Method and device for predicting degradation life of immersed power transformer

Abstract

The invention provides a method and device for predicting the degradation life of an electric transformer after water immersion, and relates to the technical field of electric transformers. The method comprises the following steps: acquiring fault rate data of the power transformer running along with time; fitting the fault rate data by adopting a least square method improved based on a whale optimization algorithm to obtain a Weibull fault rate bathtub curve when the power transformer normally operates; key indexes are selected to evaluate the health state of the power transformer and solve the fault rate in the corresponding health state, the fault rate of the power transformer after water immersion is mapped to a Weibull fault rate bathtub curve, and the equivalent degradation life of the power transformer after water immersion is solved. According to the method, quantitative analysis of the degradation degree of the immersed power transformer can be realized, and the degradation life of the immersed power transformer can be accurately predicted.

Description

Technical Field

[0001] The present invention relates to the technical field of power transformers, and particularly relates to a method and device for predicting the deterioration life of a power transformer after being immersed in water. Background Art

[0002] In the event of floods, a large number of power equipment and distribution rooms are severely flooded, resulting in serious damage to a large number of power grid metering devices. As a result, power transformers and acquisition terminals cannot continue to operate normally after being soaked in water, greatly affecting the accuracy of electric energy metering, and leading to frequent misoperations, random operations, and incorrect operations of power system control and protection devices, seriously affecting the safe, stable, and healthy operation of the power system.

[0003] Current scholars have achieved many results regarding the operating status and deterioration characteristics of power equipment such as power transformers. However, there are not many research results on the deterioration characteristics and operating status of power metering devices after being immersed in water for a long time. Some engineers, based on the Wiener model or the generalized Wiener stochastic model, used a large amount of historical data to establish evaluation methods for the remaining life of transformer aging and insulation aging. Therefore, if the historical data is extremely lacking, the accuracy of the evaluation model will be greatly reduced. Some engineers have established different evaluation models to evaluate the operating status of transformers, focusing on the historical working status of transformers, as well as the fuzziness and uncertainty of their operating status, and using the method of fuzzy mathematics to conduct a status evaluation of transformers, providing ideas for evaluating the operating status of power transformers after being immersed in water. Some scholars, based on the Weibull distribution, proposed life prediction methods and maintenance evaluation models for devices such as transformers and smart meters based on the Weibull model. There are also scholars who have carried out long-term immersion experiments on current transformers for the quality problems of current transformers after long-term immersion in water. The test results show that long-term immersion will cause irreversible degradation of the surface insulation material epoxy resin of power transformers. Some engineers have studied the aging characteristics of epoxy resin in a humid and hot environment, as well as the influence of moisture on the surface insulation characteristics of epoxy resin. All these theoretical studies and experimental tests not only provide important theoretical bases for evaluating the insulation aging of current transformers and lay an experimental foundation, but also provide reference ideas and methods for studying the failure rate of the entire life cycle of power equipment and simulating the health status.

[0004] Among power grid metering devices, power transformers are the main ones. Studying the deterioration characteristics of power transformers under long-term soaking conditions and analyzing and predicting their post-disaster recovery ability provide a decision-making basis for guiding the maintenance and replacement of post-disaster metering devices, and are of great significance for ensuring the healthy, safe, and reliable operation of the power system. Currently, the research on the deterioration characteristics of power transformers mainly focuses on the change processes and characteristics of indicators such as their insulation strength and test errors, lacking a quantitative analysis of the degree of deterioration of power transformers. Summary of the Invention

[0005] The object of the present invention is to provide a method and device for predicting the degradation life of a power transformer after being immersed in water, which are used to solve the problems in the prior art that lack of quantitative analysis of the degradation degree of the power transformer after being immersed in water and cannot accurately predict the equivalent degradation life of the power transformer after being immersed in water, so as to realize the quantitative analysis of the degradation degree of the power transformer after being immersed in water and accurately predict the degradation life of the power transformer after being immersed in water.

[0006] To achieve the above object, in a first aspect, the present invention provides a method for predicting the degradation life of a power transformer after being immersed in water, including:

[0007] Obtaining the failure rate data of the power transformer running over time;

[0008] Using the least squares method improved by the beluga optimization algorithm to fit the failure rate data to obtain the Weibull failure rate bathtub curve of the power transformer during normal operation;

[0009] Selecting key indicators to evaluate the health state of the power transformer and solving the failure rate under the corresponding health state, mapping the failure rate of the power transformer after being immersed in water into the Weibull failure rate bathtub curve, and solving the equivalent degradation life of the power transformer after being immersed in water.

[0010] According to the method for predicting the degradation life of a power transformer after being immersed in water provided by the present invention, using the least squares method improved by the beluga optimization algorithm to fit the failure rate data to obtain the Weibull failure rate bathtub curve of the power transformer during normal operation, including:

[0011] Step 21: Determine that the failure rate function of the two-parameter Weibull distribution of the power transformer is:

[0012]

[0013] In the formula, β is the shape parameter, η is the scale parameter, and t is the working time before failure;

[0014] Step 22: Using the least squares method improved by the beluga optimization algorithm to fit the failure rate data, determine the shape parameter, scale parameter, the demarcation point between the early failure period and the accidental failure period of the power transformer, and the demarcation point between the accidental failure period and the wear failure period, and then determine the Weibull failure rate bathtub curve.

