows:

[0013] 1-s=a×(lnt) 2 +b×lnt+c;(1)

[0014] Where a represents the quadratic coefficient in the quadratic fitting, b is the linear coefficient, and c is the constant term;

[0015] The time when the compression set retention rate is 50% at each temperature is calculated using the S2-fitted relationship.

[0016] Preferably, the life prediction model for the EPDM rubber seal in S4 is as follows:

[0017]

[0018] In the formula E T Here, is the activation energy of the reaction, T is the thermodynamic temperature (K), R is the gas molar constant, and B is a constant.

[0019] Preferably, in step S5, the parameters in the life prediction model of the EPDM rubber seal are adjusted using a BP neural network.

[0020] E T The process with B is as follows:

[0021] S51: Obtain failure data of EPDM rubber seals in existing converter valve cooling systems to form a training set. The data representing each training sample in the training set includes the operating environment temperature and actual failure time of the EPDM rubber seal.

[0022] S52: Initialize the parameters of the life prediction model for EPDM rubber seals;

[0023] S53: Input the i-th training sample into the life prediction model of the EPDM rubber seal to obtain the predicted failure time of the i-th training sample, i∈[1,N], where N is the number of training samples in the training set;

[0024] S54: Calculate the difference between the actual failure time and the predicted failure time of the i-th training sample, and use gradient descent to adjust E. T Update and iterate with B;

[0025] S55: When the number of iterations exceeds the preset maximum number of iterations, end the training to obtain the optimized EPDM rubber seal life prediction model; otherwise, return to S53.

[0026] Compared with the prior art, the present invention has at least the following advantages:

[0027] 1. Based on the Arrhenius equation and BP neural network algorithm, this invention proposes a method for predicting the lifespan of EPDM rubber seals used in converter valve cooling systems, providing guidance for operation and maintenance.

[0028] 2. The accelerated aging test used in this method is the aging test of the medium, which is deionized water used in the actual operating environment. This simulates the real aging environment of the rubber seals in the converter valve cooling system, making the aging data more reliable.

[0029] 3. For fitting the compression set retention rate and aging time, nonlinear fitting (quadratic fitting) is used instead of linear fitting, which provides a better fit and better reflects the relationship between the two.

[0030] 4. This method can continuously optimize the model parameters using failure data of rubber seals under actual operation, thereby making the model's prediction results more accurate.

[0031] 5. This method only requires the actual operating temperature to make a relatively accurate prediction of its lifespan. Attached Figure Description

[0032] Figure 1 This is a simplified flowchart of the overall process of the present invention.

[0033] Figure 2 The nonlinear fitting curves of 1-s and lnt (day) at 90℃ are shown.

[0034] Figure 3 The nonlinear fitting curves of 1-s and lnt (day) at 75℃ are shown.

[0035] Figure 4 The nonlinear fitting curves of 1-s and lnt (day) at 60℃ are shown.

[0036] Figure 5 This is a linear fit between lnt and 1 / T. Detailed Implementation

[0037] The present invention will now be described in further detail.

[0038] This invention proposes a life prediction method for EPDM rubber seals used in converter valve cooling systems based on the Arrhenius equation and BP neural network algorithm. The method involves conducting accelerated aging tests on the EPDM rubber seals at three typical temperatures, recording the compression set rate and corresponding aging time. Then, nonlinear fitting is performed on the compression set rate and aging time to calculate the constant term in the life prediction formula, thus obtaining a preliminary life prediction model for the rubber seals. This model is then optimized and corrected using historical failure data and the BP neural network to ultimately obtain the predicted life of the rubber seals, providing replacement recommendations for rubber seals in operating converter valve cooling systems.

[0039] See Figure 1 The method for predicting the lifespan of EPDM rubber seals used in converter valve cooling systems includes the following steps:

[0040] S1: Accelerated aging tests were conducted on EPDM rubber seals at 60℃, 75℃, and 90℃ using deionized water as the medium. The rubber seals were compressed to 25% using a fixture, and samples were taken daily at room temperature for 30 minutes to recover, recording the compression set rate (s).

