Relay protection device evaluation method and system based on GARCH model

Abstract

The invention discloses a relay protection device evaluation method and system based on a GARCH model, and the method comprises the steps: collecting the operation data of a relay protection device, and calculating the defect rate. And performing parameter estimation on the defect rate, and calculating a residual sequence. And checking whether the residual sequence is stable or not, and establishing a reliability evaluation model based on a GARCH model. Future data is predicted according to the model, and the reliable service life of the device is obtained through device reliability calculation. According to the relay protection device evaluation method and system based on the GARCH model, the GARCH model is adopted, system noise in a relay protection device reliability evaluation model is subjected to autoregressive condition heterovariance analysis and fed back to the evaluation model, the evaluation model is updated in time along with updating of acquired defect rate data, and the reliability of the relay protection device is improved. And the precision of predicting the future defect rate and the residual reliable use time of the device is improved. And the value of T is changed along with the continuous improvement of sample data, so that the prediction result is more accurate.

Description

Technical Field

[0001] The present invention relates to the technical field, and in particular to a relay protection device evaluation method and system based on a GARCH model. Background Art

[0002] As the efficiency of modern power grid systems continues to improve, the complexity and uncertainty of the system are increasing. Relay protection devices are the first line of defense to ensure the safety of power grid systems. Their refusal to operate or malfunction may not be able to curb the deterioration of system failures, and may even increase the losses caused. Therefore, the research on their reliability has become an important part of ensuring the safety of power grid systems.

[0003] The analytical method is widely used to evaluate the reliability of relay protection devices, that is, to establish a reliability model and calculate parameter indicators. There is a technology that uses the failure rate of small sample data to expand the sample, estimates the parameters of the failure rate model, and obtains the reliable service life; there is a technology that uses the actual operating status of the relay protection equipment as a basis and the corresponding status evaluation standards as a reference, and uses the health index to quantify the quality of each status of the equipment; there is a technology that establishes a life model based on the log-normal distribution, and then uses the modified maximum likelihood function method to identify different models for the truncated life data, and uses the least squares method and the average rank method for parameter estimation, and finally completes the life estimation of the relay protection device.

[0004] The above technologies are all analyzed and predicted based on the actual reliability data of existing devices. In the process of acquiring actual data, it is inevitable that there are systematic errors and measurement errors, which leads to reduced prediction accuracy. In addition, ignoring the existence of variance in the model construction process will lead to a serious underestimation of the variance of the residual, resulting in "pseudo-invalidation" errors in the parameter significance test, which will affect the accuracy of the model. Summary of the invention

[0005] In view of the above-mentioned problems, the present invention is proposed.

[0006] Therefore, the technical problem solved by the present invention is that the existing technology inevitably has systematic errors and measurement errors, which leads to reduced prediction accuracy. In addition, ignoring the existence of variance in the model construction process will lead to a serious underestimation of the variance of the residual, resulting in "false positives" errors in the parameter significance test, which affects the accuracy of the model.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions: a relay protection device evaluation method based on the GARCH model, comprising: collecting relay protection device operation data and calculating the defect rate.

[0008] Perform parameter estimation on the defect rate and calculate the residual sequence.

[0009] Test whether the residual sequence is stationary and establish a reliability assessment model based on the GARCH model.

[0010] Based on the model, future data is predicted and the reliable service life of the device is obtained through device reliability calculation.

[0011] As a preferred solution of the relay protection device evaluation method based on the GARCH model described in the present invention, wherein: the acquisition of relay protection device operation data includes acquiring the relay protection device operation data of the operational site, eliminating abnormal data, and sampling at the same time interval. The acquired data is divided into two parts by time T.

[0012] The function λ(t) of the defect rate of the relay protection device over the whole life cycle changing with time is expressed as:

[0013]

[0014] Among them, a represents the scale parameter and b represents the shape parameter.

[0015] The device reliability function is expressed as:

[0016]

[0017] Where R(t) represents the device reliability.