[0015] According to the method for predicting the degradation life of a power transformer after being immersed in water provided by the present invention, step 22 specifically includes:

[0016] Step 221: Set the population size and the maximum number of iterations of the beluga optimization algorithm;

[0017] Step 222: Input the failure rate data and calculate the initial fitness value;

[0018] Step 223: Determine the best individual and fitness;

[0019] Step 224: Calculate the balance factor and the probability of whale fall;

[0020] Step 225: If the balance factor is greater than the preset value, enter the survey stage; otherwise, enter the development stage. If the balance factor is less than the probability of whale fall, enter the whale fall stage; otherwise, recalculate the fitness and determine the best individual;

[0021] Step 226: If the maximum number of iterations is less than the maximum number threshold at this time, increment the maximum number of iterations by one and return to Step 224; otherwise, output the best individual value;

[0022] Step 227: Select the demarcation point between the early failure period and the accidental failure period, and the demarcation point between the accidental failure period and the wear-out failure period;

[0023] Step 228: Use the two demarcation points and the best individual value as the initial points of the least squares method, and use the least squares method to fit the shape parameter and the scale parameter. If the fitting error is less than the set value, output the shape parameter, the scale parameter, the demarcation point between the early failure period and the accidental failure period, and the demarcation point between the accidental failure period and the wear-out failure period; otherwise, return to Step 227.

[0024] According to a prediction method for the deteriorated life of a power transformer after being immersed in water provided by the present invention, the key indicators include primary indicators and secondary indicators. The primary indicators include insulation performance, measurement error, reliability index, and environmental factors. The secondary indicators include insulation resistance, abnormal sound, sealing integrity, ratio difference, phase difference, family defects, historical faults, environmental humidity, and environmental temperature.

[0025] According to a prediction method for the deteriorated life of a power transformer after being immersed in water provided by the present invention, key indicators are selected to evaluate the health status of the power transformer and solve the failure rate under the corresponding health status. The failure rate of the power transformer after being immersed in water is mapped into the Weibull failure rate bathtub curve, and the equivalent deteriorated life of the power transformer after being immersed in water is solved, including:

[0026] Calculate the relative deterioration degree of the key indicators, and determine the membership function of the health index of the power transformer according to the relative deterioration degree;

[0027] According to a predetermined rule, divide the health status corresponding to the key indicators into multiple levels, calculate the membership degree of the secondary indicators with respect to the levels according to the membership function, and then calculate the health index of the secondary indicators according to the membership degree;

[0028] Determine the weight values of the primary indicators and the secondary indicators according to the influence degree of the key indicators;

[0029] The health index of the power transformer is obtained based on the secondary index health index and the weight values of the primary and secondary indexes;

[0030] Determine the failure rate of the power transformer after being immersed in water, and predict the equivalent deterioration life of the power transformer based on the failure rate, shape parameter, scale parameter of the power transformer after being immersed in water, and the health index of the power transformer.

[0031] According to a method for predicting the deterioration life of a power transformer after being immersed in water provided by the present invention, if the key index is a quantitative index and the larger the key index value is, the more ideal it is, then

[0032]

[0033] In the formula, S D is the relative deterioration degree, D max and D min are the maximum and minimum values defined by the regulations respectively; D M is the actual operating value of the key index.

[0034] According to a method for predicting the deterioration life of a power transformer after being immersed in water provided by the present invention, if the key index is a quantitative index and the smaller the key index value is, the more ideal it is, then

[0035]

[0036] In the formula, S D is the relative deterioration degree, D max and D min are the maximum and minimum values defined by the regulations respectively; D M is the actual operating value of the key index.

[0037] According to a method for predicting the deterioration life of a power transformer after being immersed in water provided by the present invention, the calculation formula of the secondary index health index is:

[0038]

[0039] In the formula, HI ij is the secondary index health index, b k is the score assigned to level k; m is the number of primary indexes; is the membership degree of the secondary index ij with respect to level k;

[0040] The calculation formula of the power transformer health index is:

[0041]

[0042] In the formula, HI is the power transformer health index, n i is the number of secondary indexes corresponding to the primary index i; ω iis the weight value of the first-level index i, ω ij is the weight value of the second-level index ij;

[0043] The calculation formula for the equivalent degradation life of a power transformer is:

[0044]

[0045] Among them,

[0046] λ(HI,t) = B(t)e -C(t)HI

[0047] B(t) = (1.6×10 4 e -1.98t + 0.25e 0.26t )×10 6

[0048] C(t) = 0.344e -0.846t + 0.177e 0.012t

[0049] In the formula, TD is the equivalent degradation life of the power transformer, β is the shape parameter, η is the scale parameter, t o is the demarcation point between the accidental failure period and the wear-out failure period, and λ(HI,t) is the failure rate of the power transformer after being immersed in water.

[0050] According to a method for predicting the degradation life of a power transformer after being immersed in water provided by the present invention, the power transformer includes a voltage transformer and a current transformer.

[0051] Second, the present invention provides a device for predicting the degradation life of a power transformer after being immersed in water, including:

[0052] An acquisition unit for acquiring the failure rate data of the power transformer running over time;

[0053] A fitting unit for fitting the failure rate data by using the least squares method improved based on the beluga optimization algorithm to obtain the Weibull failure rate bathtub curve when the power transformer is running normally;

[0054] A prediction unit for selecting key indicators to evaluate the health status of the power transformer and solving the failure rate under the corresponding health status, mapping the failure rate of the power transformer after being immersed in water into the Weibull failure rate bathtub curve, and solving the equivalent degradation life of the power transformer after being immersed in water.

[0055] The technical solution of the present invention at least has the following technical effects:

[0056] A method and device for predicting the deterioration life of a power transformer after immersion in water provided by the present invention. The method includes: obtaining the failure rate data of the power transformer during operation over time; using the least squares method improved based on the beluga optimization algorithm to fit the failure rate data to obtain the Weibull failure rate bathtub curve of the power transformer during normal operation; selecting key indicators to evaluate the health status of the power transformer and solving the failure rate under the corresponding health status, mapping the failure rate of the power transformer after immersion in water into the Weibull failure rate bathtub curve, and solving the equivalent deterioration life of the power transformer after immersion in water. The present invention can achieve quantitative analysis of the deterioration degree of the power transformer after immersion in water and accurately predict the deterioration life of the power transformer after immersion in water. Description of the Drawings

[0057] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0058] In the drawings:

[0059] Figure 1 It is a process diagram for the present invention to determine the relevant fitting parameters of the Weibull failure rate bathtub curve using the BWO-OLS algorithm;

