A method for predicting the lifespan of a liquid particle counting sensor

Abstract

A method for predicting the service life of a liquid particle counting sensor disclosed by the present invention belongs to the field of predicting the service life of liquid particle counting sensors. The present invention uses the deformation ratio of the slit area in the sensor as a measurement index, conducts accelerated degradation tests in non-working and working states and collects the degradation data of the index; according to the data obtained from the tests and the usage and environmental conditions included in the tests, a functional relationship between the degradation rate and the influencing factors of the liquid particle counting sensor in two stages is established; a two-stage degradation model based on the Wiener stochastic process is established, and the maximum likelihood estimation method is used for parameter estimation to obtain the global optimal solution of the parameters; according to the conventional storage temperature and the rated working flow rate conditions of the liquid particle counting sensor, the conventional degradation rates in the non-working and working states under actual conditions are extrapolated, and further according to the failure threshold of the deformation of the slit area inside the sensor, the prediction of the service life and reliability of the liquid particle counting sensor is realized.

Description

Technical Field

[0001] The present invention relates to a method for predicting the life of a liquid particle counting sensor, and particularly to a two-stage degradation modeling and life prediction method based on Wiener random process, which is applicable to accurately predicting the life of a liquid particle counting sensor and belongs to the field of life prediction of liquid particle counting sensors. Background Art

[0002] In many industries such as aviation, aerospace, shipbuilding, and power plants, there are high requirements for the detection of the contamination degree of oil. According to historical data, 60% - 70% of hydraulic system failures are caused by oil contamination, and the cleanliness of the oil is related to the safety and reliability of the entire system. Currently, liquid particle counters are the main testing equipment for oil contamination detection. As Figure 1 shown, they are used to test the diameter size and size distribution of solid impurities in the oil, and their weak and key component is the liquid particle counting sensor.

[0003] If the performance of the particle counting sensor fails or degrades, it will lead to deviations in control commands and indications, and ultimately cause catastrophic failures in the particle counter and even the entire hydraulic system. Therefore, it is crucial to accurately evaluate the reliability of the particle counting sensor, so as to replace the particle counting sensor at the appropriate time to avoid major system failures. Therefore, a degradation modeling and life method that fits the actual use situation and environmental factors of the liquid particle counting sensor needs to be proposed.

[0004] Existing research on degradation models usually assumes that the degradation process is controlled by a single process. The traditional single-stage degradation model is only applicable to the situation where the overall degradation process is stable with small fluctuations. However, the degradation characteristics of many products show obvious multi-stage characteristics, and the traditional single-stage degradation modeling method is not applicable. In the entire life cycle of the liquid particle counting sensor, its degradation process is jointly composed of the natural degradation process during non-operation and the wear process during operation. During the storage stage of the liquid particle counting sensor in the non-operating state, natural degradation will occur, and the natural degradation is manifested as a decrease in material strength, ultimately leading to the failure of the sensor's function. When the liquid particle counting sensor is in the operating state, when the measured oil passes through the sensor, the solid particles in the oil will cause damage to its internal slit. The gradual accumulation of damage will ultimately cause deformation of the structure around the slit, and when the deformation amount increases to exceed the limit value, it will lead to sensor failure. Since the degradation processes of the liquid particle counting sensor in the non-operating and operating states are very different, its overall degradation process will show obvious two-stage degradation. Using the traditional single-stage degradation modeling method to predict the life of the liquid particle counting sensor will result in a large deviation in the prediction results.

[0005] Meanwhile, the degradation model based on the stochastic process is more suitable for characterizing the degradation process. The degradation model introducing the stochastic process has both good fitting ability and statistical characteristics and has become the mainstream method in related research. The degradation models based on the Wiener process, gamma process, and inverse Gaussian distribution are the most widely used degradation modeling methods. Considering the randomness in the failure mechanism of the liquid particle counting sensor itself and the external environmental conditions, a degradation model based on the Wiener stochastic process is established, and the diffusion parameter included in the Wiener stochastic process is used to describe the fluctuation effect caused by factors such as the environment and manufacturing materials when the particle counting sensor is in two states.

[0006] Currently, the research on liquid particle counting sensors focuses on improving the structure and process to increase the lifespan, but there is very little research on the lifespan prediction of liquid particle counting sensors. Existing methods for degradation modeling and lifespan prediction of various sensors also focus on establishing single-stage degradation models, and the stochastic process is not introduced into the degradation models, ignoring the randomness in the failure mechanism of the sensors and the external environmental conditions. The existing methods do not conform to the actual situation of the full lifespan of liquid particle counting sensors and cannot achieve accurate prediction of the lifespan of liquid particle counting sensors. Summary of the Invention

[0007] A lifespan prediction method for a liquid particle counting sensor disclosed by the present invention obtains degradation data by conducting grouped constant stress accelerated degradation tests (CSADT). When establishing the degradation model, the characteristics and influencing factors of the degradation process of the sensor in the non-working state and working state are considered respectively. The deformation ratio of the slit area in the sensor is selected as the index parameter for measuring the degradation degree, and the functional relationship between the degradation rate and the influencing factors is established. At the same time, a stochastic process is introduced into the model to describe the fluctuation effect caused by uncertain factors and the differences in the degradation processes among individuals. On this basis, a two-stage degradation model is established based on the Wiener stochastic process, comprehensively considering the degradation influencing factors in the non-working state and working state of the sensor, being more in line with the actual situation of the degradation process of the full lifespan of the liquid particle counting sensor, and achieving efficient and accurate lifespan prediction of the liquid particle counting sensor.

