A basic probability assignment method for gas turbine fault diagnosis

Abstract

The present application belongs to the field of gas turbine fault diagnosis, and discloses a basic probability assignment method for gas turbine fault diagnosis. The present application mainly comprises the following steps: step one: determining a fault identification framework according to the fault category of the gas turbine; step two: obtaining data after normalization processing of the fault sample features; step three: calculating the mean value of the historical fault samples of each fault category; step four: calculating the distance proximity coefficient according to the current fault sample; step five: calculating the direction proximity coefficient according to the current fault sample; step six: calculating the membership of the current fault sample to each fault probability according to the distance proximity coefficient and the direction proximity coefficient; and step seven: normalizing the membership fault probability to obtain the basic probability assignment of each fault category. The present application solves the problems of the existing basic probability assignment method, such as limitation, poor generalization ability, and insufficient sample information mining, when the information fusion technology based on D-S evidence theory is used in the field of gas turbine fault diagnosis.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of gas turbine fault diagnosis, and particularly relates to a basic probability assignment method for gas turbine fault diagnosis. BACKGROUND

[0002] The gas turbine is a new type of power equipment, and is widely applied to aerospace, land and sea transportation, energy, power and other fields due to high thermal efficiency, stable operation, rapid start response and other advantages. However, the gas turbine has a complex structure, and once a fault occurs, it will cause serious consequences to enterprise production and bring great economic losses. Therefore, it is of important theoretical significance and application value to find a reasonable gas turbine fault diagnosis method to timely judge the fault category and make treatment at the early stage of fault occurrence, reduce downtime maintenance time and reduce maintenance cost.

[0003] Multi-source information fusion is a front technology and research hotspot for solving the problem of gas turbine fault diagnosis, and the D-S evidence theory is most widely applied in the decision layer fusion algorithm, and has become one of the important technologies from theoretical methods to practical applications. The basic probability assignment is the most critical and difficult step of combining the multi-source information fusion technology and application based on the D-S evidence theory, and directly affects the accuracy of the fault diagnosis result.

[0004] At present, some basic probability assignment methods such as using environmental coefficient weighting, using statistical evidence and using acceleration have limitations, are limited to the use environment, cannot be widely applied, and do not fully mine the information relationship of the sample. SUMMARY

[0005] The purpose of the application is to solve the problems that the existing basic probability assignment method has limitations, poor generalization ability and insufficient mining of sample information when the information fusion technology based on the D-S evidence theory is used in the field of gas turbine fault diagnosis. A basic probability assignment method for gas turbine fault diagnosis.

[0006] A basic probability assignment method for gas turbine fault diagnosis comprises the following steps:

[0007] S1: According to the fault category of the gas turbine, a fault recognition framework Y={y1, y2,..., yn} is determined, wherein n represents the number of fault categories. n} is determined, wherein n represents the number of fault categories.

[0008] S2: The jth sample of the ith fault of the gas turbine is represented as X ij =[x ij1 ,x ij2 ,...,x ijk ] after normalization processing of the characteristics, wherein X ij ∈R 1×kis a 1xk dimensional vector, k represents the dimension of the feature;

[0009] S3: calculating the mean value of the i-th type of historical fault sample:

[0010]

[0011] Wherein, M i represents the number of samples of the i-th type of fault;

[0012] S4: for a certain fault sample Q = [q1, q2,..., q k ], wherein Q ∈ R 1×k , the distance similarity coefficient between the fault sample and the mean value of the i-th type of historical fault sample is calculated:

[0013]

[0014] S5: calculating the direction similarity coefficient between the fault sample and the mean value of the i-th type of historical fault sample:

[0015]

[0016] S6: according to the distance similarity coefficient and the direction similarity coefficient, the probability that the certain fault sample belongs to the i-th type of fault is calculated:

[0017]

[0018] S7: the probability is normalized to obtain the basic probability assignment of the i-th type of fault:

[0019]

[0020] The application has the advantages that the application has wide application range, and is not limited by environment, and can be popularized to other fields. The application fully excavates historical fault sample information, calculates the distance similarity coefficient and the direction similarity coefficient between the current fault sample and the mean value of the historical fault sample, and obtains more accurate basic probability assignment. The application has simple implementation process and is easy to realize. BRIEF DESCRIPTION OF DRAWINGS

[0021] Figure 1 The flowchart of the application. DETAILED DESCRIPTION

[0022] The application will be further described below with reference to the drawings.