[0060] Figure 2 It is a schematic diagram of the key indicators for the present invention to evaluate the health status of the power transformer;

[0061] Figure 3 It is a schematic diagram of the membership function of the health index of the power transformer for the present invention;

[0062] Figure 4 It is a fitting effect diagram of the proportionality coefficient for the present invention;

[0063] Figure 5 It is a fitting effect diagram of the curvature coefficient for the present invention;

[0064] Figure 6 It is a scatter diagram of the failure rate of the power transformer for the present invention;

[0065] Figure 7 It is a Weibull failure rate bathtub curve diagram of the power transformer for the present invention;

[0066] Figure 8 It is a bar chart of the deterioration life of the current transformer after immersion in water for the fitting method and the reference comparison method of the present invention;

[0067] Figure 9 It is a flowchart of the method for predicting the deterioration life of the power transformer after immersion in water for the present invention. Detailed implementation manners

[0068] To make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions in the present invention will be clearly and completely described below with reference to the accompanying drawings in the present invention. Apparently, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present invention without creative efforts shall fall within the protection scope of the present invention.

[0069] The following will describe in detail some implementation manners of the present invention with reference to the accompanying drawings. Without conflict, the embodiments described below and the features in the embodiments may be combined with each other.

[0070] The present invention simulates the change characteristics of the failure rate λ of power transformers under normal working conditions with the Weibull failure rate bathtub curve, uses their historical operation health data, and combines the least squares method improved by the beluga whale optimization algorithm (Beluga whale optimization-Ordinary least squares, BWO-OLS) to solve the shape parameter β and the scale parameter η of the bathtub curve of the power transformer. Key indicators are selected to evaluate their health status and solve the failure rate corresponding to this state, and the failure rate of the power transformer after being immersed in water is mapped into the Weibull failure rate bathtub curve to solve the equivalent degradation life of the power transformer after being immersed in water. It realizes the quantitative analysis of the degradation degree of the power transformer, and has relatively high prediction accuracy, which has certain reference significance for the risk assessment and maintenance of power transformers after flood disasters. The power transformers described in the present invention include voltage transformers and current transformers.

[0071] Please refer to Figure 9 , an embodiment of the present invention provides a method for predicting the degradation life of a power transformer after being immersed in water, including:

[0072] Step 1: Obtain the failure rate data of the power transformer running over time;

[0073] Step 2: Fit the failure rate data by using the least squares method improved by the beluga whale optimization algorithm to obtain the Weibull failure rate bathtub curve of the power transformer during normal operation;

[0074] It should be noted that the Weibull distribution is a commonly used distribution in reliability engineering, and its probability density function f(t) is:

[0075]

[0076] Wherein, β is a shape parameter and β > 0; η is a scale parameter and η > 0; γ is a location parameter and -∞ < γ < +∞; t is the operating time before failure and t > γ.

[0077] After the power transformer is installed, tested and operates normally, it is usually regarded as a non-repairable product. Research shows that the failure time of the power transformer conforms to the Weibull distribution model. Then, the failure probability density function f(t) of the two-parameter Weibull distribution of the power transformer can be expressed as:

[0078]

[0079] Then, the failure probability distribution function F(t) of the two-parameter Weibull distribution of the power transformer is;

[0080]

[0081] Then, the reliability function R(t) of the two-parameter Weibull distribution of the power transformer is:

[0082]

[0083] The failure rate function λ(t) of the power transformer is related to its operating reliability function R(t) and failure probability density function f(t). The failure rate function λ(t) of the two-parameter Weibull distribution is defined as the ratio of the failure probability density function f(t) to the reliability function R(t), that is

[0084]

[0085] Wherein, the value of the shape parameter β determines the curve of the failure rate of the power transformer. When the shape parameter satisfies 0 < β < 1, the initial failure rate of the power transformer decreases with the increase of time; when the shape parameter β = 1, the failure rate of the power transformer is a constant value. At this time, when the power transformer is in a stable operation state, the failures that occur are mostly accidental failures; when the shape parameter β > 1, the power transformer is in the wear-out period, and its failure rate rises steadily with time and finally reaches the failure rate threshold and is scrapped.

[0086] Taking the failure rate data of power transformers running over time as a sample, the failure rate data is fitted to obtain the failure rate curve of power transformers during normal operation. Currently, the main fitting methods include BP neural network, ordinary least squares (OLS), and Marquardt method. To ensure a relatively high fitting accuracy, the present invention uses the least squares method improved based on the beluga whale optimization (BWO) algorithm to fit the failure rate data of power transformers, obtaining the Weibull failure bathtub curve of power transformers during normal operation, and analyzing its degradation characteristics after long-term immersion in water on this basis.

[0087] Specifically, the beluga whale optimization algorithm is a swarm intelligence optimization algorithm, which is divided into three stages: exploration, exploitation, and whale fall, and introduces a balance factor B f , and a whale fall probability W f to distinguish the above three stages, and its expression is:

[0088]

[0089] In the formula, T is the current iteration number, and T max is the total number of iterations. B o is a random value within the range of (0, 1) and will change in each iteration. When B f > 0.5, the whale population enters the exploration stage, otherwise it enters the exploitation stage. When B f < W f , it enters the whale fall stage.

[0090] In the exploration stage, the beluga whales swim in pairs, and the position is updated according to the following formula, that is

[0091]

[0092] In the formula, is the position of the i-th beluga whale updated in the j-th dimension, and P j (j = 1, 2,..., D) is a random number selected from the D dimensions, r 1 , r 2 is a random value within the range of (0, 1).

[0093] In the exploitation stage, the beluga whale population shares position information and swims together to cooperate in foraging. And the Levy flight strategy is introduced to enhance the global convergence of the algorithm. The position update is as follows:

[0094]

[0095] Γ(x) = (x - 1)! (13) (13)

[0096] In the formula, is the optimal position of the beluga whale, XT r is the random position of the beluga whale, C 1 represents the random jump degree, r 3 , r 4 is a random value within the range of (0, 1). L F is a random number that conforms to the Levy distribution, where both μ and v are random numbers subject to the normal distribution, that is, μ, v ∼ N(0, 1), and τ is a constant, set to 1.5.