[0008] The object of the present invention is achieved by the following technical solutions:

[0009] A method for predicting the life of a liquid particle counting sensor disclosed by the present invention uses the deformation ratio of the slit area in the sensor as an index parameter to measure the degree of degradation, conducts accelerated degradation tests in non-working and working states and collects the degradation data of the index; according to the data obtained from the tests and the usage and environmental conditions included in the tests, establishes the functional relationship between the degradation rate and the influencing factors of the liquid particle counting sensor in two stages; further establishes a two-stage degradation model based on the Wiener random process, and uses the maximum likelihood estimation method for parameter estimation to obtain the global optimal solution of the parameters; according to the conventional storage temperature and the rated working flow rate conditions of the liquid particle counting sensor, extrapolates the conventional degradation rates in the non-working and working states in actual situations, and further realizes the prediction of the life and reliability of the liquid particle counting sensor according to the failure threshold of the deformation of the slit area inside the sensor.

[0010] A method for predicting the life of a liquid particle counting sensor disclosed by the present invention includes the following steps:

[0011] Step 1: By analyzing the characteristics and influencing factors of the degradation process in the non-working and working states, select the deformation ratio of the slit area in the sensor as the index parameter to measure the degree of degradation, conduct accelerated degradation tests in the non-working and working states respectively, and collect the degradation data of the influence degree of the degradation influencing factors on the deformation ratio of the slit area in each state.

[0012] The implementation method of Step 1 is:

[0013] To quantify the degradation degree of the liquid particle counting sensor, by measuring the area of the slit inside the sensor, take the deformation ratio X(t) of the slit area as the degradation characteristic parameter, and select the degradation characteristic parameter X(t) as the index to characterize the degradation state of the liquid particle counting sensor:

[0014]

[0015] Among them, S(t) represents the slit area inside the sensor at time t, and S0 represents the initial area of the slit.

[0016] The liquid particle counting sensor undergoes natural degradation in the storage environment in the non-working state, which is manifested as the deformation of the slit area in the storage environment of the liquid particle counting sensor. The higher the storage temperature, the faster the strength of the slit material of the liquid particle counting sensor decreases, accelerating the change of the deformation ratio X(t) of the slit area of the liquid particle counting sensor.

[0017] On the other hand, the main form of degradation of the liquid particle counting sensor in the working state is wear. When the oil to be measured passes through the sensor, the solid particles in the oil will damage the internal slit of the liquid particle counting sensor. Temperature and flow rate are the main factors affecting the wear rate of the internal slit of the sensor in the working state. That is, the degradation process of the liquid particle counting sensor is affected by the temperature and flow rate of the oil to be measured. On the premise that other conditions of the oil to be measured remain unchanged, the higher the temperature and the larger the flow rate of the oil to be measured, the faster the wear rate of the slit, accelerating the change of the deformation ratio X(t) of the slit area of the liquid particle counting sensor.

[0018] Accelerated degradation tests of the liquid particle counting sensor are carried out separately in the non-working state and the working state, and degradation data on the degree of influence of degradation influencing factors on the deformation ratio of the slit area are collected. In the non-working state, the degradation influencing factor is the storage environment temperature, and in the working state, the degradation influencing factors include the temperature and flow rate of the oil to be measured.

[0019] Considering the controllability, operability of environmental factors during the implementation of the accelerated degradation test, and the sensitivity of the degradation rate to environmental factors. Preferably, single-stress accelerated degradation tests are carried out separately in the non-working state and the working state. The flow rate of the oil to be measured is used as the acceleration stress in the working state, and the storage environment temperature is used as the acceleration stress in the non-working state. The liquid particle counting sensor samples are grouped to carry out grouped constant stress accelerated degradation tests (CSADT). After the degradation test in the non-working state, degradation data on the degree of influence of the degradation influencing factor storage environment temperature on the deformation ratio of the slit area are collected. After the degradation test in the working state, degradation data on the degree of influence of the degradation influencing factor flow rate on the deformation ratio of the slit area are collected.

