Generator insulation life prediction method and system based on Bayesian theorem

Abstract

The invention discloses a Bayesian theorem-based generator insulation life prediction method and system, and belongs to the technical field of power system insulation materials, and the method comprises the steps: predicting the insulation life of a generator through at least two insulation life prediction methods in each prediction period, and obtaining a prediction value of each method; the same initial prior probability is set for each method, and a predicted posterior probability is obtained through Bayesian theorem calculation according to a predicted value and the prior probability; and updating the posterior probability according to each prediction value, and using the posterior probability as the prior probability of next prediction until all prediction cycles are completed. According to the method, the posterior probability is dynamically updated by combining multiple prediction methods and utilizing the Bayesian theorem, the prediction precision is remarkably improved, and errors in a traditional method are reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of insulation materials for power systems, and in particular to a method and system for predicting the insulation life of a generator based on Bayesian theorem. Background Art

[0002] Generators are critical power equipment, and the health of their insulation systems directly impacts their safety and stability. A generator's insulation life refers to the time the insulation material can withstand voltage stress without failure. This lifespan is affected by a variety of factors, such as partial discharge, temperature, humidity, and workload. These factors can cause insulation material aging and degradation, shortening the generator's service life.

[0003] Traditional insulation life prediction methods are typically based on regular inspections and tests, such as partial discharge tests and electrical performance tests. However, these methods often rely on limited measured data and cannot accurately predict the long-term degradation process of insulation materials under different operating conditions. Summary of the Invention

[0004] To solve the above technical problems, a generator insulation life prediction method based on Bayesian theorem is proposed, including: in a first prediction cycle, predicting the insulation life of the generator by at least two insulation life prediction methods to obtain a first prediction value of each method; setting the same initial prior probability for each method, and calculating the posterior probability of the first prediction by Bayesian theorem based on the prediction value and the prior probability; in a second prediction cycle, using the same type of insulation life prediction method to predict the insulation life of the generator to obtain a second prediction value of each method; using the posterior probability of the first prediction result as the prior probability of the second prediction to calculate the second prediction value; in a third prediction cycle, using the same type of insulation life prediction method to predict the insulation life of the generator to obtain a third prediction value of each method; using the posterior probability of the second prediction result as the prior probability of the third prediction to calculate the third prediction value; updating the posterior probability according to each prediction value, and using the posterior probability as the prior probability of the next prediction until all prediction cycles are completed.

[0005] As a preferred embodiment of the method for predicting the insulation life of a generator based on the Bayesian theorem described in the present invention, the insulation life prediction method includes a partial discharge parameter prediction method, a D image method, and a NY image method; the partial discharge parameter prediction method measures electrical parameters, calculates the ratio of the residual breakdown voltage to the rated voltage, and predicts the remaining life of the equipment; the D image method predicts the residual breakdown voltage and the remaining life of the equipment through the discharge index and the maximum partial discharge amount; the NY image method comprehensively considers the effects of electrical aging, thermal aging, and hot and cold cycles, performs a weighted average of the residual breakdown voltage, and predicts the remaining life of the equipment.

[0006] As a preferred solution of the generator insulation life prediction method based on Bayesian theorem described in the present invention, the partial discharge parameter prediction method is expressed as:

[0007]

[0008] Among them, U BD is the residual breakdown voltage value; U n is the rated voltage value; Q m is the maximum partial discharge; tanδ0 is the dielectric loss tangent at 2kV; R1 is the insulation resistance value for 1min; C0 is the capacitance at 2kV.

[0009] As a preferred solution of the generator insulation life prediction method based on Bayesian theorem described in the present invention, the D image method is expressed as:

[0010]

[0011] Δ=Δtanδ+ΔI

[0012] Among them, U BD % is the predicted value of the residual breakdown voltage; Δ is the discharge index; Q max is the maximum partial discharge under the working voltage; Δ is the discharge index representing the average aging of the insulation; Δtanδ is the increase in the dielectric loss tangent; ΔI is the AC current growth rate.

[0013] As a preferred solution of the generator insulation life prediction method based on Bayesian theorem described in the present invention, the NY image method is expressed as:

[0014]

[0015] Wherein, U0 is the initial breakdown voltage value; is the residual breakdown voltage caused by electrical aging; is the residual breakdown voltage caused by thermal aging; is the residual breakdown voltage caused by thermal cycling.

