Electrical equipment operation monitoring method based on artificial intelligence

Abstract

The invention discloses an electrical equipment operation monitoring method based on artificial intelligence, and belongs to the technical field of electrical equipment operation monitoring. An electrical equipment operation monitoring model is constructed by adopting an artificial intelligence model, and the electrical equipment operation monitoring model is trained by taking a training data set as support and adopting an intelligent optimization algorithm; according to the method, the electrical equipment operation monitoring model is trained to obtain the trained electrical equipment operation monitoring model, and then the trained electrical equipment operation monitoring model can be used for identifying the real-time operation state parameters, so that the fault state of the electrical equipment can be quickly pre-warned, the fault discovery efficiency can be effectively improved, and the problem that manual inspection wastes time and labor is avoided.

Description

Technical Field

[0001] The present invention belongs to the technical field of electrical equipment operation monitoring, and in particular relates to an electrical equipment operation monitoring method based on artificial intelligence. Background Art

[0002] Electrical equipment refers to various types of equipment used for power generation, transmission, transformation, distribution and consumption. They are an important part of the power system. With the rapid development of my country's power industry, the number and types of electrical equipment are increasing, and the safety and reliability of equipment operation have become the focus of attention of enterprises. Traditional electrical equipment monitoring methods mainly rely on manual inspections and fixed parameter threshold warnings, which have the following shortcomings: manual inspections are inefficient and labor-intensive; fixed parameter threshold warning methods are difficult to adapt to complex and changing operating environments; and it is impossible to achieve early detection and warning of potential equipment failures. Summary of the invention

[0003] The present invention provides an electrical equipment operation monitoring method based on artificial intelligence, which is used to solve the problems of low efficiency of manual inspection and untimely fault discovery in the prior art.

[0004] An electrical equipment operation monitoring method based on artificial intelligence, comprising:

[0005] Collecting a first operating state parameter of the electrical equipment when it is operating normally and a second operating state parameter of the electrical equipment when it is operating in a faulty state, and constructing a training data set according to the first operating state parameter and the second operating state parameter;

[0006] An artificial intelligence model is used to build an electrical equipment operation monitoring model, and with the training data set as support, an intelligent optimization algorithm is used to train the electrical equipment operation monitoring model to obtain the trained electrical equipment operation monitoring model;

[0007] The real-time operating status parameters of the electrical equipment during operation are collected, and the real-time operating status parameters are identified through the trained electrical equipment operation monitoring model to obtain the electrical equipment operation monitoring results.

[0008] In a possible implementation, after the real-time operating status parameters are identified by the trained electrical equipment operation monitoring model, it also includes: when the electrical equipment operation monitoring results meet the warning conditions, text warning information and sound and light warning information are generated, and a warning is issued based on the text warning information and the sound and light warning information.

[0009] In a possible implementation manner, the first operating state parameter, the second operating state parameter, and the real-time operating state parameter have the same parameter structure and at least include an operating current and an operating voltage.

[0010] In a possible implementation, an artificial intelligence model is used to construct an electrical equipment operation monitoring model, and the electrical equipment operation monitoring model is trained using an intelligent optimization algorithm with the training data set as support, to obtain a trained electrical equipment operation monitoring model, including:

[0011] An artificial intelligence model is used to construct an electrical equipment operation monitoring model, and hyperparameters of the electrical equipment operation monitoring model are initialized, the hyperparameters are encoded into vectors, parameter individuals are obtained, and multiple different parameter individuals are obtained;

[0012] With the training data set as support, obtain the fitness value corresponding to each parameter individual, and obtain the optimal individual according to the fitness value corresponding to each parameter individual;

[0013] For each parameter individual, according to the optimal individual, the adaptive domain exploration strategy is used to perform neighborhood exploration on the parameter individual to obtain the parameter individual after neighborhood exploration;

[0014] For each parameter individual after neighborhood exploration, an adaptive diversity collaborative strategy is used to perform collaborative search on the parameter individuals to obtain the parameter individuals after collaborative search;

[0015] For each parameter individual after collaborative search, a probability acceptance global mutation strategy is used to perform a global search on the parameter individual to obtain the parameter individual after global search;

[0016] Determine whether the training end conditions are met. If so, determine the final parameters of the electrical equipment operation monitoring model based on the parameter individuals after the global search to obtain the trained electrical equipment operation monitoring model. Otherwise, return to the step of obtaining the fitness value.

