Method, device and system for monitoring concentration of dissolved gases in transformer oil

Abstract

This invention provides a method, apparatus, and system for monitoring the concentration of dissolved gases in transformer oil. The concentration monitoring method includes: obtaining initial model parameters of a gas detection model using the Firefly algorithm; training the gas detection model based on the initial model parameters and a sample training set to obtain the trained gas detection model; acquiring in real-time gas concentration feature values ​​of at least one gas in the transformer oil detected by a gas concentration feature value detection system; and applying the trained gas detection model to monitor the liquid-phase gas concentration of dissolved gases in the transformer oil in real-time based on the gas concentration feature values. This invention avoids cross-interference of dissolved gases in transformer oil and reduces the computational complexity of dissolving gas concentration in transformer oil.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of dissolved gas detection in transformer oil, and particularly relates to a concentration monitoring method, device and system for dissolved gas in transformer oil. BACKGROUND

[0002] The existing parameter calibration of the dissolved gas detection system in transformer oil mostly adopts the method of "gas concentration characteristic value-gas concentration-liquid phase concentration in transformer oil". According to the Beer-Lambert law, the Tunable Diode Laser Absorption Spectroscopy (TDLAS) is applied to obtain the gas concentration characteristic value of the measured gas according to the incident light intensity and the exit light intensity of the laser, and the linear relationship between the gas concentration characteristic value and the measured gas concentration.

[0003] In the calibration process of "gas concentration characteristic value-gas concentration", there is cumulative fitting error, therefore, the calibration model of "gas concentration characteristic value-liquid phase concentration in transformer oil" is adopted, and the parameters of the calibration model are affected by the cross interference of multi-component gases in the dissolved gas detection system in transformer oil, and the conversion parameters are difficult to identify in the process of "gas concentration-liquid phase concentration in transformer oil". The wavelet neural network method can realize the identification of the conversion parameters and the cross interference weight. Considering that the performance of the wavelet neural network is closely related to the network parameters, the gradient descent method is mostly adopted to optimize the network parameters at present, the change direction of which is relatively fixed, which limits the optimization direction of the wavelet neural network parameters, and the wavelet neural network is prone to fall into local minimum, which is not conducive to the identification of the conversion parameters and the cross interference weight. SUMMARY

[0004] The application provides a concentration monitoring method, device and system for dissolved gas in transformer oil, which solves the technical problem that the parameters are difficult to identify in the dissolved gas detection process in transformer oil, and avoids the influence of the cross interference of the dissolved gas in transformer oil.

[0005] In order to achieve the above purpose, the application provides a concentration monitoring method for dissolved gas in transformer oil, which comprises the following steps: obtaining initial model parameters of a gas detection model by using a glowworm algorithm, training the gas detection model based on the initial model parameters and a sample training set to obtain the trained gas detection model; obtaining gas concentration characteristic values of at least one kind of gas of the dissolved gas in transformer oil detected by a gas concentration characteristic value detection system in real time; and applying the trained gas detection model to the gas concentration characteristic values to monitor the liquid phase gas concentration of the dissolved gas in transformer oil in real time.

[0006] Optionally, the initial model parameter of the gas detection model is obtained by using the firefly algorithm, comprising: initializing a firefly population and determining an initial firefly individual position as a current firefly position, wherein each firefly individual in the firefly population corresponds to a model parameter of the gas detection model; calculating the fitness of all firefly individuals in a loop, and moving the current firefly randomly according to the fitness until a maximum iteration number is reached; determining an optimal position with the maximum fitness, and determining the model parameter corresponding to the firefly at the optimal position as the initial parameter of the gas detection model.

[0007] Optionally, the step of calculating the fitness of the current firefly individual in a loop and moving the current firefly randomly according to the fitness until a maximum iteration number is reached comprises: calculating the fitness of all firefly individuals and sorting them from large to small; randomly selecting a first number of fireflies at the top of the sorting, calculating the firefly position and the attraction; the current firefly moves randomly to the first number of fireflies according to the firefly position and the attraction; if the maximum iteration number is not reached, return to the step of calculating the fitness of all firefly individuals and sorting them from large to small.