[0016] As a preferred solution of the generator insulation life prediction method based on Bayesian theorem described in the present invention, wherein: the calculation of the second prediction value includes:

[0017] The distribution of the predicted value of the generator insulation life obtained by the partial discharge parameter prediction method is N(L21,1), the distribution of the predicted value of the generator insulation life obtained by the D image prediction method is N(L22,1), and the distribution of the predicted value of the generator insulation life obtained by the NY image prediction method is N(L23,1);

[0018] Calculate the probabilities of the overlapping parts of N(L21,1), N(L22,1), N(L23,1) and N(L11+R2,1), N(L12+R2,1), N(L13+R2,1) on the probability distribution graph respectively as P(2|1,1), P(2|1,2), P(2|1,3);

[0019] The posterior probabilities of the second prediction are calculated as follows:

[0020] P21=P(2|1,1)×1 / 3 / (P(2|1,1)×1 / 3+P(2|1,2)×1 / 3+P(2|1,3)×1 / 3)

[0021] P22=P(2|1,2)×1 / 3 / (P(2|1,1)×1 / 3+P(2|1,2)×1 / 3+P(2|1,3)×1 / 3)

[0022] P23=P(2|1,3)×1 / 3 / (P(2|1,1)×1 / 3+P(2|1,2)×1 / 3+P(2|1,3)×1 / 3)

[0023] The final predicted value of the generator insulation life in year T2 is L2 = L21×P21+L22×P22+L23×P23.

[0024] As a preferred solution of the generator insulation life prediction method based on Bayesian theorem described in the present invention, wherein: the step of completing all prediction cycles includes:

[0025] The Nth prediction of the generator insulation life is carried out in TN year, and the generator insulation life is predicted to be LN1 year, LN2 year and LN3 year respectively by using the partial discharge parameter prediction method, D image method and NY image method;

[0026] The posterior probabilities P(N-1)1, P(N-1)2, and P(N-1)3 calculated for the N-1th prediction are used as the prior probabilities for the Nth prediction;

[0027] In the N-1th prediction, the distribution of the predicted value of the generator insulation life obtained by the partial discharge parameter prediction method is N(L(N-1)1,1), the distribution of the predicted value of the generator insulation life obtained by the D image prediction method is N(L(N-1)2,1), and the distribution of the predicted value of the generator insulation life obtained by the NY image prediction method is N(L(N-1)3,1);

[0028] In the Nth prediction, the distribution of the generator insulation life prediction value obtained by the partial discharge parameter prediction method is N(LN1,1), the distribution of the generator insulation life prediction value obtained by the D image prediction method is N(LN2,1), and the distribution of the generator insulation life prediction value obtained by the NY image prediction method is N(LN3,1);

[0029] Calculate the probabilities of the overlapping parts of N(LN1,1), N(LN2,1), N(LN3,1) and N(L(N-1)1+RN,1), N(L(N-1)2+RN,1), N(L(N-1)3+RN,1) on the probability distribution graph respectively as P(N|N-1,1), P(N|N-1,2), P(N|N-1,3);

[0030] The posterior probabilities of the second prediction are calculated as follows:

[0031] PN1=P(N|N-1,1)×P(N-1)1 / (P(N|N-1,1)×P(N-1)1+P(N|N-1,2)×P(N-1)2+P(N|N-1,3)×P(N-1)3)

[0032] PN2=P(N|N-1,2)×P(N-1)2 / (P(N|N-1,1)×P(N-1)1+P(N|N-1,2)×P(N-1)2+P(N|N-1,3)×P(N-1)3)

[0033] PN3=P(N|N-1,3)×P(N-1)3 / (P(N|N-1,1)×P(N-1)1+P(N|N-1,2)×P(N-1)2+P(N|N-1,3)×P(N-1)3)

[0034] The final predicted value of the generator insulation life in TN years is LN=LN1×PN1+LN2×PN2+LN3×PN3.

[0035] Another object of the present invention is to provide a generator insulation life prediction system based on Bayesian theorem. The present invention solves the problem that most traditional insulation life prediction methods are based on single test data, such as partial discharge parameters, temperature and humidity. However, these methods usually ignore the combined impact of multiple factors on insulation life. Existing life prediction methods usually do not consider the long-term cumulative degradation effect, but are only based on test data at a certain moment, making it difficult to achieve accurate life prediction. The operating environment of the generator is complex, involving the interaction of multiple physical and chemical factors. It is difficult for existing technologies to accurately predict the combined impact of multiple factors.