[0097] During the whale fall stage, when the beluga whale is threatened by external life, the beluga whale transfers to other places or is shot down and falls into the deep sea. To ensure that the population quantity remains unchanged, the position of the beluga whale is updated according to the position of the beluga whale and the falling step length. At this time, the update formula for the position of the beluga whale is:

[0098]

[0099] X step =(u b -l b )exp(C 2 T / T max ) (15)

[0100] C 2 =2W f ×N (16)

[0101] In the formula, r 5 , r 6 , r 7 is a random value between (0, 1). X step is the whale fall step length, u b , l b are the search upper and lower bounds, C 2 is the step factor, and N is the number of beluga whale populations.

[0102] Use the beluga whale optimization algorithm to fit the failure rate curve. Set the fitting parameter as the position of the beluga whale, the fitness function as the sum of the squares of the curve fitting residuals, and the fitting goal as the minimum sum of the squares of the residuals.

[0103] The beluga whale optimization algorithm has the advantages of not requiring manual setting of the upper and lower limits of the fitting parameters and being easy to use. However, the fitting accuracy is not high. Although the existing least squares algorithm has high accuracy, it requires manual adjustment of the fitting settings multiple times to achieve the best fitting effect. Therefore, to overcome the disadvantages of the above two algorithms and combine their advantages, the least squares method improved based on the beluga whale optimization algorithm is used to fit the power transformer failure rate data.

[0104] In view of the piecewise characteristic of the failure rate function of the Weibull distribution, it is necessary to perform piecewise fitting on the failure rates of power transformers in different operation stages according to the actual situation. For this purpose, assume there are f failure sample points, and select the k-th data sample point as the demarcation point between the early failure period and the accidental failure period of the power transformer (k < f); similarly, select the l-th data sample point as the demarcation point between the accidental failure period and the wear-out failure period of the power transformer (k < l < f).

[0105] And perform curve fitting on the failure rate data samples distributed in the three failure periods to solve the shape parameter β and the scale parameter η. The conditions satisfied by the sample demarcation points k and l are as follows:

[0106]

[0107] In the formula, λ i is the sample failure rate; λ Ⅰ (t i ) is the failure rate of the early failure period corresponding to the sample failure rate λ i ; λ Ⅱ (t i ) is the failure rate of the accidental failure period corresponding to the sample failure rate λ i ; λ Ⅲ (t i ) is the failure rate of the wear-out failure period corresponding to the sample failure rate λ i . Determine the sample demarcation points k and l by seeking S = min(S l ).

[0108] Based on this, in some embodiments, step 2 specifically includes:

[0109] Step 21, determine the failure rate function of the two-parameter form Weibull distribution of the power transformer as:

[0110]

[0111] In the formula, β is the shape parameter, η is the scale parameter, and t is the working time before failure;

[0112] Step 22, use the least squares method improved based on the beluga optimization algorithm to fit the failure rate data, determine the shape parameter, the scale parameter, the demarcation point between the early failure period and the accidental failure period of the power transformer, and the demarcation point between the accidental failure period and the wear-out failure period, and then determine the Weibull failure rate bathtub curve.

[0113] Among them, the specific process of step 22, that is, the process of the present invention using the BWO-OLS algorithm to determine the relevant fitting parameters of the Weibull failure rate bathtub curve is as Figure 1 shown:

[0114] Step 221: Set the population size and maximum number of iterations of the beluga optimization algorithm;

[0115] Step 222: Input the failure rate data and calculate the initial fitness value;

[0116] Step 223: Determine the best individual and fitness;

[0117] Step 224: Calculate the balance factor and whale fall probability;

[0118] Step 225: If the balance factor is greater than the preset value, enter the exploration stage; otherwise, enter the development stage. If the balance factor is less than the whale fall probability, enter the whale fall stage; otherwise, recalculate the fitness and determine the best individual;

[0119] Step 226: If the maximum number of iterations is less than the maximum number threshold at this time, increment the maximum number of iterations by one and return to Step 224; otherwise, output the best individual value;

[0120] Step 227: Select the demarcation points between the early failure period and the accidental failure period, and between the accidental failure period and the wear-out failure period;

[0121] Step 228: Use the two demarcation points and the best individual value as the initial points of the least squares method, and use the least squares method to fit the shape parameter and the scale parameter. If the fitting error is less than the set value, output the shape parameter, the scale parameter, the demarcation point between the early failure period and the accidental failure period, and the demarcation point between the accidental failure period and the wear-out failure period; otherwise, return to Step 227.

[0122] Step 3: Select key indicators to evaluate the health status of the power transformer and solve the failure rate corresponding to the health status. Map the failure rate of the power transformer after immersion in water to the Weibull failure rate bathtub curve, and solve the equivalent degradation life of the power transformer after immersion in water.

[0123] Specifically, refer to the "TSMA 0036-2023 Technical Specification for 10kV~35kV Current Transformers for Metering", and consider the comprehensiveness and practicality of the health index evaluation indicators. As Figure 2 shown, select the evaluation indicators for the health status of the power transformer, such as: four first-level indicators of insulation performance, measurement error, reliability index, and environmental factors. Further supplement the secondary indicators of the power transformer to improve the evaluation index system of the power transformer.

[0124] In the measurement, the above first-level indicators can be divided into quantitative indicators and qualitative indicators. The so-called quantitative indicators refer to the indicators that can be quantified, such as: "insulation resistance", "ratio difference and phase difference", "ambient temperature" and "ambient humidity". Since the dimensions and orders of magnitude of different quantitative indicators are different, for the convenience of research, a dimensionless normalization processing method is adopted, that is, calculating the relative deterioration degree S of the quantitative indicator D .