[0008] Optionally, the sample training set comprises a plurality of historical gas concentration characteristic values and a plurality of historical liquid-phase gas concentrations corresponding to the historical gas concentration characteristic values respectively, and the model training of the gas detection model based on the initial model parameter and the sample training set comprises: taking the initial model parameter as the initial parameter of the gas detection model; applying the gas detection model to process the input historical gas concentration characteristic values to obtain a predicted liquid-phase gas concentration; iteratively calculating the error between the historical liquid-phase gas concentration corresponding to the input historical gas concentration characteristic value and the preset liquid-phase gas concentration, and iteratively adjusting the model parameter of the gas detection model according to the error until a model convergence condition is met, wherein the model convergence condition is that the error is within a preset range or a preset iteration number is reached.

[0009] Optionally, the real-time monitoring of the liquid-phase gas concentration of the dissolved gas in the transformer oil based on the trained gas detection model according to the gas concentration characteristic value comprises: receiving at least one gas concentration characteristic value of a gas transmitted in real time by the gas concentration characteristic value detection system; applying the trained gas detection model to process according to the gas concentration characteristic value to obtain the current liquid-phase gas concentration of the dissolved gas in the transformer oil output by the gas detection model.

[0010] Optionally, the gas detection model is a wavelet neural network, comprising: an input layer, a hidden layer and an output layer, the number of nodes of the input layer and the output layer is the number of gas types, the hidden layer adopts a Gaussian wavelet function, and the number of neurons of the hidden layer is determined according to the following relationship: k1 is a constant between [1, 10], and m is the number of gas types.

[0011] Based on the same inventive concept, the present application further provides a concentration monitoring device for dissolved gases in transformer oil, comprising: a model training unit, configured to obtain initial model parameters of a gas detection model by using a glowworm algorithm, to perform model training on the gas detection model based on the initial model parameters and a sample training set, and to obtain the trained gas detection model; a characteristic value acquisition unit, configured to acquire a gas concentration characteristic value of at least one gas of dissolved gases in transformer oil detected by a gas concentration characteristic value detection system in real time; and a concentration monitoring unit, configured to apply the trained gas detection model to real-time monitoring of liquid-phase gas concentration of dissolved gases in transformer oil according to the gas concentration characteristic value.

[0012] Based on the same inventive concept, the present application further provides a concentration monitoring system for dissolved gases in transformer oil, comprising: a gas concentration characteristic value detection system and the concentration monitoring device for dissolved gases in transformer oil as described above.

[0013] Based on the same inventive concept, the present application further provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the concentration monitoring method for dissolved gases in transformer oil as described above when executing the program.

[0014] Based on the same inventive concept, the present application further provides a computer storage medium, wherein at least one executable instruction is stored in the storage medium, and the executable instruction causes a processor to execute the concentration monitoring method for dissolved gases in transformer oil as described above.

[0015] From the above, it can be seen that the beneficial effects of the technical solutions provided by the present application are: the present application provides a concentration monitoring method, device and system for dissolved gas in transformer oil, the concentration monitoring method comprising: obtaining initial model parameters of a gas detection model by using a glowworm algorithm, training the gas detection model based on the initial model parameters and a sample training set to obtain the trained gas detection model; obtaining gas concentration characteristic values of at least one kind of gas in the dissolved gas in transformer oil detected by a gas concentration characteristic value detection system in real time; and applying the trained gas detection model to the gas concentration characteristic values to monitor the liquid-phase gas concentration of the dissolved gas in transformer oil in real time, the present application uses the glowworm algorithm to make the parameters of the gas detection model easy to identify, avoids the cross interference of the dissolved gas in transformer oil, and reduces the calculation complexity of the dissolved gas concentration in transformer oil. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced below. Obviously, the drawings in the following description are only embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0017] Figure 1 A flowchart of the concentration monitoring method for dissolved gas in transformer oil in the embodiments of the present application;

[0018] Figure 2 A flowchart of obtaining initial model parameters of a gas detection model by using a glowworm algorithm in the embodiments of the present application;

[0019] Figure 3 A structure diagram of a gas detection model in the embodiments of the present application;

[0020] Figure 4 A structure diagram of a concentration monitoring device for dissolved gas in transformer oil in the embodiments of the present application;

[0021] Figure 5 A hardware structure diagram of an electronic device in the embodiments of the present application. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solutions and advantages of the present disclosure more clear, the present disclosure will be further described in detail below in combination with specific embodiments and with reference to the drawings.