[0036] As a preferred solution of the generator insulation life prediction system based on Bayesian theorem described in the present invention, it is characterized by comprising: a prediction module for predicting the insulation life of the generator using at least two insulation life prediction methods, obtaining a prediction value of each method, and calculating a posterior probability of the prediction result;

[0037] The Bayesian calculation module is used to calculate the posterior probability based on the initial prior probability and predicted value of each prediction method using the Bayesian theorem, and to update the prior probability in each subsequent cycle until all prediction cycles are completed;

[0038] The data update module is used to store the updated posterior probability of each prediction value and use the updated posterior probability as the prior probability of the next prediction, thereby recursively completing the calculation of multiple prediction cycles.

[0039] A computer device includes a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of a generator insulation life prediction method based on Bayesian theorem when executing the computer program.

[0040] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a generator insulation life prediction method based on Bayesian theorem.

[0041] The beneficial effects of the present invention are as follows: First, by combining multiple prediction methods and using Bayesian theorem to dynamically update the posterior probability, the accuracy of the prediction is significantly improved, and the errors in traditional methods are reduced. Secondly, the method can be dynamically adjusted according to the previous prediction results in each prediction cycle to ensure that the prediction results are optimized as the equipment status changes, avoiding the limitations of fixed prediction models. In addition, the present invention comprehensively considers multiple factors such as partial discharge, electrical aging, thermal aging, etc., and more comprehensively evaluates the insulation life of the equipment, overcoming the problem of ignoring the interaction of multiple factors in the prior art. The method also reduces unnecessary downtime and maintenance through accurate life prediction and reasonable arrangement of maintenance cycles, reduces equipment maintenance costs, and improves the reliability and safety of the equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of 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 paying any creative work.

[0043] Figure 1 An overall flow chart of a generator insulation life prediction method based on Bayesian theorem provided in one embodiment of the present invention.

[0044] Figure 2 A flowchart of the generator insulation life prediction in year T2 of a generator insulation life prediction method based on Bayesian theorem is provided in one embodiment of the present invention.

[0045] Figure 3 A flowchart of the generator insulation life prediction for the Nth year according to a generator insulation life prediction method based on Bayesian theorem is provided in one embodiment of the present invention.

[0046] Figure 4 A probability distribution diagram of a generator insulation life prediction method based on Bayesian theorem provided in one embodiment of the present invention. DETAILED DESCRIPTION

[0047] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0048] Example 1, with reference to Figure 1 , which is the first embodiment of the present invention, provides a generator insulation life prediction method based on Bayesian theorem, comprising:

[0049] In the first prediction cycle, the insulation life of the generator is predicted using at least two insulation life prediction methods to obtain the first prediction value of each method; the same initial prior probability is set for each method, and the posterior probability of the first prediction is calculated using the Bayesian theorem based on the predicted value and the prior probability;

[0050] In the second prediction cycle, the insulation life of the generator is predicted using the same type of insulation life prediction method to obtain the second prediction value of each method; the posterior probability of the first prediction result is used as the prior probability of the second prediction to calculate the second prediction value;

[0051] In the third prediction cycle, the insulation life of the generator is predicted using the same type of insulation life prediction method to obtain the third prediction value of each method; the posterior probability of the second prediction result is used as the prior probability of the third prediction to calculate the third prediction value.

[0052] The posterior probability is updated according to each predicted value and used as the prior probability for the next prediction until all prediction cycles are completed.

[0053] The insulation life prediction method includes but is not limited to the partial discharge parameter prediction method, the D image method and the NY image method;

[0054] The partial discharge parameter prediction method measures electrical parameters, calculates the ratio of the residual breakdown voltage to the rated voltage, and predicts the remaining life of the equipment.

[0055] The D-image method predicts the residual breakdown voltage and the remaining life of the equipment through the discharge index and the maximum partial discharge amount;

[0056] The NY image method comprehensively considers the effects of electrical aging, thermal aging, and hot and cold cycles, and performs a weighted average of the residual breakdown voltage to predict the remaining life of the equipment.

[0057] The partial discharge parameter prediction method is expressed as,

[0058]

[0059] Among them, U BD is the residual breakdown voltage value; U n is the rated voltage value; Q m is the maximum partial discharge; tanδ0 is the dielectric loss tangent at 2kV; R1 is the insulation resistance value for 1min; C0 is the capacitance at 2kV.