[0125] For the indicators where the larger the test value is, the more ideal it is, such as the insulation resistance value, the relative deterioration degree S of the quantitative indicator D is normalized by Equation (18), that is

[0126]

[0127] In the formula, S D is the relative deterioration degree of the quantitative indicator; D max and D min are the maximum value and the minimum value defined by the regulations respectively; D M is the actual operating value of the corresponding indicator. For the indicators where the smaller the test value is, the more ideal it is, such as the ambient humidity, the relative deterioration degree S of the quantitative indicator D can be normalized by Equation (19), that is

[0128]

[0129] It should be noted that when the relative deterioration degree S of the quantitative indicator D < 0, take S D = 0; when S D > 1, take S D = 1, that is, ensure that the relative deterioration degree S of the quantitative indicator D takes values between 0 and 1.

[0130] For qualitative indicators, such as: "abnormal sound", "sealing integrity", "family defect", "historical fault", the relative deterioration degree S is calculated according to the method of quantitative indicators after deducting points for them by referring to the "Q / GDW 10446-2021 Current Transformer Condition Evaluation Guide" and combining experience D .

[0131] For the convenience of elaboration, the health index is represented by HI, which reflects the current health degree of the power transformer, and its value is jointly determined by 9 secondary evaluation indicators (insulation resistance, abnormal sound, sealing integrity, ratio difference, phase difference, family defect, historical fault, ambient humidity and ambient temperature), that is

[0132] (1) Calculate the membership degree of the secondary indicator with respect to level k according to the membership function and the health index HI of the secondary indicatorij ;

[0133] (2) Determine the respective weight values of the primary indicators and secondary indicators according to the degree of influence of the key indicators.

[0134] Evaluating the "reliability" and "failure" of power transformers using secondary indicators is ambiguous. Therefore, the method of fuzzy mathematics is used to establish the membership function corresponding to the secondary indicators, and the health index HI of the secondary indicators is calculated according to the membership function. ij . The trapezoidal function is selected as the membership function, and the membership function of the power transformer health index is as Figure 3 shown. Determine the parameters X D in the interval [0, 1] Figure 3 in 1 , X 2 , X 3 , X 4 , X 5 and X 6 , that is, determine the membership function of the power transformer health index according to the relative deterioration degree.

[0135] According to the predetermined rules, such as preliminary research and reference to the State Grid Power Transformer Condition Assessment Guide, divide the health status corresponding to the primary indicators and secondary indicators of the power transformer into multiple levels, such as 4 levels: normal, attention, abnormal, and serious, as shown in Table 1.

[0136] Table 1. Corresponding Status of Power Transformer Health Index

[0137]

[0138]

[0139] The calculation expression of the health index HI of the secondary indicator is: ij

[0140]

[0141] In the formula, b k is the score assigned to level k. In this embodiment, scores of 30, 67.5, 80, and 92.5 are assigned to the four levels of the power transformer status respectively. HI ij is the health index of the secondary indicator; m is the number of primary indicators. In this embodiment, m = 4; is the membership degree of the secondary indicator ij with respect to level k.

[0142] As mentioned above, 4 primary evaluation indicators are selected, and the calculation formula of the power transformer health index HI considering the multi-level and multi-factor effects is:

[0143]

[0144] In the formula, n i is the number of secondary indicators corresponding to the primary indicator i (a total of 9 secondary indicators); ω i is the weight value corresponding to the primary indicator i, and ω ij is the weight value of the secondary indicator ij.

[0145] There is an exponential relationship between the equipment failure rate and its health status. The higher the health index of the power transformer, the lower its failure rate. The "Guide for Risk Assessment of Transmission and Distribution Equipment" gives its expression, that is

[0146] λ = Be -C·HI (22)

[0147] In the formula, B is the proportionality coefficient; C is the curvature coefficient. The commonly used solution methods for the parameters B and C are the inversion method and the full-state integration method.

[0148] It should be noted that the failure rate model based on the equipment health status mentioned above only considers the impact of equipment state changes on its operation reliability. In actual engineering, the failure forms of power transformer equipment are very complex, and they also experience degradation and failure during operation. Therefore, it is more reasonable to establish a failure rate model based on time and equipment state.

[0149] Therefore, the original failure rate formula is improved, that is

[0150] λ(HI, t) = B(t)e -C(t)HI (22)

[0151] After improvement, the proportionality coefficient and the curvature coefficient are no longer constant values, but variables related to the equipment operation time. The expressions of B(t) and C(t) are solved by the method of parameter fitting. The parameters of the equipment state failure rate model obtained by using the full-state integration method are shown in Table 2.

[0152] Table 2. Parameters of the equipment state failure rate model

[0153]

[0154] The proportionality coefficient B is fitted, and the fitting effect of the proportionality coefficient B is as Figure 4 shown.

[0155] The formula for the proportionality coefficient B changing with time after solution is:

[0156] B(t) = (1.6×10 4 e -1.98t + 0.25e 0.26t )×10 6 (23)

[0157] Fitting the curvature coefficient C, the fitting effect of the curvature coefficient C is as Figure 5 shown.

[0158] The solved formula for the curvature coefficient C varying with time is:

[0159] C(t) = 0.344e -0.846t + 0.177e 0.012t (24)

[0160] Using the failure rate model (22) of the power transformer based on time and equipment status and the Weibull failure rate model (5), the equivalent deterioration life after their status change can be deduced.

[0161] When the health index of the power transformer after immersion in water is HI and the equivalent service age is t, their failure rate after immersion in water is λ(HI, t). After the power transformer is immersed in water, it accelerates deterioration and enters the loss failure period. Taking the boundary time t o between the accidental failure period and the loss failure period as the starting time of deterioration, and substituting the failure rate λ(HI, t) into Equation (5) to calculate the equivalent deterioration life TD, that is

[0162]

[0163] Analyzing Expression (26) can obtain the following conclusions:

[0164] (1) The deterioration life of the power transformer before and after immersion in water is related to the health index HI of the transformer.