[0017] In a possible implementation, an artificial intelligence model is used to construct an electrical equipment operation monitoring model, and hyperparameters of the electrical equipment operation monitoring model are initialized, the hyperparameters are encoded into vectors, parameter individuals are obtained, and multiple different parameter individuals are obtained, including:

[0018] A convolutional neural network is used to build an electrical equipment operation monitoring model, and the hyperparameters of the electrical equipment operation monitoring model are initialized. The hyperparameters are encoded into vectors to obtain parameter individuals.

[0019] Based on the obtained parameter individuals, multiple different parameter individuals are obtained: X n+1,d =X n,d *μ*(1-X n,d ), where X n,d represents the d-th dimension hyperparameter of the n-th parameter individual, X n+1,drepresents the d-th dimension hyperparameter of the n+1-th parameter individual, d=1,2,…,D, D represents the total dimension of the hyperparameter, and μ represents the chaotic mapping factor between (0,4).

[0020] In a possible implementation, the training data set is used as support to obtain the fitness value corresponding to each parameter individual, and the optimal individual is obtained according to the fitness value corresponding to each parameter individual, including:

[0021] Based on the training data set, the root mean square error function value corresponding to each parameter individual is obtained, and according to the root mean square error function value corresponding to the parameter individual, the fitness value corresponding to the parameter individual is determined as: 1 / (root mean square error function value + 0.001);

[0022] According to the fitness value corresponding to each parameter individual, the parameter individual with the largest fitness value is determined as the optimal individual.

[0023] In a possible implementation, for each parameter individual, according to the optimal individual, an adaptive domain exploration strategy is used to perform neighborhood exploration on the parameter individual to obtain the parameter individual after the neighborhood exploration, including:

[0024] For each parameter individual, a random exploration vector γ is generated, and the parameters of each dimension in the random exploration vector γ are normalized to obtain the normalized random exploration vector γ; wherein the parameter dimension of the random exploration vector γ is the same as the parameter dimension of the parameter individual, and each dimension is a random number between (0,1);

[0025] According to the normalized random exploration vector γ, the first neighborhood exploration individual and the second neighborhood exploration individual are determined as: as well as in, represents the mth parameter individual in the tth training process, m = 1, 2, ..., M, M represents the total number of parameter individuals, Represents parameter individuals The corresponding first neighborhood exploration individual, η0 represents the neighborhood exploration range factor, Represents parameter individuals The corresponding second neighborhood exploration individual;

[0026] Determine the parameter individual based on the first neighborhood exploration individual and the second neighborhood exploration individual The corresponding neighborhood information search items are: Among them, Step represents the search step length, sign represents the sign function, Represents parameter individuals The corresponding neighborhood information search item, represents the fitness value corresponding to the first neighborhood exploration individual, Indicates the fitness value corresponding to the second neighborhood exploration individual;

[0027] Based on the current number of training times, the adaptive inertia weight is obtained as: Among them, w t represents the adaptive inertia weight during the t-th training process, w max represents the maximum value of the adaptive inertia weight, w min represents the minimum value of the adaptive inertia weight, and T represents the preset maximum number of training times;

[0028] According to the adaptive inertia weight and the optimal individual, obtain the parameter individual The corresponding optimal direction search speed is: in, Indicates the parameter individual in the tth training process The corresponding optimal direction search speed, Indicates the parameter individual in the t+1th training process The corresponding optimal direction search speed, e represents a natural constant, R represents a search speed control factor between (0,1), and r1 represents a random number between (0,1). represents the optimal individual;

[0029] According to the parameters of individual The corresponding optimal direction search speed and domain information search item, for parameter individuals Perform neighborhood exploration and obtain the parameter individuals after neighborhood exploration for: in, Represents the parameter individual after neighborhood exploration λ represents the information search adjustment weight between (0,1).