[0060] The D image method is expressed as,

[0061]

[0062] Δ=Δtanδ+ΔI

[0063] Among them, U BD % is the predicted value of the residual breakdown voltage; Δ is the discharge index; Q maxis the maximum partial discharge under the working voltage; Δ is the discharge index representing the average aging of the insulation; Δtanδ is the increase in the dielectric loss tangent; ΔI is the AC current growth rate.

[0064] The NY image method is expressed as,

[0065]

[0066] Wherein, U0 is the initial breakdown voltage value; is the residual breakdown voltage caused by electrical aging; is the residual breakdown voltage caused by thermal aging; is the residual breakdown voltage caused by thermal cycling.

[0067] The generator insulation life prediction is carried out during the generator A-level maintenance. The parameters required for the partial discharge parameter prediction method, D image method and NY image method are obtained through the A-level maintenance.

[0068] Example 2, reference Figure 4 , which is the second embodiment of the present invention, provides a generator insulation life prediction method based on Bayesian theorem. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through experiments.

[0069] For a power plant generator, a Class A overhaul was performed in 2014, and the insulation life was first predicted. The insulation life of the generator was predicted to be 54 years, 56 years, and 49 years using the partial discharge parameter prediction method, the D image method, and the NY image method, respectively. The final predicted value of the generator insulation life in 2014 is L1 = 54 / 3 + 56 / 3 + 49 / 3 = 53 years.

[0070] like Figure 4 As shown in the figure, the generator underwent another Class A overhaul in 2019, and a second insulation life prediction was performed. The prior probabilities for the partial discharge parameter prediction method, the D image method, and the NY image method were 1 / 3. The partial discharge parameter prediction method, the D image method, and the NY image method predicted the generator insulation life to be 52 years, 40 years, and 45 years, respectively. The calculated probabilities of overlap between N(L21,1), N(L22,1), N(L23,1) and N(L11+R2,1), N(L12+R2,1), and N(L13+R2,1) on the probability distribution diagram were P(2|1,1) = 79%, P(2|1,2) = 61%, and P(2|1,3) = 72%, respectively.

[0071] The posterior probabilities of the second prediction are calculated as follows:

[0072] P21=P(2|1,1)×1 / 3 / (P(2|1,1)×1 / 3+P(2|1,2)×1 / 3+P(2|1,3)×1 / 3)=37%

[0073] P22=P(2|1,2)×1 / 3 / (P(2|1,1)×1 / 3+P(2|1,2)×1 / 3+P(2|1,3)×1 / 3)=29%

[0074] P23=P(2|1,3)×1 / 3 / (P(2|1,1)×1 / 3+P(2|1,2)×1 / 3+P(2|1,3)×1 / 3)=34%

[0075] Therefore, the final predicted value of the generator insulation life in 2019 is L2=L21×P21+L22×P22+L23×P23=46 years.

[0076] Example 3 is the third embodiment of the present invention, which differs from the first two embodiments in that:

[0077] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0078] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0079] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering, or processing in another suitable manner as necessary, and then stored in a computer memory.

[0080] It should be understood that various parts of the present invention can be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application-specific integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0081] Example 4, the fourth embodiment of the present invention, provides a generator insulation life prediction system based on Bayesian theorem, including.

[0082] A prediction module is used to predict the insulation life of the generator by using at least two insulation life prediction methods, obtain the prediction value of each method, and calculate the posterior probability of the prediction result;

[0083] The Bayesian calculation module is used to calculate the posterior probability based on the initial prior probability and predicted value of each prediction method using the Bayesian theorem, and to update the prior probability in each subsequent cycle until all prediction cycles are completed;

[0084] The data update module is used to store the updated posterior probability of each prediction value and use the updated posterior probability as the prior probability of the next prediction, thereby recursively completing the calculation of multiple prediction cycles.