[0165] (2) The deterioration life TD of the power transformer after immersion in water is related to both the shape parameter β and the scale parameter η of the Weibull distribution, and is also affected by the current service age of the equipment.

[0166] Therefore, in some embodiments, Step 3 specifically includes:

[0167] Calculating the relative deterioration degree of the key indicators, and determining the membership function of the health index of the power transformer according to the relative deterioration degree;

[0168] Dividing the health status corresponding to the key indicators into multiple levels according to the predetermined rules, calculating the membership degree of the secondary indicators with respect to the levels according to the membership function, and then calculating the health index of the secondary indicators according to the membership degree;

[0169] Determining the weight values of the primary indicators and the secondary indicators according to the influence degree of the key indicators;

[0170] Obtaining the health index of the power transformer based on the health index of the secondary indicators and the weight values of the primary indicators and the secondary indicators;

[0171] Determine the failure rate of the power transformer after being immersed in water, and predict the equivalent degradation life of the power transformer based on the failure rate, shape parameter, scale parameter of the power transformer after being immersed in water, and the health index of the power transformer.

[0172] Due to the extreme lack of statistical data on the failure rate of power transformers after long-term immersion in water, and considering that the operating principle of power transformers is the same as that of electromagnetic transformers, therefore, the probability data of transformer failures statistically collected by a certain power grid company is used to approximate the Weibull failure rate bathtub curve of power transformers.

[0173] Analysis Figure 6 From the scatter plot of the failure rate of the transformer shown, similar to the transformer, the failure rate of the power transformer also conforms to the Weibull failure rate curve, and as time increases, the state gradually transitions from the accidental failure period to the wear-out failure period. Substitute the data sample into the BWO-OLS algorithm, and set the BWO algorithm parameters N = 50, T max = 1000, and accordingly, the equivalent Weibull distribution failure rate curve of the power transformer is fitted and solved, as shown in Figure 7 shown.

[0174] Continuously, compare and study the segmented fitting and solving results of the improved BWO-OLS algorithm in the present invention with the traditional Beluga Whale Optimization algorithm (BWO), Least Squares Parameter Estimation (OLS), and Marquardt method. The comparison of the fitting effects is shown in Table 3.

[0175] Table 3. Comparison of fitting effects

[0176]

[0177] By comparing and analyzing Table 3, it can be known that the sum of squared residuals RSS after fitting by the improved BWO-OLS segmented fitting and solving method is more ideal, and the coefficient of determination R 2 is closer to 1, that is, the fitting effect is better.

[0178] Based on the fitting results, the following important conclusions are obtained:

[0179] (1) The dividing point l between the accidental failure period and the wear-out failure period of the power transformer is the 16th year;

[0180] (2) The shape parameter β within the accidental failure period is 1.376, and the scale parameter η is 41.736;

[0181] (3) The shape parameter β within the wear-out failure period is 4.577, and the scale parameter η is 28.287.

[0182] Therefore, the Weibull failure rate model expression of the power transformer can be obtained, that is

[0183]

[0184] It is known from analyzing the fitting effect that the failure rate of power equipment conforms to the bathtub curve. During the accidental failure period, it operates stably and the failure rate does not fluctuate greatly. During the wear-out failure period, due to wear-out factors such as aging, fatigue, wear, and corrosion, the failure rate increases sharply with the service age and finally fails.

[0185] Based on the same inventive concept, another embodiment of the present invention provides a prediction device for the deteriorated life of a power transformer after being immersed in water. This device corresponds to the method of the foregoing embodiment, and this device includes:

[0186] An acquisition unit for acquiring the failure rate data of the power transformer running over time;

[0187] A fitting unit for fitting the failure rate data by using the least squares method improved based on the beluga optimization algorithm to obtain the Weibull failure rate bathtub curve of the power transformer during normal operation;

[0188] A prediction unit for selecting key indicators to evaluate the health status of the power transformer and solving the failure rate under the corresponding health status, mapping the failure rate of the power transformer after being immersed in water into the Weibull failure rate bathtub curve, and solving the equivalent deteriorated life of the power transformer after being immersed in water.

[0189] The following is a specific embodiment of the present invention.

[0190] In order to test the key index parameter values affecting the state of the power transformer after being immersed in water, current transformers of 8 different models (one for each model, a total of 8) are selected for the immersion experiment. Two situations of setting the immersion duration of ten days and long-term drying after immersion are respectively set. Now the specific situation of the experiment is described as follows:

[0191] (1) Take eight test samples (one for each of the 8 different models) and put them into the test equipment, inject tap water, the liquid level height is 1000 mm, take them out after 10 days of immersion, wipe off the water droplets on them with a dry cloth, let them stand for 15 min, and start testing the key index parameter values affecting the state of the current transformer after drying.

[0192] (2) On the basis of step (1), dry them at room temperature for 5 days and conduct the test again after air drying.

[0193] For qualitative indicators, "abnormal sound", "sealing integrity", "family defect", and "historical fault" are scored by experts referring to the "Current Transformer State Evaluation Guide Q / GDW 10446-2021" and combining experience. The immersion experiment data of the eight current transformers are shown in Table 4.