[0030] In a possible implementation, for each parameter individual after neighborhood exploration, an adaptive diversity collaborative strategy is used to perform collaborative search on the parameter individual to obtain the parameter individual after collaborative search, including:

[0031] According to the parameter individuals after all neighborhood explorations, the parameter individual center is obtained as: in, represents the individual center of the parameters in the t-th training process, represents the parameter individuals after the kth neighborhood exploration in the tth training process, k = 1, 2, ..., M, M represents the total number of parameter individuals, Represents parameter individuals The corresponding fitness value;

[0032] For each parameter individual after neighborhood exploration, a collaborative search is performed on the parameter individual according to the center of the parameter individual, and the parameter individual after the collaborative search is obtained as follows: in, represents the parameter individuals after the i-th neighborhood exploration in the t-th training process, i = 1, 2, ..., M, M represents the total number of parameter individuals, Represents the parameter individual after collaborative search r2 represents a random number between (0,1), represents the optimal individual, and α represents the cooperation factor between (-1,1).

[0033] In a possible implementation, for each parameter individual after neighborhood exploration, an adaptive diversity collaborative strategy is used to perform collaborative search on the parameter individual to obtain the parameter individual after collaborative search, including:

[0034] For each parameter individual after neighborhood exploration, the parameter individual is mutated as follows: in, represents the parameter individual after the jth neighborhood exploration during the tth training process, Represents parameter individuals The corresponding variant individuals, Represents a random parameter individual generated by a Gaussian distribution, and each dimension is r3 represents a random number between (0,1), π represents the ratio of pi, e represents a natural constant, and t represents the current number of training times;

[0035] Determine variant individuals The fitness value of Is it greater than the parameter individual The fitness value of If so, the mutant individual as the parameter individual after collaborative search, otherwise proceed to the step of obtaining the acceptance probability of inferior solutions;

[0036] The acceptance probability of obtaining an inferior solution is: Where K is the Boltzmann constant, τ t+1 represents the probability control factor in the t+1th training process, and τ t+1 =μτ t ; μ represents the attenuation factor between (0,1), τ t represents the probability control factor in the t-th training process;

[0037] Based on the acceptance probability of inferior solutions, the roulette algorithm is used to select mutant individuals Or the original parameter individual As parameter individuals after collaborative search.

[0038] In a possible implementation, determining whether the training end condition is met includes: determining whether the current training times are greater than or equal to a preset maximum training times; if so, determining that the training end condition is met; otherwise, determining that the training end condition is not met.

[0039] The present invention provides an electrical equipment operation monitoring method based on artificial intelligence. The electrical equipment operation monitoring model is constructed by adopting an artificial intelligence model, and supported by a training data set, an intelligent optimization algorithm is used to train the electrical equipment operation monitoring model to obtain the trained electrical equipment operation monitoring model. Then, the trained electrical equipment operation monitoring model can be used to identify real-time operation status parameters, so that the fault state of the electrical equipment can be quickly warned, which can effectively improve the efficiency of fault detection and avoid the time-consuming and labor-intensive manual inspection problem. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0041] Figure 1 A flowchart of an electrical equipment operation monitoring method based on artificial intelligence is provided in an embodiment of the present invention.

[0042] The above drawings have shown clear embodiments of the present invention, which will be described in more detail below. These drawings and text descriptions are not intended to limit the scope of the present invention in any way, but to illustrate the concept of the present invention for those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0043] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Instead, they are merely examples of devices and methods consistent with some aspects of the present invention as detailed in the appended claims.

[0044] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0045] like Figure 1 As shown, the present invention provides an electrical equipment operation monitoring method based on artificial intelligence, comprising:

[0046] S1, collecting a first operating state parameter of the electrical equipment when it is operating normally and a second operating state parameter of the electrical equipment when it is operating in a faulty manner, and constructing a training data set according to the first operating state parameter and the second operating state parameter;

[0047] For example, based on the time node of the fault occurrence and the preset data sampling frequency, the operating parameters at the previous N consecutive time points can be collected to obtain the second operating state parameters, and the first operating state parameters can be collected when there is no fault, so that fault prediction can be achieved. However, it is worth noting that the operating parameters at N / 2 consecutive time points can be collected before and after the fault, and rapid fault detection can also be achieved, so that the staff can quickly find the abnormality and handle it.

[0048] S2. Use an artificial intelligence model to build an electrical equipment operation monitoring model, and use an intelligent optimization algorithm to train the electrical equipment operation monitoring model with the support of a training data set to obtain a trained electrical equipment operation monitoring model;

[0049] Since the operating state parameters may include multiple parameters and each parameter has data at multiple time points, it is easier to identify if the input is constructed as a matrix. In the embodiment of the present invention, a convolutional neural network may be preferably used to construct an electrical equipment operation monitoring model. After the electrical equipment operation monitoring model is trained using an intelligent optimization algorithm, the electrical equipment operation monitoring model may be used for fault monitoring.