[0085] Example 5, with reference to Figure 1-Figure 3 The fifth embodiment of the present invention provides a method for predicting the insulation life of a generator based on Bayesian theorem, including:

[0086] 1) If Figure 1As shown in the figure, the first prediction of the generator insulation life is made in year T1, and the generator insulation life is predicted to be L11 years, L12 years and L13 years by the partial discharge parameter prediction method, D image method and NY image method respectively;

[0087] 2) In the first prediction, the prior probabilities of the partial discharge parameter prediction method, the D image method, and the NY image method are P01, P02, and P03, respectively, and the posterior probabilities are P11, P12, and P13, respectively, with P01 = P02 = P03 = P11 = P12 = P13 = 1 / 3;

[0088] 3) The final predicted value of the generator insulation life in year T1 is L1 = L11 × P11 + L12 × P12 + L13 × P13;

[0089] 4) The predicted values ​​of generator insulation life obtained by the partial discharge parameter prediction method, D image method and NY image method are normally distributed. In the first prediction, the distribution of the predicted values ​​of generator insulation life obtained by the partial discharge parameter prediction method is N(L11,1), the distribution of the predicted values ​​of generator insulation life obtained by the D image method is N(L12,1), and the distribution of the predicted values ​​of generator insulation life obtained by the NY image method is N(L13,1);

[0090] 5) The generator insulation life prediction is carried out during the generator A-level overhaul. The parameters required for the partial discharge parameter prediction method, D image method and NY image method are obtained through the A-level overhaul. RN is the interval between the Nth A-level overhaul of the generator, N = 1, 2, 3, ..., that is, T2 = T1 + R2 years, T3 = T2 + R3 years, ..., TN = T(N-1) + RN years. The interval between A-level overhauls is usually 4 to 6 years.

[0091] 5) The posterior probabilities P(N-1)1, P(N-1)2, and P(N-1)3 calculated for the N-1th prediction are used as the prior probabilities for the Nth prediction, and the posterior probabilities PN1, PN2, and PN3 for the Nth prediction are obtained by calculation;

[0092] 6) If Figure 2 As shown in the figure, after an interval of R2 years, the second prediction of the generator insulation life is performed in year T2. The generator insulation life is predicted to be L21 years, L22 years, and L23 years by the partial discharge parameter prediction method, the D image method, and the NY image method, respectively. The prior probabilities of the partial discharge parameter prediction method, the D image method, and the NY image method in the second prediction are 1 / 3, respectively.

[0093] 7) During the second prediction, the distribution of the predicted value of the generator insulation life obtained by the partial discharge parameter prediction method is N(L21,1), the distribution of the predicted value of the generator insulation life obtained by the D image prediction method is N(L22,1), and the distribution of the predicted value of the generator insulation life obtained by the NY image prediction method is N(L23,1);

[0094] The probabilities of the overlapping parts of N(L21,1), N(L22,1), N(L23,1) and N(L11+R2,1), N(L12+R2,1), N(L13+R2,1) on the probability distribution graph are calculated as P(2|1,1), P(2|1,2), P(2|1,3) respectively.

[0095] The posterior probabilities of the second prediction are calculated as follows:

[0096] P21=P(2|1,1)×1 / 3 / (P(2|1,1)×1 / 3+P(2|1,2)×1 / 3+P(2|1,3)×1 / 3)

[0097] P22=P(2|1,2)×1 / 3 / (P(2|1,1)×1 / 3+P(2|1,2)×1 / 3+P(2|1,3)×1 / 3)

[0098] P23=P(2|1,3)×1 / 3 / (P(2|1,1)×1 / 3+P(2|1,2)×1 / 3+P(2|1,3)×1 / 3)

[0099] Therefore, the final predicted value of the generator insulation life in year T2 is L2=L21×P21+L22×P22+L23×P23

[0100] 8) If Figure 3 As shown in the figure, by analogy, the Nth prediction of the generator insulation life is performed in TN years. The generator insulation life is predicted to be LN1 years, LN2 years, and LN3 years by the partial discharge parameter prediction method, D image method, and NY image method respectively; N = 1, 2, 3, ...

[0101] 9) The posterior probabilities P(N-1)1, P(N-1)2, and P(N-1)3 calculated for the N-1th prediction are used as the prior probabilities for the Nth prediction.

[0102] 10) In the N-1th prediction, the distribution of the predicted value of the generator insulation life obtained by the partial discharge parameter prediction method is N(L(N-1)1,1), the distribution of the predicted value of the generator insulation life obtained by the D image prediction method is N(L(N-1)2,1), and the distribution of the predicted value of the generator insulation life obtained by the NY image prediction method is N(L(N-1)3,1);

[0103] 11) During the Nth prediction, the distribution of the predicted value of the generator insulation life obtained by the partial discharge parameter prediction method is N(LN1,1), the distribution of the predicted value of the generator insulation life obtained by the D image prediction method is N(LN2,1), and the distribution of the predicted value of the generator insulation life obtained by the NY image prediction method is N(LN3,1);

[0104] Calculate the probabilities of the overlapping parts of N(LN1,1), N(LN2,1), N(LN3,1) and N(L(N-1)1+RN,1), N(L(N-1)2+RN,1), N(L(N-1)3+RN,1) on the probability distribution graph respectively, which are P(N|N-1,1), P(N|N-1,2), P(N|N-1,3).