[0018] As a preferred solution of the relay protection device evaluation method based on the GARCH model described in the present invention, the calculation of the defect rate includes preprocessing the function λ(t) of the defect rate of the relay protection device over the entire life cycle over time, which is expressed as:

[0019] λ(t)=e m ·t n

[0020] Take the logarithm of both sides of the preprocessed formula and convert it into a linear form, expressed as:

[0021] ln[λ(t)]=m+nlnt

[0022] Combined with the function λ(t) of the defect rate of the relay protection device over the entire life cycle, the values ​​of the scale parameter and shape parameter are obtained, which are expressed as:

[0023]

[0024] b=n+1

[0025] Among them, m represents the coefficient related to the scale parameter a in the defect rate function and obtained by fitting, and n represents the coefficient related to the shape parameter b in the defect rate function and obtained by fitting.

[0026] As a preferred solution of the relay protection device evaluation method based on the GARCH model described in the present invention, wherein: the parameter estimation of the defect rate and the calculation of the residual sequence include that the average defect rate data of the relay protection device is a set of non-stationary time series. Any time series {Xt} is decomposed into a superposition of two parts, one part is a deterministic trend analysis determined by a polynomial, and the other part is a stable zero mean error component, expressed as:

[0027]

[0028] Among them, X t represents observations or time series data, Y t represents the trend component, z t represents the error term, μ j represents the constant coefficient, j represents the order, and w t represents a zero-mean white noise sequence, Represents a delay operator.

[0029] The {ln[λ(t)]} sequence is divided into two parts. The deterministic trend part is a polynomial model about time. The error component uses a heteroscedastic model. The model of ln[λ(t)] about lnt is expressed as:

[0030]

[0031] Among them, u t represents an independent noise sequence with mean zero.

[0032] As a preferred solution of the relay protection device evaluation method based on the GARCH model described in the present invention, wherein: the test of whether the residual sequence is stable and the establishment of a reliability evaluation model based on the GARCH model include testing whether the residual sequence is stable through the augmented Dickey-Fuller test method.

[0033] The augmented Dickey-Fuller test model is expressed as:

[0034]

[0035] Among them, Δy t Represents y t The first-order difference of t =y t -y t-1 α represents a constant term, β t Represents the time trend term. t-1 Represents the independent variable term lagged by one period. i Δy t-i represents the lagged difference term, controlling autocorrelation. ∈ t represents the error term.

[0036] The original hypothesis H0 is that the time series has a unit root and the series is non-stationary γ=0.

[0037] The alternative hypothesis H1 is that the time series does not have a unit root and the series is stationary γ<0.

[0038] The t statistic for testing γ is compared with the critical value and is expressed as:

[0039]

[0040] Where γ^ represents the estimated γ and SE(γ^) represents the standard error of γ^.

[0041] The calculated t γ Compare with the critical value in the ADF test table. γ If it is less than 5% of the critical value, the null hypothesis is rejected and the series is considered to be stable. Otherwise, the series is not stable.

[0042] As a preferred solution of the relay protection device evaluation method based on the GARCH model described in the present invention, wherein: the establishment of a reliability evaluation model based on the GARCH model includes: t A conditional autoregressive variance model is established, and the evaluation model is composed of ln[λ(t)] with respect to lnt, which is expressed as:

[0043]

[0044] The constraints satisfied by the coefficients are expressed as:

[0045]

[0046] Among them, σ t 2 represents the volatility of the time series, α0 represents the constant term, α i u represents the ARCH term coefficient, which indicates the influence of the residual square term of the lag i period. t-i 2 represents the residual square of the lag i period, β j Represents the GARCH term coefficient, which indicates the influence of the conditional variance of the lag j period. t-j 2 represents the conditional variance at lag j.

[0047] The variance of a noise at the tth moment is a quantity that changes with time, which is conditional heteroscedasticity. This variance that changes with time is a linear combination of the squares of a finite number of noise values ​​in the past, which is an autoregressive function and constitutes a regression conditional heteroscedasticity model. The variance of the noise at the tth moment is a combination of the squares of several noise values ​​in the past. If the variance of the noise increases at the previous moments, the variance of the noise at the tth moment will increase accordingly. If the variance of the noise decreases at the previous moments, the variance of the noise at the tth moment will also decrease accordingly.

[0048] As a preferred solution of the relay protection device evaluation method based on the GARCH model of the present invention, wherein: the predicting of future data according to the model and obtaining the reliable service life of the device by calculating the device reliability include using the GARCH model and estimating the model parameter μ using historical data j , α i , β j .

[0049] Calculate the residuals after model fitting and perform ADF test on the residual sequence to ensure that the residual sequence is stable.