[0041] S2: A nonlinear fitting method is used to establish the relationship between the compression set retention rate 1-s and the logarithm of aging time lnt for the data in S1;

[0042] 1-s=a*(lnt) 2 +b*lnt+c; (1)

[0043] Where a represents the coefficient of the quadratic term in the quadratic fitting, b is the coefficient of the linear term, and c is the constant term.

[0044] S3: Calculate the time when the compression set retention rate is 50% at each temperature using the relationship fitted by S2;

[0045] S4: Substitute the calculation time in S3 into the life prediction model after deformation processing based on the Arrhenius equation to obtain the life prediction model of EPDM rubber seal.

[0046]

[0047] In the formula E T Here, T is the activation energy of the reaction, T is the thermodynamic temperature (K), R is the gas molar constant, R = 8.34 J / mol·K, and B is a constant.

[0048] S5: Obtain failure data of EPDM rubber seals in existing converter valve cooling systems, continuously optimize the EPDM rubber seal life prediction model using a BP neural network model, and adjust the parameters of the EPDM rubber seal life prediction model to obtain an optimized EPDM rubber seal life prediction model, making the model more accurate.

[0049] Using a backpropagation neural network to adjust the parameter E in the life prediction model of EPDM rubber seals T The process with B is as follows:

[0050] S51: Obtain failure data of EPDM rubber seals in existing converter valve cooling systems to form a training set. The data representing each training sample in the training set includes the operating environment temperature and actual failure time of the EPDM rubber seal.

[0051] S52: Initialize the parameters for the life prediction model of EPDM rubber seals:

[0052] S53: Input the i-th training sample into the life prediction model of the EPDM rubber seal to obtain the predicted failure time of the i-th training sample, i∈[1,N], where N is the number of training samples in the training set;

[0053] S54: Calculate the loss between the actual failure time and the predicted failure time for the i-th training sample, and apply gradient descent to E. T Update with B;

[0054] S55: When the number of iterations exceeds the preset maximum number of iterations, end the training to obtain the optimized EPDM rubber seal life prediction model; otherwise, return to S53.

[0055] S6: Collect the operating ambient temperature of the EPDM rubber seal in the cooling system of the converter valve to be predicted, and input the operating ambient temperature into the optimized EPDM rubber seal life prediction model obtained in S5 to calculate the service life of the EPDM rubber seal in the cooling system of the converter valve to be predicted. Based on this, maintenance or replacement measures can be taken for the EPDM rubber seal.

[0056] See the example. Figures 2-5 :

[0057] The nonlinear fitting curves of 1-s and lnt (day) for EPDM rubber sealing rings at 90℃ are as follows:

[0058] 1-s = -0.00819*(lnt) 2 -0.02601*lnt+0.92808; (3)

[0059] The nonlinear fitting curves of 1-s and lnt (day) for EPDM rubber sealing rings at 75℃ are as follows:

[0060] 1-s = -0.00681*(lnt) 2 -0.01488*lnt+0.93905; (4)

[0061] The nonlinear fitting curves of 1-s and lnt (day) for EPDM rubber sealing rings at 60℃ are as follows:

[0062] 1-s = -0.00673*(lnt) 2 -0.00231*lnt+0.95754; (5)

[0063] Table 1. Logarithm of time (lnt) to reach aging critical value at different temperatures

[0064] Temperature / °C 60 75 90 1 / T(1 / K) 1 / 60 1 / 75 1 / 90 lnt (day) 8.0750 7.0190 6.1792

[0065] The optimized life prediction model for EPDM rubber seals is as follows:

[0066]

[0067] Substituting the values ​​into the calculation, we can see that the lifespan of the EPDM rubber seal is 17.64 years when the ambient temperature is 50℃, and 37.72 years when the ambient temperature is 40℃.