[0050] S3. Collect real-time operating status parameters of the electrical equipment during operation, and identify the real-time operating status parameters through the trained electrical equipment operation monitoring model to obtain electrical equipment operation monitoring results.

[0051] By identifying the real-time operating status parameters through the trained electrical equipment operation monitoring model, automatic fault identification can be achieved, which not only solves the time-consuming and labor-intensive technical problems caused by manual inspections, but also improves the efficiency of fault discovery.

[0052] In a possible implementation, after the real-time operating status parameters are identified by the trained electrical equipment operation monitoring model, it also includes: when the electrical equipment operation monitoring results meet the warning conditions, text warning information and sound and light warning information are generated, and a warning is issued based on the text warning information and the sound and light warning information.

[0053] For example, text warning information can be transmitted to the staff's mobile device for warning, and sound and light warning information can be transmitted to the sound and light warning device at the electrical equipment work site for warning, so that the staff can quickly deal with the abnormality.

[0054] In a possible implementation manner, the first operating state parameter, the second operating state parameter, and the real-time operating state parameter have the same parameter structure and at least include an operating current and an operating voltage.

[0055] However, it is worth noting that the operating current and operating voltage are only examples of the embodiments of the present invention, and other operating parameters may also be included. For example, when the electrical device is a motor, vibration parameters may also be collected as operating state parameters to identify various faults of the motor.

[0056] The prior art often uses a gradient descent intelligent optimization algorithm to optimize the hyperparameters of an artificial intelligence model, which is prone to falling into a local optimal value, resulting in the electrical equipment operation monitoring model being unable to effectively identify faults or having a low recognition accuracy. Therefore, an embodiment of the present invention provides a new intelligent optimization algorithm to solve the technical problems in the prior art and improve the accuracy of fault identification.

[0057] In a possible implementation, an artificial intelligence model is used to construct an electrical equipment operation monitoring model, and the electrical equipment operation monitoring model is trained using an intelligent optimization algorithm with the training data set as support, to obtain a trained electrical equipment operation monitoring model, including:

[0058] An artificial intelligence model is used to construct an electrical equipment operation monitoring model, and hyperparameters of the electrical equipment operation monitoring model are initialized, the hyperparameters are encoded into vectors, parameter individuals are obtained, and multiple different parameter individuals are obtained;

[0059] Optionally, a random initialization method may be used to randomly initialize the hyperparameters (such as weight parameters) of the electrical equipment operation monitoring model.

[0060] With the training data set as support, obtain the fitness value corresponding to each parameter individual, and obtain the optimal individual according to the fitness value corresponding to each parameter individual;

[0061] For each parameter individual, according to the optimal individual, the adaptive domain exploration strategy is used to perform neighborhood exploration on the parameter individual to obtain the parameter individual after neighborhood exploration;

[0062] For each parameter individual after neighborhood exploration, an adaptive diversity collaborative strategy is used to perform collaborative search on the parameter individuals to obtain the parameter individuals after collaborative search;

[0063] For each parameter individual after collaborative search, a probability acceptance global mutation strategy is used to perform a global search on the parameter individual to obtain the parameter individual after global search;

[0064] Determine whether the training end conditions are met. If so, determine the final parameters of the electrical equipment operation monitoring model based on the parameter individuals after the global search to obtain the trained electrical equipment operation monitoring model. Otherwise, return to the step of obtaining the fitness value.

[0065] Optionally, after neighborhood exploration, collaborative search, and global search, the individual parameters may be processed once for out-of-bounds, and the parameters exceeding the upper and lower limits may be pulled back to the nearest boundary value (e.g., if the upper limit is exceeded, the exceeding parameter is set to the upper limit) or randomly generated within the upper and lower limits.

[0066] In a possible implementation, an artificial intelligence model is used to construct an electrical equipment operation monitoring model, and hyperparameters of the electrical equipment operation monitoring model are initialized, the hyperparameters are encoded into vectors, parameter individuals are obtained, and multiple different parameter individuals are obtained, including:

[0067] A convolutional neural network is used to build an electrical equipment operation monitoring model, and the hyperparameters of the electrical equipment operation monitoring model are initialized. The hyperparameters are encoded into vectors to obtain parameter individuals.