[0023] It should be noted that the technical terms or scientific terms used in the embodiments of the present application should be the general meanings understood by those skilled in the art to which the present disclosure belongs, unless otherwise defined. The terms "first", "second", and similar words used in the embodiments of the present application do not represent any order, number or importance, but are only used to distinguish different components. The terms "include" or "contain" and similar words mean that the elements or objects before the words cover the elements or objects listed after the words and their equivalents, and do not exclude other elements or objects. The terms "connected" or "connected" and similar words are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to represent relative positional relationships, and when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0024] The embodiments of the present application implement a concentration monitoring method for dissolved gases in transformer oil, as shown in Figure 1 The method comprises the following steps:

[0025] Step S1: obtaining initial model parameters of the gas detection model by using the glowworm algorithm, training the gas detection model based on the initial model parameters and the sample training set, and obtaining the trained gas detection model.

[0026] The initial model parameters of the gas detection model are obtained by using the glowworm algorithm parameters, the gas concentration characteristic values are detected by using the gas concentration characteristic value detection system, the dissolved gases in the transformer oil are detected to obtain a plurality of gas concentration characteristic values, and the liquid phase gas concentration corresponding to each gas concentration characteristic value is obtained. A plurality of measurement results of dissolved gases in transformer oil are obtained to form a sample training set, and then the gas detection model is trained according to the sample training set to obtain a trained gas detection model.

[0027] Step S2: Real-time acquisition of the gas concentration characteristic value of at least one gas of the dissolved gas in the transformer oil detected by the gas concentration characteristic value detection system.

[0028] The dissolved gas in the transformer oil is a mixed gas, including but not limited to carbon monoxide, carbon dioxide, methane, ethane, ethylene, and acetylene. The gas concentration characteristic value of at least one gas in the mixed gas is obtained by using the gas concentration characteristic value detection system.

[0029] Step S3: Real-time monitoring of the liquid phase gas concentration of the dissolved gas in the transformer oil according to the gas concentration characteristic value by using the trained gas detection model.

[0030] Based on the trained gas detection model, the liquid phase gas concentration of the dissolved gas in the transformer oil is monitored to obtain a real-time liquid phase gas concentration value.

[0031] This invention employs the firefly algorithm to make the parameters of the gas detection model easy to identify, avoids cross-interference of dissolved gases in transformer oil, and reduces the computational complexity of dissolved gas concentration in transformer oil.

[0032] In step S1 of this embodiment of the invention, when using the firefly algorithm to obtain the initial model parameters of the gas detection model, such as... Figure 2 As shown, it includes:

[0033] Step S21: Initialize the firefly population and determine the initial location of individual fireflies, and determine the current location of fireflies. Each individual firefly in the firefly population corresponds to a model parameter of the gas detection model.

[0034] To initialize the firefly population, it is preferable to set the population size to N, determine the initial positions of individual fireflies in the population, and determine the maximum number of iterations (Max). Calculate and sort the fitness of the initial firefly individuals, and use the position with the highest fitness as the current firefly position.

[0035] Step S22: Iteratively calculate the fitness of all individual fireflies, and randomly move the current firefly according to its fitness until the maximum number of iterations is reached.

[0036] In this embodiment of the invention, firstly, the fitness of all individual fireflies is calculated and sorted from highest to lowest; secondly, the first r fireflies in the sorted order are randomly selected, and their positions and attractiveness are calculated. The firefly positions can be determined using the formula... The calculation is performed, where α is the step size factor, taking values ​​[0,1] and following a Gaussian distribution. Let k be the position of the firefly at time k, which is the global optimal solution. The random number obtained at time k+1, the attraction between fireflies can be calculated using the following formula. Taking firefly e and firefly f as an example, the attraction between firefly e and firefly f is: Where β0 represents the attractive force at the light source, γ ef Let λ represent the distance between firefly e and firefly f, and λ represent the light absorption rate of the medium, usually set to 1. Then, the current firefly randomly moves towards the first set of fireflies based on its position and attractiveness. If the maximum number of iterations (Max) has not been reached, the algorithm returns to the step of calculating the fitness of all individual fireflies and sorting them in descending order. By reducing the number of firefly movements and incorporating the idea of ​​global optima in particle swarm optimization, the algorithm's computational complexity is reduced while avoiding getting trapped in local optima. If the preset maximum number of iterations is reached, the firefly stops moving.