[0050] The GARCH model is used to model the residual series and obtain the conditional variance.

[0051] Use the fitted model to predict future time series data and predict future conditional variance based on the GARCH model.

[0052] Determine the failure threshold based on the device's usage requirements and historical failure data. Calculate the reliability of each time point using the predicted time series data and conditional variance.

[0053] The reliability threshold is determined to be 95% as the criterion for determining the reliable service life.

[0054] According to the reliability calculation results, find the time point T when the reliability reaches the set standard life , this time point is the reliable service life of the device.

[0055] A relay protection device evaluation system based on a GARCH model, characterized in that: it includes:

[0056] The calculation module collects the operation data of the relay protection device and calculates the defect rate. It estimates the parameters of the defect rate and calculates the residual sequence.

[0057] Establish a model module, test whether the residual sequence is stable, and establish a reliability assessment model based on the GARCH model.

[0058] The model prediction module predicts future data based on the model and obtains the reliable service life of the device through device reliability calculation.

[0059] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above method when executing the computer program.

[0060] A computer-readable storage medium stores a computer program, which implements the steps of the method described above when executed by a processor.

[0061] The beneficial effects of the present invention are as follows: the GARCH model is used to perform autoregressive conditional heteroscedastic analysis on the system noise in the reliability evaluation model of the relay protection device, and the system noise is fed back into the evaluation model. With the update of the defect rate data, the evaluation model is updated in time, which improves the accuracy of the prediction of the future defect rate and the remaining reliable use time of the device. And with the continuous improvement of the sample data, the value of T is changed, which can make the prediction result more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0063] Figure 1 An overall flow chart of a relay protection device evaluation method based on a GARCH model provided for the first embodiment of the present invention.

[0064] Figure 2 A fitting diagram of data before time T of a relay protection device evaluation method based on a GARCH model provided in the second embodiment of the present invention.

[0065] Figure 3 A fitting diagram of original and predicted data of a relay protection device evaluation method based on a GARCH model provided in the second embodiment of the present invention. DETAILED DESCRIPTION

[0066] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0067] Example 1, reference Figure 1, as an embodiment of the present invention, provides a relay protection device evaluation method based on a GARCH model, comprising:

[0068] S1: Collect the operating data of the relay protection device and calculate the defect rate.

[0069] Collect the operation data of the relay protection device at the operation site, remove the abnormal data, and sample at the same time interval. The acquired data is divided into two parts by time T.

[0070] The function λ(t) of the defect rate of the relay protection device over the whole life cycle changing with time is expressed as:

[0071]

[0072] Among them, a represents the scale parameter and b represents the shape parameter.

[0073] The device reliability function is expressed as:

[0074]

[0075] Where R(t) represents the device reliability.

[0076] The function λ(t) of the defect rate of the relay protection device over the whole life cycle changing with time is preprocessed and expressed as:

[0077] λ(t)=e m ·t n

[0078] Take the logarithm of both sides of the preprocessed formula and convert it into a linear form, expressed as:

[0079] ln[λ(t)]=m+nlnt

[0080] Combined with the function λ(t) of the defect rate of the relay protection device over the entire life cycle, the values ​​of the scale parameter and shape parameter are obtained, which are expressed as:

[0081]

[0082] b=n+1

[0083] Among them, m represents the coefficient related to the scale parameter a in the defect rate function and obtained by fitting, and n represents the coefficient related to the shape parameter b in the defect rate function and obtained by fitting.

[0084] S2: Estimate the parameters of the defect rate and calculate the residual sequence.

[0085] It should be noted that the average defect rate data of relay protection devices is a set of non-stationary time series. According to Cramer decomposition theorem, any time series {Xt} can be decomposed into the superposition of two parts: one part is the deterministic trend analysis determined by the polynomial, and the other part is the stable zero mean error component, which can be expressed as:

[0086]

[0087] Among them, X t represents observations or time series data, Y t represents the trend component, z t represents the error term, μ j represents the constant coefficient, j represents the order, and w t represents a zero-mean white noise sequence, Represents a delay operator.

[0088] The {ln[λ(t)} sequence is divided into two parts. The deterministic trend part is a polynomial model about time. The error component uses a heteroscedastic model. The model of ln[λ(t)] about lnt is established as:

[0089]

[0090] Among them, u t represents an independent noise sequence with mean zero.