[0068] Based on the obtained parameter individuals, multiple different parameter individuals are obtained: X n+1,d =X n,d *μ*(1-X n,d ), where X n,d represents the d-th dimension hyperparameter of the n-th parameter individual, X n+1,d represents the d-th dimension hyperparameter of the n+1-th parameter individual, d=1,2,…,D, D represents the total dimension of the hyperparameter, and μ represents the chaotic mapping factor between (0,4).

[0069] The initialization method provided by the embodiment of the present invention can make the parameter individuals more evenly distributed in the solution space, and can effectively improve the convergence speed of the algorithm.

[0070] In a possible implementation, the training data set is used as support to obtain the fitness value corresponding to each parameter individual, and the optimal individual is obtained according to the fitness value corresponding to each parameter individual, including:

[0071] With the training data set as support, the root mean square error function value corresponding to each parameter individual is obtained, and according to the root mean square error function value corresponding to the parameter individual (it is worth noting that other error functions can also be used to obtain it), the fitness value corresponding to the parameter individual is determined as: 1 / (root mean square error function value + 0.001);

[0072] According to the fitness value corresponding to each parameter individual, the parameter individual with the largest fitness value is determined as the optimal individual.

[0073] In a possible implementation, for each parameter individual, according to the optimal individual, an adaptive domain exploration strategy is used to perform neighborhood exploration on the parameter individual to obtain the parameter individual after the neighborhood exploration, including:

[0074] For each parameter individual, a random exploration vector γ is generated, and the parameters of each dimension in the random exploration vector γ are normalized to obtain the normalized random exploration vector γ; wherein the parameter dimension of the random exploration vector γ is the same as the parameter dimension of the parameter individual, and each dimension is a random number between (0,1);

[0075] According to the normalized random exploration vector γ, the first neighborhood exploration individual and the second neighborhood exploration individual are determined as: as well as in, represents the mth parameter individual in the tth training process, m = 1, 2, ..., M, M represents the total number of parameter individuals, Represents parameter individuals The corresponding first neighborhood exploration individual, η0 represents the neighborhood exploration range factor, Represents parameter individuals The corresponding second neighborhood exploration individual;

[0076] Determine the parameter individual based on the first neighborhood exploration individual and the second neighborhood exploration individual The corresponding neighborhood information search items are: Among them, Step represents the search step length, sign represents the sign function, Represents parameter individuals The corresponding neighborhood information search item, represents the fitness value corresponding to the first neighborhood exploration individual, Indicates the fitness value corresponding to the second neighborhood exploration individual;

[0077] Based on the current number of training times, the adaptive inertia weight is obtained as: Among them, w t represents the adaptive inertia weight during the t-th training process, w max represents the maximum value of the adaptive inertia weight, w min represents the minimum value of the adaptive inertia weight, and T represents the preset maximum number of training times;

[0078] According to the adaptive inertia weight and the optimal individual, obtain the parameter individual The corresponding optimal direction search speed is: in, Indicates the parameter individual in the tth training process The corresponding optimal direction search speed, Indicates the parameter individual in the t+1th training process The corresponding optimal direction search speed, e represents a natural constant, R represents a search speed control factor between (0,1), and r1 represents a random number between (0,1). represents the optimal individual;

[0079] According to the parameters of individual The corresponding optimal direction search speed and domain information search item, for parameter individuals Perform neighborhood exploration and obtain the parameter individuals after neighborhood exploration for: in, Represents the parameter individual after neighborhood exploration λ represents the information search adjustment weight between (0,1).

[0080] The embodiment of the present invention adopts an adaptive field exploration strategy to perform neighborhood exploration on parameter individuals, which can effectively explore better positions within the neighborhood of parameter individuals and learn the optimal positions at the same time, which can effectively improve the convergence speed and convergence accuracy of the algorithm, and at the same time has a certain probability of jumping out of the local optimal solution.