[0020] As a further preference, taking multiple particle counting sensors with consistent states as samples, the temperature of l groups of tests in the accelerated test in the non-working state is set as T1 < T2 < … < T l , all greater than the conventional storage temperature T0; the flow rates of l groups of tests in the accelerated degradation test in the working state are set as v1 < v2 < … < v l , all greater than the rated flow rate v0 of the oil. During the implementation of the accelerated degradation test, the internal slit area of the sensor of each group of test samples is collected at equal time intervals, and the accelerated degradation test data of each group of test samples are calculated according to formula (1). The accelerated degradation test data are the deformation ratio of the internal slit area of the sensor.

[0021] Step 2: Based on the characteristics and influencing factors of the degradation process in the non-operating state and the operating state analyzed in Step 1, divide the degradation process into two degradation stages in the non-operating state and the operating state. At the same time, introduce a stochastic process to describe the fluctuation effect caused by uncertain factors and the differences in the degradation processes among individuals. According to the degradation data on the degree of influence of the degradation influencing factors collected in Step 1 on the deformation ratio of the slit area, establish a functional relationship between the degradation rate and the influencing factors of the liquid particle counting sensor in the two stages. This functional relationship is the degradation rate function model in the two states.

[0022] Considering the controllability, operability of environmental factors when implementing the accelerated degradation test, and the sensitivity of the degradation rate to environmental factors, preferably, corresponding to the degradation test data obtained from the single-stress accelerated tests conducted in the non-operating and operating states respectively in Step 1.

[0023] The implementation method of Step 2 is as follows:

[0024] Step 2.1: Establish the degradation rate function model in the non-operating state.

[0025] The Arrhenius acceleration model describes the relationship between the degradation rate and the temperature T:

[0026]

[0027] where M is the degradation amount of a certain characteristic value, r Arrhenius represents the degradation rate at temperature T, k b is the Boltzmann constant, C0 is a parameter to be determined, and Ea is the activation energy of the reaction, in eV, which is a constant for the same failure mode of the same type of product.

[0028] The liquid particle counting sensor undergoes a natural degradation process under the storage conditions in the non-operating state. Based on the form of the above Arrhenius acceleration model, establish the degradation rate function model in the non-operating state. In the non-operating state, the storage temperature is used as the single acceleration stress, and other influencing factors are combined into one influencing factor. The degradation rate r1 in the non-operating state is expressed as:

[0029]

[0030] where T represents the storage environmental temperature in the non-operating state, and A and B are parameters to be estimated.

[0031] Step 2.2: Establish the degradation rate function model in the operating state.

[0032] Wear is an irreversible damage process on the material surface. The Archard model for the wear amount is:

[0033]

[0034] where r m represents the wear rate, W represents the wear amount, t represents the wear time, k represents the wear coefficient under certain working conditions, P and v represent the pressure and relative linear velocity on the friction surface, and m and n represent undetermined correlation coefficients.

[0035] The degradation of the liquid particle counting sensor in the working state is mainly manifested as wear. Based on the above Archard model form, a degradation rate function model in the working state is established. The flow rate of the measured oil through the slit is a single acceleration stress, and other influencing factors are combined into one influencing factor. The degradation rate r2 in the working state is expressed as:

[0036] r2 = C·(υ) D (3)

[0037] where υ represents the liquid flow rate in the working state, and C and D are parameters to be estimated.

[0038] Step 3: Based on the cumulative damage theory, accumulate the natural aging degradation effect in the non-working state and the degradation effect caused by wear in the working state. According to the functional relationship between the degradation rate and the influencing factor of the liquid particle counting sensor in the two stages established in Step 2, establish a two-stage degradation model of the liquid particle counting sensor based on the Wiener random process.

[0039] The implementation method of Step 3 is as follows:

[0040] The two-stage degradation process processed based on the cumulative damage theory has different degradation rates. Therefore, the two-stage Wiener processes in the two-stage degradation model of the liquid particle counting sensor have different drift coefficients, and the drift coefficient term is represented by the degradation rate function in Step 2. The diffusion parameter of the Wiener process characterizes the differences between individuals due to materials and process levels, and is used to describe the overall fluctuation characteristics of the degradation process. The diffusion parameters of the two-stage Wiener processes can be considered to be close.

[0041]

[0042] where r1 represents the natural degradation rate in the non-working state, r2 represents the degradation rate of wear in the working state; σ represents the fluctuation effect of the degradation process, and B(t) is a standard Brownian motion. Define the initial degradation amount X(0) = 0, and X(J) is the cumulative degradation amount corresponding to the cumulative working time t J as the initial degradation amount in the non-working stage.

[0043] According to the functional relationship between the degradation rate and the influencing factor of the liquid particle counting sensor in the two stages established in Step 2, the two-stage degradation model of the liquid particle counting sensor based on the Wiener random process is expressed as:

[0044]

[0045] Step 4: Estimate the parameters of the two-stage degradation model of the liquid particle counting sensor established in Step 3. Use the maximum likelihood estimation method for parameter estimation to obtain the global optimal solution of the parameters of the two-stage degradation model of the liquid particle counting sensor. The parameters include the parameters in the degradation rate function relationship and the diffusion parameter of the stochastic process.