[0091] S3: Check whether the residual sequence is stationary and establish a reliability assessment model based on the GARCH model.

[0092] The augmented Dickey-Fuller test is used to test whether the residual series is stationary.

[0093] The augmented Dickey-Fuller test model is expressed as:

[0094]

[0095] Among them, Δy t Represents y t The first-order difference of t =y t -y t-1 α represents a constant term, β t Represents the time trend term. t-1 Represents the independent variable term lagged by one period. i Δy t-i represents the lagged difference term, controlling autocorrelation. ∈ t represents the error term.

[0096] The original hypothesis H0 is that the time series has a unit root and the series is non-stationary γ=0.

[0097] The alternative hypothesis H1 is that the time series does not have a unit root and the series is stationary γ<0.

[0098] The t statistic for testing γ is compared with the critical value and is expressed as:

[0099]

[0100] Where γ^ represents the estimated γ and SE(γ^) represents the standard error of γ^.

[0101] The calculated t γ Compare with the critical value in the ADF test table. γ If it is less than 5% of the critical value, the null hypothesis is rejected and the series is considered to be stable. Otherwise, the series is not stable.

[0102] It should be noted that in order to estimate the heteroscedastic function more accurately, when establishing a time series model, the residual sequence is required to satisfy zero mean, pure randomness and variance homogeneity. Zero mean and pure randomness are easy to achieve. Variance homogeneity is homoscedasticity, which means that the residual sequence has the same variance. If the variance of the residual does not satisfy homoscedasticity, that is, the variance of the residual is not a constant, the regression model is said to have heteroscedasticity.

[0103] For the error component u t A conditional autoregressive variance model is established, and the evaluation model is composed of ln[λ(t)] with respect to lnt, which is expressed as:

[0104]

[0105] The constraints satisfied by the coefficients are expressed as:

[0106]

[0107] Among them, σ t 2 represents the volatility of the time series, α0 represents the constant term, α i u represents the ARCH term coefficient, which indicates the influence of the residual square term of the lag i period. t-i 2 represents the residual square of the lag i period, β j Represents the GARCH term coefficient, which indicates the influence of the conditional variance of the lag j period. t-j 2 represents the conditional variance at lag j.

[0108] It should be noted that the variance of a noise at the tth moment is a quantity that changes with time, which is conditional heteroscedasticity, and this variance that changes with time is a linear combination of the squares of a finite number of noise values ​​in the past, which is an autoregressive function and constitutes a regression conditional heteroscedasticity model. The variance of the noise at the tth moment is a combination of the squares of several noise values ​​in the past. If the variance of the noise increases at the previous moments, the variance of the noise at the tth moment will increase accordingly. If the variance of the noise decreases at the previous moments, the variance of the noise at the tth moment will also decrease accordingly.

[0109] S4: Predict future data based on the model and calculate the device reliability to obtain the device's reliable service life.

[0110] Use the GARCH model to estimate the model parameters μ using historical data j , α i , β j .

[0111] Calculate the residuals after model fitting and perform ADF test on the residual sequence to ensure that the residual sequence is stable.

[0112] The GARCH model is used to model the residual series and obtain the conditional variance.

[0113] Use the fitted model to predict future time series data and predict future conditional variance based on the GARCH model.

[0114] Determine the failure threshold based on the device's usage requirements and historical failure data. Calculate the reliability of each time point using the predicted time series data and conditional variance.

[0115] The reliability threshold is determined to be 95% as the criterion for determining the reliable service life.

[0116] According to the reliability calculation results, find the time point T when the reliability reaches the set standard life , this time point is the reliable service life of the device.

[0117] The above embodiments also include a relay protection device evaluation system based on the GARCH model, specifically:

[0118] The calculation module collects the operation data of the relay protection device and calculates the defect rate. It estimates the parameters of the defect rate and calculates the residual sequence.

[0119] Establish a model module, test whether the residual sequence is stable, and establish a reliability assessment model based on the GARCH model.

[0120] The model prediction module predicts future data based on the model and obtains the reliable service life of the device through device reliability calculation.

[0121] The computer device may be a server. The computer device includes a processor, a memory, an input / output interface (I / O for short) and a communication interface. The processor, the memory and the input / output interface are connected via a system bus, and the communication interface is connected to the system bus via the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data cluster data of a power monitoring system. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a method is implemented.