[0081] In a possible implementation, for each parameter individual after neighborhood exploration, an adaptive diversity collaborative strategy is used to perform collaborative search on the parameter individual to obtain the parameter individual after collaborative search, including:

[0082] According to the parameter individuals after all neighborhood explorations, the parameter individual center is obtained as: in, represents the individual center of the parameters in the t-th training process, represents the parameter individuals after the kth neighborhood exploration in the tth training process, k = 1, 2, ..., M, M represents the total number of parameter individuals, Represents parameter individuals The corresponding fitness value;

[0083] For each parameter individual after neighborhood exploration, a collaborative search is performed on the parameter individual according to the center of the parameter individual, and the parameter individual after the collaborative search is obtained as follows: in, represents the parameter individuals after the i-th neighborhood exploration in the t-th training process, i = 1, 2, ..., M, M represents the total number of parameter individuals, Represents the parameter individual after collaborative search r2 represents a random number between (0,1), represents the optimal individual, and α represents the cooperation factor between (-1,1).

[0084] The embodiment of the present invention adopts an adaptive diversity collaboration strategy to perform collaborative search on parameter individuals. It can perform adaptive search based on the overall search situation, have a faster search speed in the early stage of the algorithm, and gradually convert to fine search when the algorithm gradually converges. It can effectively improve the convergence speed of the algorithm while ensuring the diversity of the algorithm in the early and middle stages.

[0085] In a possible implementation, for each parameter individual after neighborhood exploration, an adaptive diversity collaborative strategy is used to perform collaborative search on the parameter individual to obtain the parameter individual after collaborative search, including:

[0086] For each parameter individual after neighborhood exploration, the parameter individual is mutated as follows: in, represents the parameter individual after the jth neighborhood exploration during the tth training process, Represents parameter individuals The corresponding variant individuals, Represents a random parameter individual generated by a Gaussian distribution, and each dimension is r3 represents a random number between (0,1), π represents the ratio of pi, e represents a natural constant, and t represents the current number of training times;

[0087] Determine variant individuals The fitness value of Is it greater than the parameter individual The fitness value of If so, the mutant individual as the parameter individual after collaborative search, otherwise proceed to the step of obtaining the acceptance probability of inferior solutions;

[0088] The acceptance probability of obtaining an inferior solution is: Where K is the Boltzmann constant, τ t+1 represents the probability control factor in the t+1th training process, and τ t+1 =μτ t ; μ represents the attenuation factor between (0,1), τ t represents the probability control factor in the t-th training process;

[0089] Based on the acceptance probability of inferior solutions, the roulette algorithm is used to select mutant individuals Or the original parameter individual As a parameter individual after collaborative search. For example, a random number can be generated between (0,1). When this random number is less than the acceptance probability of the inferior solution, the mutation individual is selected. As the parameter individual after collaborative search, otherwise select the original parameter individual As parameter individuals after collaborative search.

[0090] The embodiment of the present invention adopts an adaptive diversity collaboration strategy to perform collaborative search on parameter individuals, which can effectively assist the algorithm to jump out of the local optimal solution and have a stronger global search capability in the early and middle stages of the algorithm. In the later stages of the algorithm, the probability of accepting inferior solutions gradually decreases, which can ensure the stability of the algorithm.

[0091] In a possible implementation, determining whether the training end condition is met includes: determining whether the current training times are greater than or equal to a preset maximum training times; if so, determining that the training end condition is met; otherwise, determining that the training end condition is not met.

[0092] The present invention provides an electrical equipment operation monitoring method based on artificial intelligence. The electrical equipment operation monitoring model is constructed by adopting an artificial intelligence model, and supported by a training data set, an intelligent optimization algorithm is used to train the electrical equipment operation monitoring model to obtain the trained electrical equipment operation monitoring model. Then, the trained electrical equipment operation monitoring model can be used to identify real-time operation status parameters, so that the fault state of the electrical equipment can be quickly warned, which can effectively improve the efficiency of fault detection and avoid the time-consuming and labor-intensive manual inspection problem.

[0093] Those skilled in the art will readily appreciate other embodiments of the present invention after considering the specification and practicing the invention disclosed herein. The present invention is intended to cover any variations, uses or adaptations of the present invention that follow the general principles of the present invention and include common knowledge or customary techniques in the art that are not disclosed by the present invention. It should be understood that the present invention is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from the scope thereof. The scope of the present invention is limited only by the appended claims.