[0046] Use the maximum likelihood estimation method to estimate the parameters of the two-stage degradation model of the liquid particle counting sensor and obtain the global optimal solution of the model parameters in one step. The implementation method of Step 4 is as follows:

[0047] The total sample size invested in the degradation test in Step 1 is n, the total number of tests for the area deformation ratio in the test is m times, the cumulative number of tests in the non-working state in the first stage is J times, and the cumulative number of tests in the working state in the second stage is (m - J) times. The test data obtained from the degradation test is expressed as:

[0048]

[0049] The degradation amount of sample i between the (j - 1)-th and j-th tests is denoted as ΔX(t ij -t i(j-1) ) = ΔX ij . Based on the independent increment property of the Wiener process, we can obtain:

[0050] ΔX ij ~N(r1·Δt j ,σ 2 ·Δt j ) 1 ≤ j ≤ J

[0051] ΔX ij ~N(r2·Δt j ,σ 2 ·Δt j ) J + 1 ≤ j ≤ m

[0052] In the entire life cycle of the particle counting sensor, when the ratio of the cumulative non-working time to the cumulative working time is known, according to the degradation data obtained in the two stages, use the maximum likelihood method to estimate the parameters in stages. The likelihood function is expressed as follows;

[0053]

[0054] Maximize the likelihood function L 1 (·), L 2 (·) to obtain the estimated values of parameters A, B, C, D, and σ;

[0055]

[0056]

[0057] Step 5: According to the parameters of the two-stage degradation model of the liquid particle counting sensor estimated in Step 4, extrapolate the degradation rates in the non-working and working states of the actual situation, and derive the failure distribution function based on the failure threshold of the deformation ratio of the internal slit area of the sensor, so as to achieve the accurate prediction of the reliability and life of the liquid particle counting sensor.

[0058] The implementation method of Step 5 is as follows:

[0059] According to the normal temperature condition during storage, extrapolate the degradation rate r of the particle counting sensor in the non-working state under actual conditions s` ; according to the rated flow rate of the measured oil fluid, extrapolate the degradation rate r of the particle counting sensor in the working state under actual conditions s2 .

[0060] The first time the area deformation ratio of the degradation state of the particle counting sensor exceeds the failure threshold W is regarded as the out-of-tolerance deformation of the internal slit, which will cause the sensor to fail. The corresponding time T w is regarded as the life of the particle counting sensor:

[0061] T w = inf{t≥t J :X(t)≥W}

[0062] R(t)=P(T w ≥t≥t J )=P(t≥t J :X(t)≤W) (7)

[0063] Based on the statistical characteristics of the two-stage Wiener random process, the time to first exceed the failure threshold follows a two-stage inverse Gaussian distribution. The derived failure distribution function is:

[0064]

[0065] Among them, represents the estimated value of the degradation amount at the turning moment accumulated during the entire working time of the particle counting sensor, that is,

[0066] Predict the life of the liquid particle counting sensor according to the failure distribution F(t). The median life t(0.5) of the particle counting sensor is the time corresponding to R(t)=1 - F(t)=0.5.

[0067] Beneficial effects:

[0068] 1. A method for predicting the life of a liquid particle counting sensor disclosed by the present invention considers the degradation influencing factors in both non-working and working states, and respectively establishes the functional relationship between the degradation rate and the storage temperature in the non-working state, and the functional relationship between the degradation rate and the rated flow rate of the measured oil in the working state. By establishing a reasonable functional relationship between the degradation rate and the actual storage environment and working conditions, the extrapolation of the degradation rate in the actual situation is made more accurate, achieving the purpose of accurately predicting the life of the particle counting sensor.

[0069] 2. A method for predicting the life of a liquid particle counting sensor disclosed by the present invention considers the differences in the degradation processes of the particle counting sensor in the non-working state and the working state. Based on the cumulative damage theory, a two-stage degradation model for non-working and working states is established. Compared with the single-stage degradation model, it is closer to the actual degradation law of the sensor, effectively improving the accuracy of the life prediction result.

[0070] 3. A method for predicting the life of a liquid particle counting sensor disclosed by the present invention uses a grouped constant acceleration degradation test to obtain degradation data, which can obtain a large amount of degradation data in a short time, improving the efficiency of the test and life prediction.