[0105] The posterior probabilities of the second prediction are calculated as follows:

[0106] PN1=P(N|N-1,1)×P(N-1)1 / (P(N|N-1,1)×P(N-1)1+P(N|N-1,2)×P(N-1)2+P(N|N-1,3)×P(N-1)3)

[0107] PN2=P(N|N-1,2)×P(N-1)2 / (P(N|N-1,1)×P(N-1)1+P(N|N-1,2)×P(N-1)2+P(N|N-1,3)×P(N-1)3)

[0108] PN3=P(N|N-1,3)×P(N-1)3 / (P(N|N-1,1)×P(N-1)1+P(N|N-1,2)×P(N-1)2+P(N|N-1,3)×P(N-1)3)

[0109] 10) The final predicted value of the generator insulation life in TN years is LN = LN1×PN1+LN2×PN2+LN3×PN3.

[0110] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to 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 method for predicting the insulation life of a generator based on Bayesian theorem, characterized by: include, In the first prediction cycle, the insulation life of the generator is predicted using at least two insulation life prediction methods to obtain the first prediction value of each method; the same initial prior probability is set for each method, and the posterior probability of the first prediction is calculated using the Bayesian theorem based on the predicted value and the prior probability; In the second prediction cycle, the insulation life of the generator is predicted using the same type of insulation life prediction method to obtain the second prediction value of each method; the posterior probability of the first prediction result is used as the prior probability of the second prediction to calculate the second prediction value; In the third prediction cycle, the insulation life of the generator is predicted using the same type of insulation life prediction method to obtain the third prediction value of each method; the posterior probability of the second prediction result is used as the prior probability of the third prediction to calculate the third prediction value; The posterior probability is updated according to each predicted value and used as the prior probability for the next prediction until all prediction cycles are completed.

2. The method for predicting the insulation life of a generator based on Bayesian theorem according to claim 1, wherein: The insulation life prediction method includes a partial discharge parameter prediction method, a D image method and a NY image method; The partial discharge parameter prediction method measures electrical parameters and calculates the ratio of the residual breakdown voltage to the rated voltage to predict the remaining life of the equipment. The D-image method predicts the residual breakdown voltage and the remaining life of the device through the discharge index and the maximum partial discharge amount; The NY image method comprehensively considers the effects of electrical aging, thermal aging, and thermal cycling, performs a weighted average of the residual breakdown voltage, and predicts the remaining life of the device.

3. The method for predicting the insulation life of a generator based on Bayesian theorem according to claim 2, characterized in that: The partial discharge parameter prediction method is expressed as: Among them, U BD is the residual breakdown voltage value; U n is the rated voltage value; Q m is the maximum partial discharge; tanδ0 is the dielectric loss tangent at 2kV; R1 is the insulation resistance value for 1min; C0 is the capacitance at 2kV.

4. The method for predicting the insulation life of a generator based on Bayesian theorem according to claim 3, wherein: The D image method is expressed as, Δ=Δtanδ+ΔI Among them, U BD % is the predicted value of the residual breakdown voltage; Δ is the discharge index; Q max is the maximum partial discharge under the working voltage; Δ is the discharge index representing the average aging of the insulation; Δtanδ is the increase in the dielectric loss tangent; ΔI is the AC current growth rate.

5. The method for predicting the insulation life of a generator based on Bayesian theorem according to claim 4, characterized in that: The NY image method is expressed as, Wherein, U0 is the initial breakdown voltage value; is the residual breakdown voltage caused by electrical aging; is the residual breakdown voltage caused by thermal aging; is the residual breakdown voltage caused by thermal cycling.