Claims

1. An electrical equipment operation monitoring method based on artificial intelligence, characterized in that: include: Collecting a first operating state parameter of the electrical equipment when it is operating normally and a second operating state parameter of the electrical equipment when it is operating in a faulty state, and constructing a training data set according to the first operating state parameter and the second operating state parameter; An artificial intelligence model is used to build an electrical equipment operation monitoring model, and with the training data set as support, an intelligent optimization algorithm is used to train the electrical equipment operation monitoring model to obtain the trained electrical equipment operation monitoring model; The real-time operating status parameters of the electrical equipment during operation are collected, and the real-time operating status parameters are identified through the trained electrical equipment operation monitoring model to obtain the electrical equipment operation monitoring results.

2. The method for monitoring the operation of electrical equipment based on artificial intelligence according to claim 1, characterized in that: After the real-time operation status parameters are identified by the trained electrical equipment operation monitoring model, it also includes: when the electrical equipment operation monitoring results meet the warning conditions, text warning information and sound and light warning information are generated, and a warning is issued based on the text warning information and the sound and light warning information.

3. The method for monitoring electrical equipment operation based on artificial intelligence according to claim 1, characterized in that: The first operating state parameter, the second operating state parameter and the real-time operating state parameter have the same parameter structure and at least include an operating current and an operating voltage.

4. The method for monitoring electrical equipment operation based on artificial intelligence according to claim 1, characterized in that: An artificial intelligence model is used to build an electrical equipment operation monitoring model, and with the training data set as support, an intelligent optimization algorithm is used to train the electrical equipment operation monitoring model to obtain the trained electrical equipment operation monitoring model, including: An artificial intelligence model is used to construct an electrical equipment operation monitoring model, and hyperparameters of the electrical equipment operation monitoring model are initialized, the hyperparameters are encoded into vectors, parameter individuals are obtained, and multiple different parameter individuals are obtained; With the training data set as support, obtain the fitness value corresponding to each parameter individual, and obtain the optimal individual according to the fitness value corresponding to each parameter individual; For each parameter individual, according to the optimal individual, the adaptive domain exploration strategy is used to perform neighborhood exploration on the parameter individual to obtain the parameter individual after neighborhood exploration; For each parameter individual after neighborhood exploration, an adaptive diversity collaborative strategy is used to perform collaborative search on the parameter individuals to obtain the parameter individuals after collaborative search; For each parameter individual after collaborative search, a probability acceptance global mutation strategy is used to perform a global search on the parameter individual to obtain the parameter individual after global search; Determine whether the training end conditions are met. If so, determine the final parameters of the electrical equipment operation monitoring model based on the parameter individuals after the global search to obtain the trained electrical equipment operation monitoring model. Otherwise, return to the step of obtaining the fitness value.

5. The method for monitoring the operation of electrical equipment based on artificial intelligence according to claim 4, characterized in that: An artificial intelligence model is used to build an electrical equipment operation monitoring model, and the hyperparameters of the electrical equipment operation monitoring model are initialized. The hyperparameters are encoded into vectors to obtain parameter individuals, and multiple different parameter individuals are obtained, including: A convolutional neural network is used to build an electrical equipment operation monitoring model, and the hyperparameters of the electrical equipment operation monitoring model are initialized. The hyperparameters are encoded into vectors to obtain parameter individuals. Based on the obtained parameter individuals, multiple different parameter individuals are obtained: X n+1,d =X n,d *μ*(1-X n,d ), where X n,d represents the d-th dimension hyperparameter of the n-th parameter individual, X n+1,d represents the d-th dimension hyperparameter of the n+1-th parameter individual, d=1,2,…,D, D represents the total dimension of the hyperparameter, and μ represents the chaotic mapping factor between (0,4).

6. The method for monitoring the operation of electrical equipment based on artificial intelligence according to claim 5, characterized in that: With the training data set as support, obtain the fitness value corresponding to each parameter individual, and obtain the optimal individual according to the fitness value corresponding to each parameter individual, including: Based on the training data set, the root mean square error function value corresponding to each parameter individual is obtained, and according to the root mean square error function value corresponding to the parameter individual, the fitness value corresponding to the parameter individual is determined as: 1 / (root mean square error function value + 0.001); According to the fitness value corresponding to each parameter individual, the parameter individual with the largest fitness value is determined as the optimal individual.