[0071] 4. A method for predicting the life of a liquid particle counting sensor disclosed by the present invention provides an evaluation method for predicting the life of the particle counting sensor, avoiding large deviations when testing the pollution degree level of the oil, and providing guarantee for the safety of the entire hydraulic system. Description of the Drawings

[0072] Figure 1 Working principle diagram of the particle counting sensor;

[0073] Figure 2 General flow schematic diagram of a method for predicting the life of a liquid particle counting sensor disclosed by the present invention;

[0074] Figure 3 Simulation results of the temperature acceleration degradation test in the non-working stage in the example of the present invention;

[0075] Figure 4 Simulation results of the flow acceleration degradation test in the working stage in the example of the present invention;

[0076] Figure 5 Simulation results of the two-stage degradation process under normal conditions in the example of the present invention;

[0077] Figure 6 Comparison diagram of the reliability prediction results in the example of the present invention. Detailed Embodiments

[0078] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention.

[0079] To verify the effectiveness and feasibility of the method, first, accelerated degradation simulation tests are carried out in two states: non-working and working. Further, based on the data of the accelerated degradation simulation tests, using the life prediction method of the liquid particle counting sensor disclosed in the present invention, a two-stage degradation model is established and a reliability prediction curve is obtained. Finally, the test results are compared and analyzed. On the one hand, by comparing with the K-M curve in the actual situation, the feasibility of this method is verified; on the other hand, by comparing with the reliability prediction curve based on the single-stage degradation model, it is verified that the method of the present invention can improve the accuracy of life prediction.

[0080] A life prediction method for a liquid particle counting sensor disclosed in this embodiment, the overall process of the method is as Figure 2 shown, and the embodiment includes the following steps:

[0081] Step 1: Conduct grouped accelerated degradation simulation tests in the non-working state and the working state, and collect the degradation data of the influence degree of the degradation influencing factors on the slit area deformation ratio in each state.

[0082] (1) Accelerated degradation simulation test in the non-working state

[0083] In the non-working state, taking temperature as the single acceleration stress, conduct a temperature acceleration test simulation in the non-working state. The acceleration test is divided into 3 groups according to the temperature gradient, and the set temperatures are 50°C, 60°C, and 70°C respectively. Each group has 4 samples, and the set simulation test duration is 36 time units. The other parameters in the degradation model and the degradation rate acceleration model are set as shown in Table 1. Collect the degradation data obtained from the non-working state simulation test. The 3 groups of randomly generated degradation data are as Figure 3 .

[0084] Table 1 Parameter settings for temperature acceleration test simulation in the non-working state

[0085] A B Sigma1 2000 -3000 0.1

[0086] (2) Accelerated degradation simulation test in the working state

[0087] In the working state, taking the oil flow rate as the single acceleration stress, conduct a flow acceleration test simulation in the working state. The acceleration test is divided into 3 groups according to the flow gradient, and the set acceleration flow rates are 60 mL / min, 70 mL / min, and 80 mL / min respectively. Each group has 4 samples, and the set simulation test duration is 36 units. The other parameters in the degradation model and the degradation rate acceleration model are set as shown in Table 2. Collect the degradation data obtained from the working state simulation test. The 3 groups of randomly generated degradation data are as Figure 4 .

[0088] Table 2 Simulation Parameter Settings for Flow Acceleration Test of Working State

[0089] C D Sigma2 0.1 0.5 0.1

[0090] Step 2: Establish the degradation rate function models of the liquid particle counter sensor in the non - working and working states.

[0091] (1) Establish the degradation rate function model in the non - working state

[0092] Establish the degradation rate function model in the non - working state. In the non - working state, the storage temperature is taken as the single acceleration stress, and other influencing factors are combined into one influencing factor. The degradation rate r1 in the non - working state is expressed as:

[0093]

[0094] where T represents the storage environment temperature in the non - working state, and A and B are parameters to be estimated.

[0095] (2) Establish the degradation rate function model in the working state

[0096] Establish the degradation rate function model in the working state. The flow rate of the measured oil through the slit is taken as the single acceleration stress, and other influencing factors are combined into one influencing factor. The degradation rate r2 in the working state is expressed as:

[0097] r2 = C·(υ) D (2)

[0098] where υ represents the liquid flow rate in the working state, and C and D are parameters to be estimated.

[0099] Step 3: Establish a two - stage degradation model of the non - working and working states based on the Wiener stochastic process.

[0100] According to the functional relationship between the degradation rate and the influencing factor of the liquid particle counter sensor established in Step 2 in the two - stage, establish a two - stage degradation model of the liquid particle counter sensor based on the Wiener stochastic process:

[0101]

[0102] where σ represents the fluctuation effect of the degradation process, B(t) is the standard Brownian motion; it is defined that the initial degradation amount X(0)=0, X(J) is the cumulative degradation amount corresponding to the cumulative working time t J as the initial degradation amount in the non - working stage.

[0103] Step 4: Use the maximum likelihood estimation method for parameter estimation to obtain the global optimal solution of the two - stage degradation model parameters of the liquid particle counter sensor.