[0023] As Figure 1 shown, a basic probability assignment method for gas turbine fault diagnosis comprises the following steps:

[0024] S1: Based on the gas turbine fault category, determine the fault identification framework Y = {y1, y2, ..., y} n}, where n represents the number of fault categories;

[0025] S2: The j-th sample of the i-th type of fault in the gas turbine is represented by X after feature normalization. ij =[x ij1 ,x ij2 ,...,x ijk ], where X ij ∈R 1×k It is a 1×k dimensional vector, where k represents the dimension of the feature;

[0026] S3: Calculate the mean for the i-th type of historical fault samples:

[0027]

[0028] Among them, M i This represents the number of samples of the i-th type of fault;

[0029] S4: For a certain fault sample Q = [q1, q2, ..., q...] k ], where Q∈R 1×k Calculate the similarity coefficient between the fault sample and the mean of the i-th type of historical fault samples:

[0030]

[0031] S5: Calculate the directional similarity coefficient between this fault sample and the mean of the historical fault samples of the i-th type:

[0032]

[0033] S6: Based on the proximity coefficient and direction proximity coefficient, calculate the probability that a certain fault sample belongs to the i-th type of fault:

[0034]

[0035] S7: Normalize the probabilities to obtain the basic probability assignment for the i-th type of fault:

[0036]

[0037] The specific embodiments of the present invention will be described in detail below through examples:

[0038] ① Based on the gas turbine fault categories, determine the fault identification framework Y = {y1, y2, y3}.

[0039] ②The following is the data after obtaining the fault samples corresponding to each fault category and normalizing the features:

[0040]

[0041] ③Calculate the mean value of historical fault samples according to formula (1):

[0042]

[0043]

[0044]

[0045] ④Assume that the current fault sample is Q = [0.5, 0.7], calculate the distance proximity coefficient according to formula (2):

[0046]

[0047]

[0048]

[0049] ⑤Calculate the direction proximity coefficient according to formula (3):

[0050]

[0051]

[0052]

[0053] ⑥Calculate the probability that the current fault sample belongs to different fault categories according to formula (4):

[0054] p1 = 0.565 + 0.5 x 0.997 = 1.064 (15)

[0055] p2 = 0.932 + 0.5 x 1.000 = 1.432 (16)

[0056] p3 = 0.668 + 0.5 x 0.991 = 1.164 (17)

[0057] ⑦Calculate the basic probability assignment of each fault category according to formula (5):

[0058]

[0059]

[0060]

[0061] Take this as the basic probability assignment of the fault sample belonging to each fault category.

[0062] The above merely provides the preferred embodiments of the present application, and is not used to limit the present application. For those skilled in the art, the present application can have various modifications and changes. Any modifications, equivalent replacements, improvements, etc. made within the principles and technical scope of the present application shall fall into the scope of the present application.

Claims

1. A basic probability assignment method for gas turbine fault diagnosis, characterized by, The method comprises the following steps: S1: Based on the gas turbine fault category, determine the fault identification framework Y = {y1, y2, ..., y} n }, where n represents the number of fault categories; S2: the i-th type of fault of the gas turbine, the j-th sample of the feature after normalization is represented as X ij = [x ij1 ,x ij2 ,...,x ijk ] , wherein X ij ∈R 1×k is a 1×k dimensional vector, and k represents the dimension of the feature; S3: calculating the mean value of the historical fault samples of the i-th type: wherein M i represents the number of samples of the i-th type of failure; S4: For a current fault sample Q = [q1, q2,..., q k ], where Q ∈ R 1×k , compute the distance proximity coefficient between the current fault sample and the mean of the historical fault samples of the ith class: S5: calculating the direction similarity coefficient between the fault sample and the mean value of the historical fault samples of the i-th type: S6: calculating the probability that the current fault sample belongs to the i-th type according to the distance similarity coefficient and the direction similarity coefficient: S7: normalizing the probability to obtain the basic probability assignment of the i-th type of fault: 。

Images