[0037] Step S23: Determine the optimal position with the highest fitness, and use the model parameters corresponding to the firefly at the optimal position as the initial parameters of the gas detection model.

[0038] During the firefly's movement, the fitness of individual fireflies is ranked, and the optimal position is found based on the maximum fitness. The model parameters corresponding to the firefly at the optimal position are then used as the initial parameters for the gas detection model.

[0039] In this embodiment of the invention, the sample training set includes multiple historical gas concentration feature values ​​and multiple historical liquid phase gas concentrations corresponding to the historical gas concentration feature values. A sample training set T = {(X1,Y1),K,(X...} can be created based on multiple sets of historical gas concentration feature values ​​and historical liquid phase gas concentrations. n ,Y n )}, where X i =(x1,x2,K,x) m ) T Y is the i-th gas concentration characteristic value. i =(y1,y2,K,y m ) T Let i be the corresponding liquid phase gas concentration value in the transformer oil, i = 1, 2, ..., n, where n is the total number of samples and m is the number of dissolved gas types in the transformer oil.

[0040] In this embodiment of the invention, the gas detection model is preferably a wavelet neural network, such as... Figure 3 As shown, the wavelet neural network consists of an input layer, hidden layers, and an output layer. The number of nodes in the input and output layers is equal to the number of gas types. The hidden layers use Gaussian wavelet functions, which can be expressed by the formula: Where a is the translation parameter and b is the scaling parameter, the connection weights and thresholds from the input layer to the hidden layer are replaced by the scaling and translation parameters of the Gaussian wavelet function, respectively. The number of neurons in the hidden layer is determined according to the following relationship: Where k1 is a constant between [1, 10], and m is the number of gas species.

[0041] In step S1, when training the gas detection model based on the initial model parameters and the sample training set, the initial model parameters are used as the initial parameters of the gas detection model. The gas detection model is then applied to process the input historical gas concentration feature values ​​to obtain the predicted liquid phase gas concentration. The error between the historical liquid phase gas concentration corresponding to the input historical gas concentration feature values ​​and the preset liquid phase gas concentration is iteratively calculated, and the model parameters of the gas detection model are adjusted iteratively based on the error until the model convergence condition is met. The model convergence condition is that the error is within a preset range or reaches a preset number of iterations. The error function is... y i This is the output gas concentration value. This represents the true concentration of the i-th gas in the gas mixture.

[0042] In the embodiment of the present application, in step S3, the gas concentration characteristic value of at least one kind of gas transmitted by the gas concentration characteristic value detection system in real time is first received; and then the trained gas detection model is applied for processing according to the gas concentration characteristic value, so as to obtain the current liquid-phase gas concentration of the dissolved gas in the transformer oil output by the gas detection model.

[0043] The gas concentration characteristic value detection system comprises, in sequence, an oil-gas separation unit, a filtering unit, a gas detection unit, and a data processing unit connected with the gas detection unit. The oil-gas separation unit is used for the separation of the dissolved gas in the transformer oil, and can be realized by setting a degassing chamber, a liquid level sensor, a temperature sensor, and a waste oil treatment device. The filtering unit is used for filtering the oil and other impurities contained in the separated gas. The gas detection unit is used for detecting the dissolved gas concentration characteristic value in the transformer oil. The gas detection unit can be realized by setting a signal generator, a near-infrared distributed feedback laser, a laser controller, an optical switch, a collimator, a gas chamber, a photodetector, and a preamplifier. The signal generator can generate a low-frequency sawtooth wave scanning signal and a high-frequency sinusoidal modulation signal, the near-infrared distributed feedback laser can emit laser to irradiate the dissolved gas in the transformer oil to obtain outgoing light, the laser controller can tune the optical signal, the optical switch can physically switch or logically operate the optical signal in the optical transmission line or integrated optical path, the collimator can maximize the efficiency of light signal reception, the gas chamber can be used to store the dissolved gas in the transformer oil, the light signal output by the collimator is received by the photodetector after passing through the gas chamber, the photodetector can convert the light signal into an electrical signal, and the preamplifier can preamplify the electrical signal output by the photodetector. The data processing unit is used for gas concentration characteristic value detection, and related data display and storage.