[0104] The parameters of the degradation model and the degradation rate acceleration model are estimated by the maximum likelihood estimation method. According to the simulation test data, the maximum likelihood function is maximized to obtain the parameter estimation results, as shown in Table 3:

[0105] Table 3. Parameter Estimation Results

[0106] A B C D Sigma 1871 -2976 0.1020 0.4947 0.0966

[0107] Step 5: Extrapolate the degradation rates in the non-working and working states of the actual situation, and predict the reliability and life of the liquid particle counting sensor.

[0108] According to the conventional storage temperature and the rated oil flow rate conditions, the non-working degradation rate r s1 , the working degradation rate r s2 and the Brownian motion diffusion parameter σ s are calculated. When the ratio of the cumulative working time to the cumulative non-working time is 1 / 2, 10 samples of two-stage degradation data are generated with the cumulative non-working time of 200 units and the cumulative working time of 100 units in the simulation. The results are as Figure 5 shown.

[0109] Calculate the failure distribution function:

[0110]

[0111] where, represents the estimated value of the degradation amount at the turning moment accumulated during all the working time of the particle counting sensor, that is,

[0112] In this test, the failure threshold of the area deformation ratio is set as W = 40‰. According to the above failure distribution function, the reliability prediction result is further obtained, and the reliability result curve within 0 - 300 unit time is plotted. The result is the two-stage reliability prediction curve as shown in Figure 6 .

[0113] Step 6: Verify the effectiveness and accuracy of a life prediction method for a liquid particle counting sensor disclosed in this embodiment.

[0114] (1) Compare the results of a life prediction method for a liquid particle counting sensor disclosed in this embodiment with the K-M curve to verify the feasibility of the method

[0115] According to the time when the actual first exceeds the failure threshold in the simulated degradation data, the K-M comparison curve is plotted, as shown in Table 4 from small to large, and the K-M curve is further plotted. The result is the K-M curve as shown in Figure 6 .

[0116] Table 4. Sample Life Results

[0117] 245 245 247 248 249 251 253 255 256 263

[0118] (2) Establish a single-stage degradation model to obtain the single-stage reliability prediction results, and compare them with the results of a life prediction method for liquid particle counting sensors disclosed in this patent to verify the accuracy of the method.

[0119] Establish a single-stage degradation model. The degradation process of the particle counting sensor is stable, and the degradation rate is considered constant throughout the process, denoted by r0. The single-stage degradation model of the particle counting sensor is:

[0120] X(t) = X(0) + r0·t + σB(t) (5)

[0121] Based on the generated two-stage degradation data, estimate the parameters of the single-stage degradation model, and further obtain the single-stage reliability prediction results based on the single-stage degradation modeling method. Draw the result curve, and the result is shown in Figure 6 the single-stage reliability prediction curve in

[0122] Through Figure 6 the comparative analysis of the 3 curves in

[0123] It can be verified that the reliability evaluation curve obtained based on the two-stage degradation modeling method disclosed in this patent is very close to the K-M curve, verifying the feasibility and effectiveness of this method. At the same time, compared with the reliability evaluation results obtained by the single-stage degradation model, it is significantly closer to the K-M curve, verifying that this method has higher accuracy than the single-stage degradation modeling and life prediction methods. The above specific description further details the purpose, technical solution, and beneficial effects of the invention. It should be understood that the above are only specific embodiments of the present invention and are not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for predicting the lifespan of a liquid particle counting sensor, characterized in that: including the following steps: Step 1: By analyzing the characteristics of the degradation process and its influencing factors in the non-working state and working state, select the deformation ratio of the slit area inside the sensor as the index parameter to measure the degree of degradation. Conduct accelerated degradation tests in the non-working state and working state respectively, and collect the degradation data of the influence degree of the degradation influencing factors on the deformation ratio of the slit area in each state. When in the non-working state, the degradation influencing factor is the storage environment temperature, and when in the working state, the degradation influencing factors include the temperature and flow rate of the measured oil fluid; Step 2: Based on the characteristics of the degradation process and its influencing factors analyzed in Step 1 in the non-working state and working state, divide the degradation process into two degradation stages in the non-working state and working state. At the same time, introduce a stochastic process to describe the fluctuation effect caused by uncertain factors and the differences in the degradation processes among individuals. According to the degradation data of the influence degree of the degradation influencing factors on the deformation ratio of the slit area collected in Step 1, establish the functional relationship between the degradation rate and the influencing factors of the liquid particle counter sensor in the two stages. The functional relationship is the degradation rate function model in the two states; Step 3: Based on the cumulative damage theory, accumulate the natural aging degradation effect in the non-working state and the degradation effect caused by wear in the working state. According to the functional relationship between the degradation rate and the influencing factors of the liquid particle counter sensor in the two stages established in Step 2, establish a two-stage degradation model of the particle counter sensor based on the Wiener stochastic process; Step 4: Estimate the parameters of the two-stage degradation model of the particle counter sensor established in Step 3, and use the maximum likelihood estimation method for parameter estimation to obtain the global optimal solution of the parameters of the two-stage degradation model of the particle counter sensor. The parameters included are the parameters in the degradation rate function relationship and the diffusion parameters of the stochastic process; Step 5: According to the parameters of the two-stage degradation model of the particle counter sensor estimated in Step 4, extrapolate the conventional degradation rates in the non-working and working states of the actual situation. According to the failure threshold of the deformation ratio of the slit area inside the sensor, deduce the failure distribution function to achieve the accurate prediction of the reliability and life of the liquid particle counter sensor.

