Method for evaluating the performance of the internal insulating medium of oil-immersed capacitors

Abstract

The application discloses a performance evaluation method for an internal insulating medium of an oil-immersed capacitor based on a RBF neural network, and the method comprises the following steps: measuring characteristic variables of the internal insulating medium of the oil-immersed capacitor to construct training samples; constructing a RBF neural network model; optimizing initial weights and thresholds of the RBF neural network model; iteratively training the RBF neural network model by using the training samples to obtain a trained RBF neural network model; and evaluating the performance of the RBF neural network model by using the trained RBF neural network model.

Description

Technical Field

[0001] This invention belongs to the technical field of performance evaluation of internal insulating media in oil-immersed capacitors, and in particular, a performance evaluation method for internal insulating media in oil-immersed capacitors based on RBF neural networks. Background Technology

[0002] Oil-immersed AC capacitors are the primary reactive power compensation and filtering devices in AC and DC power transmission technologies. While my country has mastered the key technologies for oil-immersed capacitors, the polypropylene resin used in the roughening film remains entirely dependent on imports. The production volume and price of polypropylene resin, especially high-end products, are thus subject to foreign control.

[0003] The ash content, isotacticity, molecular weight distribution, and molecular chain structure of polypropylene resin determine the structure and properties of biaxially oriented polypropylene (BOPP) film. Compared with dry capacitor film materials, BOPP film for oil-immersed capacitors requires a higher surface roughness. Its rough structure is beneficial to the bonding and compatibility between aluminum foil and insulating oil. This unique roughened surface structure and its impregnation characteristics in capacitor oil are closely related to the isotacticity and molecular structure of polypropylene resin. Simultaneously, the balance between ash content, temperature resistance, and roughness is achieved by controlling the resin's ash content, isotacticity, and molecular weight distribution. Different requirements for film and resin properties lead to incompatibility in polypropylene resin polymerization processes and additive formulations. Optimizing the molecular chain configuration and arrangement of BOPP film is crucial to reducing electrical weaknesses, improving the aggregated structure, and enhancing insulation performance and impregnation characteristics. Furthermore, the morphology and distribution of electrical weaknesses, surface roughness, and crystallinity of the roughened BOPP film for oil-immersed capacitors have a significant impact on the capacitor's insulation design and vacuum impregnation process, thus affecting key performance characteristics such as dielectric loss, partial discharge, thermal stability, and durability. The insulating dielectric material of oil-immersed capacitors mainly consists of insulating oil and a thin film. However, there is currently a lack of testing and evaluation methods for the internal insulating dielectric material of capacitors, resulting in insufficient assessment of the insulation performance of capacitors during long-term operation. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a performance evaluation method for the internal insulating medium of oil-immersed capacitors based on RBF neural networks, which more accurately evaluates the performance of the internal insulating medium of oil-immersed capacitors operating for extended periods.

[0005] The performance evaluation methods for the internal insulating dielectric of oil-immersed capacitors based on RBF neural networks include:

[0006] Step S1: Measure the characteristic variables of the internal insulating medium of the oil-immersed capacitor to construct training samples;

[0007] Step S2: Construct the RBF neural network model;

[0008] Step S3: Optimize the initial weights and thresholds of the RBF neural network model;

[0009] Step S4: Iteratively train the RBF neural network model using training samples to obtain a trained RBF neural network model.

[0010] Step S5: Evaluate the performance of the RBF neural network model using the trained RBF neural network model.

[0011] In the aforementioned performance evaluation method for the internal insulating medium of an oil-immersed capacitor based on an RBF neural network, step S1 involves measuring the glass transition temperature, crystallinity, DC resistivity, and dielectric loss of the internal insulating medium of the oil-immersed capacitor as characteristic variables.

[0012] In the aforementioned performance evaluation method for the internal insulating medium of an oil-immersed capacitor based on an RBF neural network, glass transition temperature, crystallinity, DC resistivity, dielectric loss, and service time are used as input parameters, and the measured DC breakdown field strength of the insulating medium is used as the output parameter. The input and output parameters are normalized to obtain training samples.

[0013] In the performance evaluation method of the internal insulating medium of oil-immersed capacitor based on RBF neural network, the RBF neural network model includes an input layer, a hidden layer, and an output layer. The number of input parameters is used as the number of nodes in the input layer N, the number of output parameters is used as the number of nodes in the output layer M, and the number of nodes in the hidden layer L is determined according to the training results of the training samples.