[0044] When the dissolved gas detection system in transformer insulation oil based on TDLAS measures multiple-component gases, there are problems such as gas cross interference of the measured gas concentration characteristic value, and difficulty in identifying the conversion model parameters of the gas-phase concentration and the liquid-phase gas concentration. The concentration monitoring method for the dissolved gas in the transformer oil in the embodiment directly adopts a "gas concentration characteristic value-liquid-phase gas concentration in the transformer oil" calibration model, and uses a wavelet neural network for parameter identification of the above model, specifically, the wavelet neural network is used to process the detected gas concentration characteristic value to obtain the liquid-phase gas concentration in the transformer oil. Considering that the performance of the wavelet neural network is closely related to the network parameters, at present, the gradient descent method is mostly used to optimize the network parameters, the change direction is relatively fixed, which limits the optimization direction of the wavelet neural network parameters, and the wavelet neural network is prone to fall into a local minimum. In view of this, the wavelet neural network is optimized by improving the glowworm algorithm in the embodiment, and the global optimum is avoided by reducing the number of glowworm shifts, so that the algorithm calculation complexity is reduced and the algorithm is prevented from falling into a local optimum.

[0045] The embodiment of the present application provides a concentration monitoring method of dissolved gas in transformer oil, comprising the following steps: obtaining initial model parameters of a gas detection model by using a glowworm algorithm, performing model training on the gas detection model based on the initial model parameters and a sample training set, and obtaining the trained gas detection model; obtaining gas concentration characteristic values of at least one kind of gas in the dissolved gas in transformer oil detected by a gas concentration characteristic value detection system in real time; and monitoring the liquid-phase gas concentration of the dissolved gas in transformer oil in real time by using the trained gas detection model according to the gas concentration characteristic values. The glowworm algorithm is used to make the parameters of the gas detection model easy to identify, avoid the cross interference of the dissolved gas in transformer oil, and reduce the calculation complexity of the dissolved gas concentration in transformer oil.

[0046] The embodiment of the present application further provides a concentration monitoring device of dissolved gas in transformer oil, as shown in the figure, comprising a model training unit 401, a characteristic value obtaining unit 402 and a concentration monitoring unit 403. Figure 4

[0047] The model training unit 401 is used for obtaining initial model parameters of a gas detection model by using a glowworm algorithm, performing model training on the gas detection model based on the initial model parameters and a sample training set, and obtaining the trained gas detection model.

[0048] The characteristic value obtaining unit 402 is used for obtaining gas concentration characteristic values of at least one kind of gas in the dissolved gas in transformer oil detected by a gas concentration characteristic value detection system in real time.

[0049] The concentration monitoring unit 403 is used for monitoring the liquid-phase gas concentration of the dissolved gas in transformer oil in real time by using the trained gas detection model according to the gas concentration characteristic values.

[0050] For the convenience of description, the concentration monitoring device of the dissolved gas in transformer oil is described by dividing various units according to functions. Of course, the functions of the units can be realized in the same or multiple software and / or hardware when the embodiment of the present application is implemented.

[0051] The embodiment of the present application further provides a concentration monitoring system of dissolved gas in transformer oil, comprising a gas concentration characteristic value detection system and the aforementioned concentration monitoring device of dissolved gas in transformer oil. The gas concentration characteristic value detection system obtains gas concentration characteristic values of the dissolved gas in transformer oil, and the concentration monitoring device can obtain the liquid-phase gas concentration of the dissolved gas in transformer oil in real time by using the gas detection model according to the obtained gas concentration characteristic values.

[0052] ​The embodiment of the present application further provides an electronic device, which comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the concentration monitoring method of dissolved gas in transformer oil when executing the program.