2. The life prediction method for a liquid particle counting sensor according to claim 1, characterized in that: The implementation method of Step 1 is as follows: To quantify the degradation degree of the liquid particle counter sensor, by measuring the area of the slit inside the sensor, use the deformation ratio X(t) of the slit area as the degradation characteristic parameter, and select the degradation characteristic parameter X(t) as the index to characterize the degradation state of the liquid particle counter sensor: where S(t) represents the area of the slit inside the sensor at time t, and S0 represents the initial area of the slit; The liquid particle counter sensor undergoes natural degradation in the storage environment in the non-working state, manifested as the deformation of the slit area of the liquid particle counter sensor in the storage environment. The higher the storage temperature, the faster the strength of the slit material of the liquid particle counter sensor decreases, accelerating the change of the deformation ratio X(t) of the slit area of the liquid particle counter sensor; When the oil to be measured passes through the sensor, solid particles in the oil will damage the internal slit of the liquid particle counting sensor. That is, the degradation process of the liquid particle counting sensor is affected by the temperature and flow rate of the oil to be measured. Under the premise that other conditions of the oil to be measured remain unchanged, the higher the temperature and the greater the flow rate of the oil to be measured, the faster the wear rate of the slit, accelerating the change of the deformation ratio X(t) of the slit area of the liquid particle counting sensor. Accelerated degradation tests of the liquid particle counting sensor are carried out separately in the non-working state and the working state, and degradation data on the influence degree of degradation influencing factors on the deformation ratio of the slit area are collected. When in the non-working state, the degradation influencing factor is the storage environment temperature, and when in the working state, the degradation influencing factors include the temperature and flow rate of the oil to be measured.

3. A method for predicting the lifespan of a liquid particle counting sensor according to claim 2, characterized in that: Considering the controllability, operability of environmental factors during the implementation of the accelerated degradation test, and the sensitivity of the degradation rate to environmental factors; single-stress accelerated degradation tests are carried out separately in the non-working state and the working state. The flow rate of the oil to be measured is used as the acceleration stress in the working state, and the storage environment temperature is used as the acceleration stress in the non-working state. The liquid particle counting sensor samples are grouped to carry out the grouped constant accelerated degradation test (CSADT); after the degradation test in the non-working state, degradation data on the influence degree of the storage environment temperature, the degradation influencing factor, on the deformation ratio of the slit area are collected, and after the degradation test in the working state, degradation data on the influence degree of the oil flow rate, the degradation influencing factor, on the deformation ratio of the slit area are collected.

4. A method for predicting the lifespan of a liquid particle counting sensor according to claim 3, characterized in that: Taking multiple particle counting sensors with consistent states as samples, the temperature of l groups of tests in the acceleration test under the non-working state is set as T1 < T2 < … < T l , all of which are greater than the conventional storage temperature T0; the flow rates of l groups of tests in the accelerated degradation test under the working state are set as v1 < v2 < … < v l , all of which are greater than the rated oil flow rate v0; during the implementation of the accelerated degradation test, the slit area inside the sensors of each group of test samples is collected at equal time intervals, and the accelerated degradation test data of each group of test samples are calculated according to formula (1), and the accelerated degradation test data are the deformation ratio of the slit area inside the sensors.

5. A method for predicting the lifespan of a liquid particle counting sensor according to claim 1, 2, 3 or 4, characterized in that: Considering the controllability and operability of environmental factors during the implementation of the accelerated degradation test, and the sensitivity of the degradation rate to environmental factors, for the degradation test data obtained from the single-stress acceleration tests carried out separately in the non-working and working states in step one, the implementation method in step two is as follows. Step 2.1: Establish a degradation rate function model in the non-working state. The Arrhenius acceleration model describes the relationship between the degradation rate and the temperature T: where M is the degradation amount of a certain eigenvalue, r Arrhenius represents the degradation rate at temperature T, k b is the Boltzmann constant, C0 is a parameter to be determined, Ea is the activation energy of the reaction, in eV, and is a constant for the same failure mode of the same type of product; The particle counting sensor undergoes a natural degradation process under the storage conditions in the non-working state. Based on the form of the above Arrhenius acceleration model, a degradation rate function model in the non-working state is established. The storage temperature is used as the single acceleration stress in the non-working state, and other influencing factors are combined into one influencing factor. The degradation rate r1 in the non-working state is expressed as: where T represents the storage environment temperature in the non-working state, and A and B are parameters to be estimated. Step 2.2: Establish a degradation rate function model in the working state. Wear is an irreversible damage process on the material surface, and the Archard model for the wear amount is: where r m represents the wear rate, W represents the wear amount, t represents the wear time, k represents the wear coefficient under certain working conditions, P and v represent the pressure and relative linear velocity on the friction surface, and m and n represent undetermined correlation coefficients; The degradation of the particle counting sensor in the working state is mainly manifested as wear. Based on the form of the above Archard model, a degradation rate function model in the working state is established. The flow rate of the oil to be measured passing through the slit is used as the single acceleration stress, and other influencing factors are combined into one influencing factor. The degradation rate r2 in the working state is expressed as: r2 = C·(υ) D (3) where υ represents the liquid flow rate in the working state, and C and D are parameters to be estimated.