[0014] In the aforementioned performance evaluation method for the internal insulating dielectric of an oil-immersed capacitor based on an RBF neural network, step S3 includes:

[0015] S31, Population initialization, uses four parts—the connection weights between the input and hidden layers, the weights between the hidden and output layers, the hidden layer threshold, and the output layer threshold—as the genetic information that an individual should possess, and encodes them using real numbers. The chromosome encoding length S satisfies:

[0016] S = N·L + M·L + N + M + L,

[0017] S32, the reciprocal of the sum of the absolute values ​​of the errors of each prediction is defined as the fitness function;

[0018] S33 uses a roulette wheel selection method, where individuals with higher fitness values ​​have a greater probability of being selected.

[0019] S34, using real number encoding, employs real number crossover to perform crossover operations;

[0020] S35: Randomly select individuals from the population and mutate them randomly according to a predetermined probability to obtain new individuals.

[0021] Compared with the prior art, the present invention has the following advantages:

[0022] RBF neural networks belong to the feedforward neural network category, and their network structure is similar to that of multilayer feedforward networks. The first layer is the input layer, consisting of signal source nodes; the second layer is the hidden layer, the number of which depends on the needs of the problem being described. The transformation function of the neurons in the hidden layer, i.e., the radial basis function, is a non-negative nonlinear function that is radially symmetric and decays around the center point; the third layer is the output layer, which responds to the input pattern. Using RBF neural networks can more accurately evaluate the performance of the internal insulating dielectric of oil-immersed capacitors operating for long periods. Attached Figure Description

[0023] The accompanying drawings illustrate exemplary embodiments of the invention and, together with the description thereof, serve to explain the principles of the invention. These drawings are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification.

[0024] Figure 1 This is a flowchart illustrating a method for evaluating the performance of the internal insulating medium of an oil-immersed capacitor based on an RBF neural network, according to one embodiment of the present invention.

[0025] Figure 2 A schematic diagram of the RBF neural network model of the performance evaluation method for the internal insulating medium of an oil-immersed capacitor based on the RBF neural network in one embodiment of the present invention.

[0026] Figure 3 This is a schematic diagram illustrating the prediction results of a performance evaluation method for the internal insulating medium of an oil-immersed capacitor based on an RBF neural network, according to one embodiment of the present invention. Detailed Implementation

[0027] The following is in conjunction with the appendix Figures 1 to 3 The present invention will be further described in detail below with reference to the embodiments. It is to be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention. Furthermore, it should be noted that, for ease of description, only the parts relevant to the present invention are shown in the accompanying drawings.

[0028] It should be noted that, unless otherwise specified, the embodiments and features described in this invention can be combined with each other. The technical solution of this invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0029] Unless otherwise stated, the exemplary embodiments / exemplifications shown are to be understood as providing exemplary features of various details that provide ways in which the technical concept of the invention can be implemented in practice. Therefore, unless otherwise stated, the features of the various embodiments / exemplifications may be additionally combined, separated, interchanged and / or rearranged without departing from the technical concept of the invention.

[0030] Crosshairs and / or shading may be used in the accompanying drawings to clarify the boundaries between adjacent components. Thus, unless otherwise stated, the presence or absence of crosshairs or shading does not convey or indicate any preference or requirement for the specific material, material properties, dimensions, proportions, commonalities between the illustrated components, or any other characteristics, properties, etc., of the components. Furthermore, the dimensions and relative dimensions of components may be exaggerated in the accompanying drawings for clarity and / or descriptive purposes. When exemplary embodiments can be implemented differently, a specific process sequence may be performed in a manner different from the described order of steps. For example, two consecutively described processes may be performed substantially simultaneously or in the reverse order of the described process. Moreover, the same reference numerals denote the same components.

[0031] When a component is referred to as being "on" or "above" another component, "connected to," or "joined to" another component, the component may be directly on, directly connected to, or directly joined to the other component, or there may be intermediate components. However, when a component is referred to as being "directly on" another component, "directly connected to," or "directly joined to" another component, there are no intermediate components. Therefore, the term "connection" can refer to a physical connection, an electrical connection, etc., and may or may not have intermediate components.