[0053] Figure 5 A more specific hardware structure of an electronic device provided by the embodiment is shown, which can comprise a processor 501, a memory 502, an input / output interface 503, a communication interface 504 and a bus 505. The processor 501, the memory 502, the input / output interface 503 and the communication interface 504 are connected to each other through the bus 505.

[0054] The processor 501 can be implemented by a general CPU (Central Processing Unit), a microprocessor, an ASIC (Application Specific Integrated Circuit) or one or more integrated circuits, and is used to execute related programs to implement the technical solutions provided by the embodiment of the present application.

[0055] The memory 502 can be implemented by a ROM (Read Only Memory), a RAM (Random Access Memory), a static storage device, a dynamic storage device, etc. The memory 502 can store an operating system and other application programs, and when the technical solutions provided by the embodiment of the present application are implemented by software or firmware, the related program codes are stored in the memory 502 and are called and executed by the processor 501.

[0056] The input / output interface 503 is used to connect input / output modules to realize information input and output. The input / output modules can be configured as components in the device (not shown in the figure) or can be externally connected to the device to provide corresponding functions. The input device can include a keyboard, a mouse, a touch screen, a microphone, various sensors, etc., and the output device can include a display, a speaker, a vibrator, an indicator light, etc.

[0057] The communication interface 504 is used to connect a communication module (not shown in the figure) to realize the communication interaction between the device and other devices. The communication module can realize communication through a wired manner (such as USB, network cable, etc.) or through a wireless manner (such as mobile network, WIFI, Bluetooth, etc.).

[0058] The bus 505 includes a path for transferring information between the various components (e.g., the processor 501, the memory 502, the input / output interface 503, and the communication interface 504) of the device.

[0059] It should be noted that although the above device only shows the processor 501, the memory 502, the input / output interface 503, the communication interface 504 and the bus 505, in the specific implementation process, the device can also include other components necessary for normal operation. In addition, those skilled in the art can understand that the above device can also only contain the components necessary to implement the embodiments of the present application, and does not have to contain all the components shown in the figure.

[0060] The embodiment of the present application also provides a computer storage medium, and the storage medium stores at least one executable instruction, and the executable instruction enables the processor to execute the concentration monitoring method of the dissolved gas in the transformer oil.

[0061] The above describes specific embodiments of the present application, and other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be performed in a different order than the order in which they are recited and still achieve the desired results. In addition, the processes depicted in the figures do not necessarily require the particular order shown, or sequential order, to achieve the desired results. In certain implementations, multitasking and parallel processing can be advantageous.

[0062] Those skilled in the art should understand that the above discussion of any embodiment is only exemplary and is not intended to limit the scope of the disclosure (including claims) to these examples; under the idea of the present disclosure, the above embodiments or technical features in different embodiments can also be combined, the steps can be implemented in any order, and there are many other changes of different aspects of the embodiments of the present application as described above. For the sake of brevity, they are not provided in detail.

[0063] The embodiments of the present application are intended to cover all such alternatives, modifications and variations as fall within the broad scope of the appended claims. Accordingly, any and all such modifications, variations, omissions, and equivalents should be included within the scope of the present disclosure.

Claims

1. A method of monitoring the concentration of dissolved gases in transformer oil, characterized by, The concentration monitoring method includes: The initial model parameters of the gas detection model are obtained using the firefly algorithm. The gas detection model is then trained based on the initial model parameters and the sample training set to obtain the trained gas detection model. The gas concentration characteristic value of at least one gas in the transformer oil is acquired in real time by a gas concentration characteristic value detection system. The liquid phase gas concentration of dissolved gases in transformer oil is monitored in real time using the trained gas detection model based on the gas concentration characteristic values. The initial model parameters for obtaining the gas detection model using the firefly algorithm include: Initialize the firefly population and determine the initial location of individual fireflies, and determine the current location of fireflies, wherein each individual firefly in the firefly population corresponds to a model parameter of the gas detection model; The fitness of all individual fireflies is calculated iteratively, and the current firefly is randomly moved according to its fitness until the maximum number of iterations is reached. The optimal position with the highest fitness is determined, and the model parameters corresponding to the firefly at the optimal position are determined as the initial parameters of the gas detection model. The process of iteratively calculating the fitness of the current firefly individual and randomly moving the current firefly based on its fitness until the maximum number of iterations is reached includes: Calculate the fitness of all firefly individuals and sort them from largest to smallest; Randomly select the first number of fireflies in the sorted list and calculate their position and attractiveness. The current firefly moves randomly to the first number of fireflies based on the firefly's location and the attraction force. If the maximum number of iterations has not been reached, return to the step of calculating the fitness of all firefly individuals and sorting them from largest to smallest.