[0068] This invention effectively combines a BP neural network, a lifetime prediction model, and the relationship between deformation retention rate and the logarithm of aging time: 1. The compressed permanent deformation retention rate data is obtained by simulating the actual operating environment, and its reliability is the foundation for ensuring the accuracy of the prediction model; 2. The relationship between deformation retention rate and the logarithm of aging time provides fitting parameters for the lifetime prediction model. The quality of the fitting effect of the relationship between deformation retention rate and the logarithm of aging time is the key to ensuring the lifetime prediction model, and the fitting effect of the quadratic term is excellent; 3. The BP neural network processes historical failure data to adjust the model parameters, making the lifetime model obtained through experiments more suitable for the actual operating conditions in the field.

[0069] This method employs accelerated aging tests with a medium, specifically deionized water used in actual operating environments, to simulate the real aging environment of rubber seals in converter valve cooling systems, thus making the aging data more reliable. For fitting the compression set retention rate and aging time, a nonlinear fitting (quadratic term fitting) is used instead of a linear fitting, resulting in better fitting and a more accurate representation of the relationship between the two. This method continuously optimizes model parameters using failure data of rubber seals under actual operating conditions, leading to more accurate predictions. Furthermore, this method only requires the actual operating temperature to accurately predict the lifespan of the seals.

[0070] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for predicting the lifespan of EPDM rubber seals used in converter valve cooling systems, characterized in that: Includes the following steps: S1: Accelerated aging tests were conducted on EPDM rubber seals at n temperatures. The rubber seals were compressed to 25% using a fixture, and samples were taken daily at room temperature for t min to record the compression set of the rubber seals. S2: Using a nonlinear fitting method, establish the relationship between the compression set retention rate 1-s and the logarithm of the aging time lnt based on the data in S1; the established relationship between 1-s and lnt is as follows: 1-s=a×(lnt) 2 +b×lnt+c;(1) Where a represents the quadratic coefficient in the quadratic fitting, b is the linear coefficient, and c is the constant term; The time when the compression set retention rate is 50% at each temperature is calculated using the S2-fitted relationship. S3: Calculate the time when the compression permanent deformation retention rate is P% at each operating environment temperature using the relationship fitted by S2; S4: Substitute the calculation time in S3 into the life prediction model after deformation processing based on the Arrhenius equation to obtain the life prediction model of EPDM rubber seal. The life prediction model for the EPDM rubber seal is as follows: ；（2） In the formula E T Here, T is the activation energy of the reaction, T is the thermodynamic temperature (K), R is the gas molar constant, and B is a constant. S5: Obtain failure data of EPDM rubber seals in existing converter valve cooling systems, continuously optimize the life prediction model of EPDM rubber seals using a BP neural network model, and adjust the parameters of the life prediction model of EPDM rubber seals to obtain the optimized life prediction model of EPDM rubber seals. S6: Collect the operating ambient temperature of the EPDM rubber seal in the cooling system of the converter valve to be predicted, and input the operating ambient temperature into the optimized EPDM rubber seal life prediction model obtained in S5 to calculate the life of the EPDM rubber seal in the cooling system of the converter valve to be predicted.

2. The method for predicting the lifespan of EPDM rubber seals for converter valve cooling systems as described in claim 1, characterized in that: In step S5, a BP neural network is used to adjust the parameter E in the life prediction model of the EPDM rubber seal. T The process with B is as follows: S51: Obtain failure data of EPDM rubber seals in existing converter valve cooling systems to form a training set. The data representing each training sample in the training set includes the operating environment temperature and actual failure time of the EPDM rubber seal. S52: Initialize the parameters of the life prediction model for EPDM rubber seals; S53: Input the i-th training sample into the life prediction model of the EPDM rubber seal to obtain the predicted failure time of the i-th training sample. N is the number of training samples in the training set; S54: Calculate the difference between the actual failure time and the predicted failure time of the i-th training sample, and use gradient descent to adjust E. T Update and iterate with B; S55: When the number of iterations exceeds the preset maximum number of iterations, end the training to obtain the optimized EPDM rubber seal life prediction model; otherwise, return to S53.

Images