[0032] For descriptive purposes, the present invention may use spatial relative terms such as “below,” “under,” “below,” “down,” “above,” “above,” “higher,” and “side (e.g., as in a “sidewall”)” to describe the relationship between one component and another component as shown in the accompanying drawings. In addition to the orientations depicted in the drawings, the spatial relative terms are also intended to encompass different orientations of the device during use, operation, and / or manufacture. For example, if the device in the drawings is flipped, a component described as “below” or “under” another component or feature would subsequently be positioned “above” said other component or feature. Thus, the exemplary term “below” can encompass both “above” and “below” orientations. Furthermore, the device may be otherwise positioned (e.g., rotated 90 degrees or in other orientations), thus interpreting the spatial relative descriptive terms used herein accordingly.

[0033] The terminology used herein is for the purpose of describing particular embodiments and is not intended to be limiting. As used herein, unless the context clearly indicates otherwise, the singular forms “a” and “the” are intended to include the plural forms as well. Furthermore, when the terms “comprising” and / or “including” and variations thereof are used in this specification, it indicates the presence of the stated features, integrals, steps, operations, parts, components, and / or groups thereof, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, parts, components, and / or groups thereof. It should also be noted that, as used herein, the terms “substantially,” “about,” and other similar terms are used as approximate terms rather than as terms of degree, thus explaining the inherent biases in measurements, calculated values, and / or provided values ​​that would be recognized by one of ordinary skill in the art.

[0034] See Figures 1 to 3 In one embodiment, the performance evaluation method for the internal insulating dielectric of an oil-immersed capacitor based on an RBF neural network according to the present invention includes:

[0035] Step S1: Measure the characteristic variables of the internal insulating medium of the oil-immersed capacitor to construct training samples;

[0036] Step S2: Construct the RBF neural network model;

[0037] Step S3: Optimize the initial weights and thresholds of the RBF neural network model;

[0038] Step S4: Iteratively train the RBF neural network model using training samples to obtain a trained RBF neural network model.

[0039] Step S5: Evaluate the performance of the RBF neural network model using the trained RBF neural network model.

[0040] In a preferred embodiment of the performance evaluation method for the internal insulating medium of an oil-immersed capacitor based on an RBF neural network, step S1 involves measuring the glass transition temperature, crystallinity, DC resistivity, and dielectric loss of the internal insulating medium of the oil-immersed capacitor as characteristic variables.

[0041] In a preferred embodiment of the performance evaluation method for the internal insulating medium of an oil-immersed capacitor based on an RBF neural network, the glass transition temperature, crystallinity, DC resistivity, dielectric loss, and service time are used as input parameters, and the measured DC breakdown field strength of the insulating medium is used as the output parameter. The input and output parameters are normalized to obtain training samples.

[0042] In a preferred embodiment of the performance evaluation method for the internal insulating medium of an oil-immersed capacitor based on an RBF neural network, the RBF neural network model includes an input layer, a hidden layer, and an output layer. The number of input parameters is used as the number of nodes N in the input layer, and the number of output parameters is used as the number of nodes M in the output layer. The number of nodes L in the hidden layer is determined based on the training results of the training samples.

[0043] In a preferred embodiment of the performance evaluation method for the internal insulating dielectric of an oil-immersed capacitor based on an RBF neural network, step S3 includes:

[0044] S31, Population initialization, uses four parts—the connection weights between the input and hidden layers, the weights between the hidden and output layers, the hidden layer threshold, and the output layer threshold—as the genetic information that an individual should possess, and encodes them using real numbers. The chromosome encoding length S satisfies:

[0045] S = N·L + M·L + N + M + L,

[0046] S32, the reciprocal of the sum of the absolute values ​​of the errors of each prediction is defined as the fitness function;

[0047] S33 uses a roulette wheel selection method, where individuals with higher fitness values ​​have a greater probability of being selected.

[0048] S34, using real number encoding, employs real number crossover to perform crossover operations;

[0049] S35: Randomly select individuals from the population and mutate them randomly according to a predetermined probability to obtain new individuals.

[0050] like Figure 1 As shown, the characteristic variables of the insulating dielectric of an oil-immersed capacitor are used as training samples to establish the topology of an RBF neural network model. A genetic algorithm is then used to optimize the weights and thresholds of the RBF neural network. The elements of the genetic algorithm for optimizing the RBF neural network include population initialization, fitness function, selection operation, crossover operation, and mutation operation.