[0122] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory may include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0123] Example 2, an embodiment of the present invention, provides a relay protection device evaluation method and system based on the GARCH model. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through simulation experiments.

[0124] The defect rates of relay protection devices with different service years at 220KV voltage level in a certain area are shown in Table 1.

[0125] Table 1 Defect rate

[0126]

[0127]

[0128] To ensure the stability of the residual sequence, the time T is set to 8, and the first 16 data in Table 1 are fitted using MATLAB software.

[0129] The fitting curve is as follows Figure 2 As shown, the curve formula is:

[0130] ln[λ(t)]=-3.9857-0.0211lnt

[0131] It can be concluded that m = -3.9857, n = -0.0221. The fitting residuals are calculated to meet the critical value requirement of 5%, which is a stable series.

[0132] A conditional heteroskedasticity model is established. After experiments, the GARCH (1, 2) model is determined to be the optimal model. The stability of this model is tested. The reliability assessment model is constructed together with the ln [λ (t)] model. The data set after time T obtained by the model is obtained. The data fitting curve is as follows: Figure 3 The curve formula is:

[0133] ln[λ(t)]=-4.6241+0.5332lnt

[0134] It can be concluded that m = -4.5997, n = 0.5413. From formula 5, it can be concluded that a = 26.2895, b = 1.5413. According to the reliability threshold of the relay protection device given by expert experience is 0.735, and the reliable service life t of the device is calculated to be 12.189 years.

[0135] If we directly perform curve fitting on all data samples, we can get m = -4.6241, n = 0.5332. From formula 5, we can get a = 26.9656, b = 1.5332. According to the reliability threshold of the relay protection device given by expert experience is 0.735, and the reliable service life t of the device is calculated to be 12.513 years, which is recorded as the reference life t0.

[0136] Take different time T, fit the data, calculate the residual, establish the conditional heteroscedasticity model, and finally establish the evaluation model to calculate the reliable service life. The results are shown in Table 2. It can be seen that as time T gradually increases, the error between the obtained reliable service life and the reference life t0 becomes smaller and smaller, and the error is the smallest when T is 12.

[0137] Table 2 Reliable life

[0138]

[0139] To verify the effectiveness of this patented method, the first 24 defect rate sample data were directly fitted using the least squares method, and the reliability life t was obtained to be 13.221 years, with a relative error of 5.7% with the reference life t0. In comparison, the evaluation model that also used the first 24 defect rate sample data for prediction had a smaller error.

[0140] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A relay protection device evaluation method based on GARCH model, characterized in that: include: Collect the operation data of relay protection devices and calculate the defect rate; Estimate the parameters of the defect rate and calculate the residual sequence; Check whether the residual sequence is stable and establish a reliability assessment model based on the GARCH model; Based on the model to predict future data and calculate the device reliability, the reliable service life of the device can be obtained.

2. The relay protection device evaluation method based on the GARCH model according to claim 1, characterized in that: The collecting of the relay protection device operation data includes collecting the relay protection device operation data at the operation site, eliminating abnormal data, and sampling at the same time interval; the acquired data is divided into two parts by time T; The function λ(t) of the defect rate of the relay protection device over the whole life cycle changing with time is expressed as: Among them, a represents the scale parameter and b represents the shape parameter; The device reliability function is expressed as: Where R(t) represents the device reliability.

3. The relay protection device evaluation method based on the GARCH model according to claim 2, characterized in that: The calculation of the defect rate includes preprocessing the function λ(t) of the defect rate of the relay protection device over the entire life cycle over time, which is expressed as: λ(t)=e m ·t n Take the logarithm of both sides of the preprocessed formula and convert it into a linear form, expressed as: ln[λ(t)]=m+nlnt Combined with the function λ(t) of the defect rate of the relay protection device over the entire life cycle, the values ​​of the scale parameter and shape parameter are obtained, which are expressed as: b=n+1 Among them, m represents the coefficient related to the scale parameter a in the defect rate function and obtained by fitting, and n represents the coefficient related to the shape parameter b in the defect rate function and obtained by fitting.