[0194] Table 4. Immersion experiment data of eight current transformers

[0195]

[0196] According to the analytic hierarchy process and combined with the experience accumulated in maintenance, the weight ω of the first-level index is set i =[0.4 0.1 0.2 0.3], and the corresponding weights of the second-level indexes are [0.5 0.375 0.125], [0.5 0.5], [0.5 0.5], [0.5 0.5] (consistent with the order of the second-level indexes in Figure 2 ). After calculating the relative deterioration degree S for the above data D , substituting it into the corresponding membership function, the membership degrees and health indexes HI of the corresponding states of each current transformer are solved, and the results are shown in Table 5

[0197] Table 5. Fuzzy comprehensive evaluation results of eight different types of current transformers in different states

[0198]

[0199] Referring to the corresponding states of the health indexes of the power transformers in Table 1, it can be obtained that the #5 transformer is in a state of attention after being immersed in water for ten days, and the rest of the transformers are in an abnormal state. After drying for a long time, the #3 transformer is in an abnormal state, and the rest of the transformers return to the normal state. The evaluation results of the operation state of the transformers using the evaluation method and the experimental data of the transformers show that: the state grades of the transformers after being immersed in water for ten days are [5 5 6 5 5 5 5 5]. At this grade, the #3 transformer is between the abnormal and serious states, and the rest of the transformers are in an abnormal state; the state grades of the transformers after drying for a long time are [1 1 5 1 1 1 1 1]. Except for the #3 transformer being in an abnormal state, the rest of the transformers are in a normal state. By comparison, the effectiveness of the method of the present invention is fully verified

[0200] The current transformer is set within the accidental failure period, the equivalent service age is 5 years, and the corresponding shape factor B = 1.73×10 6 , and the curvature factor C = 0.193. There is a demarcation time t 0 = 16 between the accidental failure period and the wear failure period. To verify the effectiveness of the evaluation, the method of equivalent service age is used for comparison. The equivalent deterioration life of the transformer after being immersed in water is shown in Table 6 Figure 8 as shown

[0201] Table 6. Equivalent deterioration life of the current transformer after being immersed in water

[0202]

[0203] Analyzing Table 6 Figure 8 the following conclusions can be obtained

[0204] (1) Under long-term immersion conditions, the health index score HI of Transformer #3 is 74.13, which is in an abnormal state. One or more of its evaluation indicators exceed the safe range, and maintenance should be arranged immediately (refer to Table 1); after the experiment, a careful inspection of Transformer #3 found that there were obvious water traces inside the joint between its bottom plate and epoxy resin. In a long-term immersed and humid environment, water enters the current transformer through the gap and cannot precipitate, resulting in a decline in its insulation performance. The actual detection of the current transformer is consistent with the evaluation status, indicating the effectiveness of the evaluation method;

[0205] (2) The fitting method adopted in the present invention shows a smaller error compared with the reference Marquardt method, thus making the prediction of the deterioration life more accurate. In addition, the evaluation method of the present invention for the state of the current transformer is more comprehensive and has higher objectivity, which is superior to the simple point deduction method;

[0206] (3) After long-term immersion, the current transformer has undergone irreversible degradation, and the failure rate has increased significantly, from less than 0.02 (times / year) during the accidental failure period to an average failure rate of 0.03 (times / year), and the average deterioration life reaches 2.05 years.

[0207] In summary, aiming at the problem that it is difficult to evaluate the quality and deterioration degree of power transformers, which are typical metering devices in the power grid, after flood disasters, the present invention proposes a method for predicting the deterioration life of power transformers after immersion based on the health index and the Weibull failure rate model, including:

[0208] (1) Use the least squares method improved by the beluga whale optimization algorithm (BWO-OLS) to fit the failure rate curve of the power transformer. The improved fitting method has higher accuracy and overcomes the shortcoming of the least squares method that requires manual multiple parameter settings;

[0209] (2) Use the fuzzy comprehensive evaluation method to evaluate the health status of the power transformer, taking into account the fuzziness and uncertainty of the evaluation indicators, and having stronger universality and applicability;

[0210] (3) Improve the original failure rate model of the equipment health status, taking into account the influence of the equipment service life, and innovatively establish a failure rate model based on time and equipment status, and use the fitting method to solve the proportionality coefficient B(t) and the curvature coefficient C(t);

[0211] (4) A method for predicting the deterioration life of the power transformer based on the health index and the Weibull failure rate model, and combining the immersion experiment to predict that the average deterioration life of the current transformer after immersion is 2.05 years. The failure rate has increased significantly, from less than 0.02 (times / year) during the accidental failure period to an average failure rate of 0.03 (times / year);

[0212] In summary, conducting research on the degradation characteristics of power transformers in power grid metering devices after immersion in water is crucial for the power grid company's risk assessment and maintenance decision-making of power transformers after flood disasters. Therefore, the present invention proposes a method for predicting the degradation life of power transformers after immersion in water based on the health index and Weibull failure rate model, that is, by using the fuzzy comprehensive evaluation method to solve the health index of the transformer, describing the operating state of the transformer in fuzzy language, and using the improved failure rate model based on time and equipment state to determine the failure rate of the transformer in the current state. Further, the least squares method improved by the beluga optimization algorithm is used to piecewise fit the Weibull failure rate bathtub curve of the power transformer, map the failure rate in its state after immersion in water into the Weibull failure rate bathtub curve, and solve the equivalent degradation life of the power transformer after immersion in water. Eight different types of current transformers are selected to conduct immersion experiments, and the test results show that the current transformers will undergo irreversible degradation after immersion in water, and their operating states after immersion in water are consistent with the model evaluation results, verifying the effectiveness of the degradation transformer life prediction method after immersion in water. It has certain reference significance for the power grid company's risk assessment and maintenance of power transformers after flood disasters.

[0213] After considering the specification and practicing the disclosed embodiments herein, those skilled in the art will readily conceive of other embodiments of the present invention. The present invention is intended to cover any variations, uses, or adaptations of the present invention, which follow the general principles of the present invention and include known common knowledge or conventional technical means in the technical field not disclosed by the present invention. It should be understood that the present invention is not limited to the precise structures described above and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of the present invention is only limited by the appended claims.

Claims

1. A method for predicting the degradation life of a power transformer after immersion in water, characterized in that: include: Obtaining failure rate data of power transformers over time; The fault rate data are fitted using a least square method improved based on the White Whale optimization algorithm to obtain a Weibull fault rate bathtub curve when the power transformer is operating normally; Key indicators are selected to evaluate the health status of the power transformer and solve the failure rate under the corresponding health status, the failure rate of the power transformer after immersion in water is mapped to the Weibull failure rate bathtub curve, and the equivalent degradation life of the power transformer after immersion in water is solved.