[0051] 1. Population initialization: The individual encoding method is real number encoding, which consists of four parts: the connection weight between the input layer and the hidden layer, the hidden layer threshold, the connection weight between the hidden layer and the output layer, and the output layer threshold.

[0052] 2. Fitness function: Based on the initial weights and thresholds of the RBF neural network obtained from the individual, the neural network is trained with training data to predict the system output, and the absolute value of the error between the predicted output and the expected output is used as the individual fitness value F.

[0053] 3. Selection Operation: The selection strategy is a roulette wheel selection, i.e., a selection strategy based on fitness ratios. The selection probability pi for each individual i is:

[0054]

[0055]

[0056] In the formula, Fi is the fitness value of individual i. Since a smaller fitness value is better, the reciprocal of the fitness value is taken before individual selection; k is a coefficient; and N is the number of individuals in the population.

[0057] 4. Crossover operation: The crossover operation method uses real number crossover. The crossover operation method for the k-th chromosome ak and the 1st chromosome al at position j is as follows:

[0058]

[0059] In the formula, b is a random number between [0, 1].

[0060] 5. Mutation operation: Randomly select an individual from the population and mutate it with a 20% probability to obtain a new individual.

[0061] The characteristic variables of the internal dielectric of different oil-immersed capacitors were measured, and a total of 45 sets of data were obtained, of which 36 sets were used as the training set and 9 sets were used as the test set. Some of the original data are shown in the table below:

[0062]

[0063] After obtaining the performance evaluation model of the insulating dielectric of the oil-immersed capacitor through the training set, the DC breakdown field strength of the insulating dielectric is predicted using the test set. The prediction results are as follows: Figure 3 As shown in the figure, the actual values ​​and predicted values ​​are basically consistent. The prediction error remains within 5%, indicating high accuracy.

[0064] In the description of this specification, the references to terms such as "one embodiment / mode," "some embodiments / modes," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment / mode or example is included in at least one embodiment / mode or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment / mode or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments / modes or examples. Furthermore, without contradiction, those skilled in the art can combine and integrate the different embodiments / modes or examples described in this specification, as well as the features of different embodiments / modes or examples.

[0065] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0066] Those skilled in the art should understand that the above embodiments are merely for illustrating the present invention and are not intended to limit the scope of the invention. Those skilled in the art can make other changes or modifications based on the above disclosure, and these changes or modifications still fall within the scope of the present invention.

Claims

1. A method for evaluating the performance of the internal insulating dielectric of an oil-immersed capacitor based on an RBF neural network, characterized in that, It includes the following steps: Step S1: Measure the characteristic variables of the internal insulating medium of the oil-immersed capacitor to construct training samples; Step S2: Construct the RBF neural network model; Step S3: Optimize the initial weights and thresholds of the RBF neural network model; Step S4: Iteratively train the RBF neural network model using training samples to obtain a trained RBF neural network model. Step S5: Evaluate the performance of the RBF neural network model using the trained RBF neural network model; The glass transition temperature, crystallinity, DC resistivity, dielectric loss, and service time are used as input parameters, and the measured DC breakdown field strength of the insulating medium is used as the output parameter. The input and output parameters are normalized to obtain training samples.

2. The performance evaluation method for the internal insulating dielectric of an oil-immersed capacitor based on an RBF neural network according to claim 1, characterized in that, The RBF neural network model includes an input layer, a hidden layer, and an output layer. The number of input parameters is used as the number of nodes in the input layer, N, and the number of output parameters is used as the number of nodes in the output layer, M. The number of hidden layer nodes L is determined based on the training results of the training samples.

3. The performance evaluation method for the internal insulating dielectric of an oil-immersed capacitor based on an RBF neural network according to claim 1, characterized in that, Step S3 includes, S31, Population initialization, uses four parts—the connection weights between the input and hidden layers, the weights between the hidden and output layers, the hidden layer threshold, and the output layer threshold—as the genetic information that an individual should possess, and encodes them using real numbers. The chromosome encoding length S satisfies: ， S32, the reciprocal of the sum of the absolute values ​​of the errors of each prediction is defined as the fitness function; S33 uses a roulette wheel selection method, where individuals with higher fitness values ​​have a greater probability of being selected. S34, using real number encoding, employs real number crossover to perform crossover operations; S35: Randomly select individuals from the population and mutate them randomly according to a predetermined probability to obtain new individuals.

Images