6. The method for predicting the insulation life of a generator based on Bayesian theorem according to claim 5, characterized in that: The calculating of the second prediction value includes: The distribution of the predicted value of the generator insulation life obtained by the partial discharge parameter prediction method is N(L21,1), the distribution of the predicted value of the generator insulation life obtained by the D image prediction method is N(L22,1), and the distribution of the predicted value of the generator insulation life obtained by the NY image prediction method is N(L23,1); Calculate the probabilities of the overlapping parts of N(L21,1), N(L22,1), N(L23,1) and N(L11+R2,1), N(L12+R2,1), N(L13+R2,1) on the probability distribution graph respectively as P(2|1,1), P(2|1,2), P(2|1,3); The posterior probabilities of the second prediction are calculated as follows: P21=P(2|1,1)×1 / 3 / (P(2|1,1)×1 / 3+P(2|1,2)×1 / 3+P(2|1,3)×1 / 3) P22=P(2|1,2)×1 / 3 / (P(2|1,1)×1 / 3+P(2|1,2)×1 / 3+P(2|1,3)×1 / 3) P23=P(2|1,3)×1 / 3 / (P(2|1,1)×1 / 3+P(2|1,2)×1 / 3+P(2|1,3)×1 / 3) The final predicted value of the generator insulation life in year T2 is L2 = L21×P21+L22×P22+L23×P23.

7. The method for predicting the insulation life of a generator based on Bayesian theorem according to claim 6, characterized in that: The above-mentioned period until all forecast cycles are completed includes, The Nth prediction of the generator insulation life is carried out in TN year, and the generator insulation life is predicted to be LN1 year, LN2 year and LN3 year by using the partial discharge parameter prediction method, D image method and NY image method respectively; The posterior probabilities P(N-1)1, P(N-1)2, and P(N-1)3 calculated for the N-1th prediction are used as the prior probabilities for the Nth prediction; In the N-1th prediction, the distribution of the predicted value of the generator insulation life obtained by the partial discharge parameter prediction method is N(L(N-1)1,1), the distribution of the predicted value of the generator insulation life obtained by the D image prediction method is N(L(N-1)2,1), and the distribution of the predicted value of the generator insulation life obtained by the NY image prediction method is N(L(N-1)3,1); In the Nth prediction, the distribution of the generator insulation life prediction value obtained by the partial discharge parameter prediction method is N(LN1,1), the distribution of the generator insulation life prediction value obtained by the D image prediction method is N(LN2,1), and the distribution of the generator insulation life prediction value obtained by the NY image prediction method is N(LN3,1); Calculate the probabilities of the overlapping parts of N(LN1,1), N(LN2,1), N(LN3,1) and N(L(N-1)1+RN,1), N(L(N-1)2+RN,1), N(L(N-1)3+RN,1) on the probability distribution graph respectively as P(N|N-1,1), P(N|N-1,2), P(N|N-1,3); The posterior probabilities of the second prediction are calculated as follows: PN1=P(N|N-1,1)×P(N-1)1 / (P(N|N-1,1)×P(N-1)1+P(N|N-1,2)×P(N-1)2+P(N|N-1,3)×P(N-1)3) PN2=P(N|N-1,2)×P(N-1)2 / (P(N|N-1,1)×P(N-1)1+P(N|N-1,2)×P(N-1)2+P(N|N-1,3)×P(N-1)3) PN3=P(N|N-1,3)×P(N-1)3 / (P(N|N-1,1)×P(N-1)1+P(N|N-1,2)×P(N-1)2+P(N|N-1,3)×P(N-1)3) The final predicted value of the generator insulation life in TN years is LN=LN1×PN1+LN2×PN2+LN3×PN3.

8. A generator insulation life prediction system based on Bayesian theorem, applying a generator insulation life prediction method based on Bayesian theorem as claimed in any one of claims 1 to 7, characterized in that: include: A prediction module is used to predict the insulation life of the generator by using at least two insulation life prediction methods, obtain the prediction value of each method, and calculate the posterior probability of the prediction result; The Bayesian calculation module is used to calculate the posterior probability based on the initial prior probability and predicted value of each prediction method using the Bayesian theorem, and to update the prior probability in each subsequent cycle until all prediction cycles are completed; The data update module is used to store the updated posterior probability of each prediction value and use the updated posterior probability as the prior probability of the next prediction, thereby recursively completing the calculation of multiple prediction cycles.

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 processor implements the steps of a generator insulation life prediction method based on Bayesian theorem according to any one of claims 1 to 7.

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 a generator insulation life prediction method based on Bayesian theorem according to any one of claims 1 to 7 are implemented.

Images