2. The concentration monitoring method of claim 1, wherein The sample training set includes multiple historical gas concentration feature values ​​and multiple historical liquid phase gas concentrations corresponding to the historical gas concentration feature values. The step of training the gas detection model based on the initial model parameters and the sample training set includes: The initial model parameters are used as the initial parameters of the gas detection model; The gas detection model is applied to process the input historical gas concentration feature values ​​to obtain the predicted liquid phase gas concentration. The error between the historical liquid gas concentration and the predicted liquid gas concentration corresponding to the input historical gas concentration feature value is calculated iteratively, and the model parameters of the gas detection model are adjusted iteratively according to the error until the model convergence condition is met, wherein the model convergence condition is that the error is within a preset range or reaches a preset number of iterations.

3. The method of claim 1, wherein the concentration of the analyte is determined by measuring the intensity of the light emitted by the analyte. The step of applying the trained gas detection model based on the gas concentration characteristic value to perform real-time monitoring of the liquid phase gas concentration of dissolved gases in transformer oil includes: Receive the gas concentration characteristic value of at least one gas transmitted in real time by the gas concentration characteristic value detection system; The gas concentration feature value is processed using the trained gas detection model to obtain the current liquid phase gas concentration of dissolved gas in transformer oil output by the gas detection model.

4. The concentration monitoring method according to any one of claims 1-3, characterized in that, The gas detection model is a wavelet neural network, comprising an input layer, a hidden layer and an output layer, the number of nodes of the input layer and the output layer is the number of gas types, the hidden layer adopts a Gaussian wavelet function, and the number of neurons of the hidden layer is determined according to the following relationship: , is a constant between [1, 10], and m is the number of gas types.

5. A device for monitoring the concentration of dissolved gases in transformer oil, characterized in that, The device includes: The model training unit is configured to obtain initial model parameters of the gas detection model by using the glowworm swarm optimization algorithm, and train the gas detection model based on the initial model parameters and a sample training set to obtain the trained gas detection model. The feature value acquisition unit is configured to acquire, in real time, a gas concentration feature value of at least one gas in the dissolved gas in transformer oil detected by the gas concentration feature value detection system. The concentration monitoring unit is configured to apply the trained gas detection model to the gas concentration feature value to monitor, in real time, a liquid-phase gas concentration of the dissolved gas in transformer oil. The initial model parameters of the gas detection model are obtained by using the glowworm swarm optimization algorithm, including: initializing a glowworm swarm, determining initial positions of glowworm individuals, and determining a current glowworm position, wherein each glowworm individual in the glowworm swarm corresponds to a model parameter of the gas detection model; recursively calculating fitness of all glowworm individuals, and randomly moving the current glowworm according to the fitness until a maximum iteration number is reached; determining an optimal position with the maximum fitness, and determining a model parameter corresponding to the glowworm at the optimal position as the initial parameter of the gas detection model; the recursive calculation of the fitness of the current glowworm individual and the random movement of the current glowworm according to the fitness until the maximum iteration number is reached, including: calculating the fitness of all glowworm individuals and sorting them in descending order; randomly selecting a first number of glowworms with high ranking, calculating a glowworm position and an attractive force; the current glowworm randomly moves to the first number of glowworms according to the glowworm position and the attractive force; if the maximum iteration number is not reached, returning to the step of calculating the fitness of all glowworm individuals and sorting them in descending order.

6. A system for monitoring the concentration of dissolved gases in transformer oil, characterized by The system comprises a gas concentration feature value detection system and the concentration monitoring device of the dissolved gas in transformer oil according to claim 5.

7. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor implements the concentration monitoring method according to any one of claims 1-4 when executing the program.

8. A computer storage medium, characterized by, The storage medium stores at least one executable instruction, and the executable instruction causes the processor to execute the concentration monitoring method according to any one of claims 1-4.

Images