6. A method for predicting the lifespan of a liquid particle counting sensor according to claim 5, characterized in that: The implementation method in step three is as follows. The two-stage degradation process after being processed based on the cumulative damage theory has different degradation rates. Therefore, the two-stage Wiener processes in the two-stage degradation model of the particle counting sensor have different drift coefficients, and the drift coefficient term is represented by the degradation rate function in Step 2; the diffusion parameter of the Wiener process characterizes the differences among individuals due to materials and process levels, and is used to describe the overall fluctuation characteristics of the degradation process. The diffusion parameters of the two-stage Wiener process can be considered to be close; Among them, r1 represents the natural degradation rate in the non-operating state, and r2 represents the wear degradation rate in the operating state; σ represents the fluctuation effect of the degradation process, and B(t) is a standard Brownian motion; the initial degradation amount X(0) = 0 is defined, and X(J) is the cumulative degradation amount corresponding to the cumulative working time t, which serves as the initial degradation amount in the non-working stage. J The corresponding cumulative degradation amount is used as the initial degradation amount in the non-working stage. Based on the functional relationship between the degradation rate and the influencing factors of the liquid particle counting sensor established in Step 2, a two-stage degradation model of the particle counting sensor is established based on the Wiener random process and is expressed as: 。 7. A method for predicting the lifespan of a liquid particle counting sensor according to claim 5, characterized in that: The maximum likelihood estimation method is used to estimate the parameters of the two-stage degradation model of the particle counting sensor, and the global optimal solution of the model parameters is obtained in one step. The implementation method of Step 4 is as follows: In Step 1, the total number of samples invested in the degradation test is n, and the total number of tests for the area deformation ratio in the test is m times. The cumulative number of tests in the non-working state in the first stage is J times, and the cumulative number of tests in the working state in the second stage is (m - J) times. The test data obtained from the degradation test are expressed as: The degradation amount of sample i between the (j - 1)-th and j-th tests is denoted as ΔX(t ij - t i(j-1) ) = ΔX ij , and based on the independent increment property of the Wiener process, we can obtain: ΔX ij ~N(r1·Δt j ,σ 2 ·Δt j )1≤j≤J ΔX ij ~N(r2·Δt j ,σ 2 ·Δt j )J+1≤j≤m In the entire life cycle of the particle counting sensor, when the ratio of the cumulative time in the non-working state to the cumulative time in the working state is known, based on the degradation data obtained in the two stages, the maximum likelihood method is used to estimate the parameters in stages, and the likelihood function is expressed as follows; Maximize the likelihood function L 1 (·), L 2 (·) to obtain the estimated values of parameters A, B, C, D, and σ; 8. A method for predicting the lifespan of a liquid particle counting sensor according to claim 7, characterized in that: The implementation method of Step 5 is as follows: Extrapolate the degradation rate r of the particle count sensor in the non-operating state according to the normal temperature conditions during actual storage s` ; extrapolate the degradation rate r of the particle count sensor in the operating state according to the rated flow rate of the measured oil under normal conditions s2 ; When the area deformation ratio of the degradation state of the particle counting sensor first exceeds the failure threshold W, it is regarded as the out-of-tolerance of the internal slit deformation, which will cause the sensor to malfunction, and the corresponding time T w is regarded as the lifespan of the particle counting sensor: T w = inf{t ≥ t J : X(t) ≥ W} R(t) = P(T w ≥ t ≥ t J ) = P(t ≥ t J : X(t) ≤ W) (7) Based on the statistical characteristics of the two-stage Wiener random process, the time to first exceed the failure threshold follows a two-stage inverse Gaussian distribution, and the derived failure distribution function is: Among them, represents the estimated degradation amount of the turning moment accumulated by the particle counting sensor during all working hours, that is, Based on the failure distribution F(t), the life of the liquid particle counting sensor is predicted. The median life t(0.5) of the particle counting sensor is the time corresponding to R(t) = 1 - F(t) = 0.5.

Images