4. The relay protection device evaluation method based on the GARCH model according to claim 3, characterized in that: The parameter estimation of the defect rate and the calculation of the residual sequence include that the average defect rate data of the relay protection device is a set of non-stationary time series; any time series {Xt} is decomposed into the superposition of two parts, one part is the deterministic trend analysis determined by the polynomial, and the other part is the stable zero mean error component, which is expressed as: Among them, X t represents observations or time series data, Y t represents the trend component, z t represents the error term, μ j represents the constant coefficient, j represents the order, and w t represents a zero-mean white noise sequence, represents the delay operator; The {ln[λ(t)]} sequence is divided into two parts. The deterministic trend part establishes a polynomial model about time; the error component uses a heteroscedastic model; the model of ln[λ(t)] about lnt is established as: Among them, u t represents an independent noise sequence with mean zero.

5. The relay protection device evaluation method based on the GARCH model according to claim 4, characterized in that: The testing of whether the residual sequence is stable and establishing a reliability assessment model based on the GARCH model include testing whether the residual sequence is stable by using an augmented Dickey-Fuller test method; The augmented Dickey-Fuller test model is expressed as: Among them, Δy t Represents y t The first-order difference of t =y t -y t-1 ; α represents a constant term, β t represents the time trend term; γy t-1 represents the independent variable term lagged by one period; δ i Δy t-i represents the lagged difference term, controlling autocorrelation; ∈ t represents the error term; The original hypothesis H0 is that the time series has a unit root and the series is non-stationary γ = 0; The alternative hypothesis H1 is that the time series does not have a unit root and the series is stationary γ<0; The t statistic for testing γ is compared with the critical value and is expressed as: Where γ^ represents the estimated γ, SE(γ^) represents the standard error of γ^; The calculated t γ Compare with the critical value in the ADF test table; if t γ If it is less than 5% of the critical value, the null hypothesis is rejected and the sequence is considered to be stationary; otherwise, the sequence is not stationary.

6. The relay protection device evaluation method based on the GARCH model according to claim 5, characterized in that: The reliability evaluation model based on the GARCH model is established by: t A conditional autoregressive variance model is established, and the evaluation model is composed of ln[λ(t)] with respect to lnt, which is expressed as: The constraints satisfied by the coefficients are expressed as: Among them, σ t 2 represents the volatility of the time series, α0 represents the constant term, α i represents the ARCH term coefficient, which indicates the influence of the residual square term of the lag i period; u t-i 2 represents the residual square of the lag i period, β j represents the GARCH term coefficient, which indicates the influence of the conditional variance of the j-period lag; σ t-j 2 represents the conditional variance at lag j; The variance of a noise at the tth moment is a quantity that changes with time, which is conditional heteroskedasticity, and this variance that changes with time is a linear combination of the squares of a finite number of noise values ​​in the past, which is an autoregressive function and constitutes a regression conditional heteroskedasticity model; the variance of the noise at the tth moment is a combination of the squares of several noise values ​​in the past. If the variance of the noise increases in the previous several moments, the variance of the noise at the tth moment will increase accordingly; if the variance of the noise decreases in the previous several moments, the variance of the noise at the tth moment will also decrease accordingly.

7. The relay protection device evaluation method based on the GARCH model according to claim 6, characterized in that: The method of predicting future data based on the model and calculating the device reliability to obtain the device reliable service life includes using the GARCH model and estimating the model parameter μ using historical data. j , α i , β j ; Calculate the residuals after model fitting and perform ADF test on the residual sequence to ensure that the residual sequence is stable; Use the GARCH model to model the residual sequence and obtain the conditional variance; Use the fitted model to predict future time series data, and predict future conditional variance based on the GARCH model; Determine the failure threshold based on the device's usage requirements and historical failure data; calculate the reliability at each time point using the predicted time series data and conditional variance; The reliability threshold is determined to be 95% as the criterion for determining reliable service life; According to the reliability calculation results, find the time point T when the reliability reaches the set standard life , this time point is the reliable service life of the device.

8. A relay protection device evaluation system based on a GARCH model using the method according to any one of claims 1 to 7, characterized in that: The calculation module collects the operation data of the relay protection device and calculates the defect rate; estimates the parameters of the defect rate and calculates the residual sequence; Establish a model module, test whether the residual sequence is stable, and establish a reliability assessment model based on the GARCH model; The model prediction module predicts future data based on the model and obtains the reliable service life of the device through device reliability calculation.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images