2. The method for predicting the degradation life of a power transformer after immersion in water according to claim 1, characterized in that: The least square method improved based on the White Whale optimization algorithm is used to fit the failure rate data to obtain the Weibull failure rate bathtub curve when the power transformer is operating normally, including: Step 21, determine the failure rate function of the two-parameter Weibull distribution of the power transformer as: In the formula, β is the shape parameter, η is the scale parameter, and t is the working time before failure; Step 22: Use the least square method improved based on the White Whale optimization algorithm to fit the failure rate data, determine the shape parameter, the proportional parameter, and the dividing point between the early failure period and the accidental failure period, and the dividing point between the accidental failure period and the loss failure period of the power transformer, and then determine the Weibull failure rate bathtub curve.

3. The method for predicting the degradation life of a power transformer after immersion in water according to claim 2, characterized in that: The step 22 specifically includes: Step 221, setting the population size and maximum number of iterations of the White Whale optimization algorithm; Step 222, input the failure rate data and calculate the initial suitability value; Step 223, determine the best individual and fitness; Step 224, calculating the balance factor and the whale fall probability; Step 225: If the balance factor is greater than the preset value, the exploration phase is entered, otherwise the development phase is entered; if the balance factor is less than the whale fall probability, the whale fall phase is entered, otherwise the fitness is recalculated and the best individual is determined; Step 226: If the maximum number of iterations is less than the maximum number threshold, then the maximum number of iterations is increased by one and the process returns to step 224; otherwise, the best individual value is output; Step 227, selecting the dividing point between the early failure period and the accidental failure period and the dividing point between the accidental failure period and the loss failure period; Step 228, taking the two dividing points and the best individual value as the initial points of the least squares method, and using the least squares method to fit the shape parameters and proportion parameters, if the fitting error is less than the set value, output the shape parameters, proportion parameters, and the dividing point between the early failure period and the accidental failure period, and the dividing point between the accidental failure period and the loss failure period, otherwise return to step 227.

4. The method for predicting the degradation life of a power transformer after immersion in water according to claim 3 is characterized in that: The key indicators include primary indicators and secondary indicators. The primary indicators include insulation performance, measurement error, reliability indicators and environmental factors. The secondary indicators include insulation resistance, abnormal sound, sealing integrity, ratio difference, phase difference, family defects, historical failures, ambient humidity and ambient temperature.

5. The method for predicting the degradation life of a power transformer after immersion in water according to claim 4, characterized in that: The selecting key indicators to evaluate the health status of the power transformer and solving the failure rate under the corresponding health status, mapping the failure rate of the power transformer after being immersed in water to the Weibull failure rate bathtub curve, and solving the equivalent degradation life of the power transformer after being immersed in water, includes: Calculating the relative degradation degree of the key indicator, and determining the membership function of the power transformer health index according to the relative degradation degree; Dividing the health status corresponding to the key indicator into multiple levels according to a predetermined rule, calculating the membership of the secondary indicator with respect to the level according to the membership function, and then calculating the secondary indicator health index according to the membership; Determine the weight values ​​of the primary and secondary indicators according to the impact of the key indicators; Obtaining a power transformer health index based on the secondary indicator health index and the weight values ​​of the primary indicator and the secondary indicator; The failure rate of the power transformer after being immersed in water is determined, and the equivalent degradation life of the power transformer is predicted based on the failure rate, shape parameter, proportional parameter and power transformer health index of the power transformer after being immersed in water.

6. The method for predicting the degradation life of a power transformer after immersion in water according to claim 5, characterized in that: If the key indicator is a quantitative indicator and the larger the key indicator value, the better, then In the formula, S D is the relative degradation degree, D max and D min They are the maximum and minimum values ​​defined by the regulations respectively; D M It is the actual operating value of the key indicator.

7. The method for predicting the degradation life of a power transformer after immersion in water according to claim 5, characterized in that: If the key indicator is a quantitative indicator and the smaller the key indicator value, the better, then In the formula, S D is the relative degradation degree, D max and D min They are the maximum and minimum values ​​defined by the regulations respectively; D M It is the actual operating value of the key indicator.

8. The method for predicting the degradation life of a power transformer after immersion in water according to claim 5, characterized in that: The calculation formula of the secondary indicator health index is: In the formula, HI ij is the secondary indicator health index, b k is the assigned score of level k; m is the number of first-level indicators; is the membership degree of the secondary index ij to level k; The calculation formula of the power transformer health index is: Where HI is the health index of the power transformer, n i is the number of secondary indicators corresponding to the primary indicator i; ω i is the weight value of the first-level index i, ω ij is the weight value of the secondary index ij; The calculation formula for the equivalent degradation life of the power transformer is: in, λ(HI,t)=B(t)e -C(t)HI B(t)=(1.6×10 4 e -1.98t +0.25e 0.26t )×10 6 C(t)=0.344e -0.846t +0.177e 0.012t Where TD is the equivalent degradation life of the power transformer, β is the shape parameter, η is the proportional parameter, and t o is the dividing point between the accidental failure period and the loss failure period, and λ(HI,t) is the failure rate of the power transformer after being immersed in water.

9. The method for predicting the degradation life of a power transformer after immersion in water according to claim 1, characterized in that: The power transformer includes a voltage transformer and a current transformer.

10. A device for predicting the degradation life of a power transformer after being immersed in water, characterized in that: include: An acquisition unit, used for acquiring failure rate data of the power transformer running over time; A fitting unit, used for fitting the fault rate data by using a least square method improved based on the White Whale optimization algorithm, to obtain a Weibull fault rate bathtub curve when the power transformer is operating normally; The prediction unit is used to select key indicators to evaluate the health status of the power transformer and solve the failure rate under the corresponding health status, map the failure rate of the power transformer after immersion in water to the Weibull failure rate bathtub curve, and solve the equivalent degradation life of the power transformer after immersion in water.

Images