7. The method for monitoring electrical equipment operation based on artificial intelligence according to claim 6, characterized in that: For each parameter individual, according to the optimal individual, the adaptive domain exploration strategy is used to perform neighborhood exploration on the parameter individual, and the parameter individuals after neighborhood exploration are obtained, including: For each parameter individual, a random exploration vector γ is generated, and the parameters of each dimension in the random exploration vector γ are normalized to obtain the normalized random exploration vector γ; wherein the parameter dimension of the random exploration vector γ is the same as the parameter dimension of the parameter individual, and each dimension is a random number between (0,1); According to the normalized random exploration vector γ, the first neighborhood exploration individual and the second neighborhood exploration individual are determined as: as well as in, represents the mth parameter individual in the tth training process, m = 1, 2, ..., M, M represents the total number of parameter individuals, Represents parameter individual The corresponding first neighborhood exploration individual, η0 represents the neighborhood exploration range factor, Represents parameter individuals The corresponding second neighborhood exploration individual; Determine the parameter individual based on the first neighborhood exploration individual and the second neighborhood exploration individual The corresponding neighborhood information search items are: Among them, Step represents the search step length, sign represents the sign function, Represents parameter individuals The corresponding neighborhood information search item, represents the fitness value corresponding to the first neighborhood exploration individual, Indicates the fitness value corresponding to the second neighborhood exploration individual; Based on the current number of training times, the adaptive inertia weight is obtained as: Among them, w t represents the adaptive inertia weight during the t-th training process, w max represents the maximum value of the adaptive inertia weight, w min represents the minimum value of the adaptive inertia weight, and T represents the preset maximum number of training times; According to the adaptive inertia weight and the optimal individual, obtain the parameter individual The corresponding optimal direction search speed is: in, Indicates the parameter individual in the tth training process The corresponding optimal direction search speed, Indicates the parameter individual in the t+1th training process The corresponding optimal direction search speed, e represents a natural constant, R represents a search speed control factor between (0,1), and r1 represents a random number between (0,1). represents the optimal individual; According to the parameters The corresponding optimal direction search speed and domain information search item, for parameter individuals Perform neighborhood exploration and obtain the parameter individuals after neighborhood exploration for: in, Represents the parameter individual after neighborhood exploration λ represents the information search adjustment weight between (0,1).

8. The method for monitoring electrical equipment operation based on artificial intelligence according to claim 7, characterized in that: For each parameter individual after neighborhood exploration, an adaptive diversity collaborative strategy is used to perform collaborative search on the parameter individuals to obtain the parameter individuals after collaborative search, including: According to the parameter individuals after all neighborhood explorations, the parameter individual center is obtained as: in, represents the individual center of the parameters in the t-th training process, represents the parameter individuals after the kth neighborhood exploration in the tth training process, k = 1, 2, ..., M, M represents the total number of parameter individuals, Represents parameter individuals The corresponding fitness value; For each parameter individual after neighborhood exploration, a collaborative search is performed on the parameter individual according to the center of the parameter individual, and the parameter individual after the collaborative search is obtained as follows: in, represents the parameter individuals after the i-th neighborhood exploration in the t-th training process, i = 1, 2, ..., M, M represents the total number of parameter individuals, Represents the parameter individual after collaborative search r2 represents a random number between (0,1), represents the optimal individual, and α represents the cooperation factor between (-1,1).

9. The method for monitoring the operation of electrical equipment based on artificial intelligence according to claim 8, characterized in that: For each parameter individual after neighborhood exploration, an adaptive diversity collaborative strategy is used to perform collaborative search on the parameter individuals to obtain the parameter individuals after collaborative search, including: For each parameter individual after neighborhood exploration, the parameter individual is mutated as follows: in, represents the parameter individual after the jth neighborhood exploration during the tth training process, Represents parameter individuals The corresponding variant individuals, Represents a random parameter individual generated by a Gaussian distribution, and each dimension is r3 represents a random number between (0,1), π represents the ratio of pi, e represents a natural constant, and t represents the current number of training times; Determine variant individuals The fitness value of Is it greater than the parameter individual The fitness value of If so, the mutant individual as the parameter individual after collaborative search, otherwise enter the step of obtaining the acceptance probability of inferior solutions; The acceptance probability of obtaining an inferior solution is: Where K is the Boltzmann constant, τ t+1 represents the probability control factor in the t+1th training process, and τ t+1 =μτ t ; μ represents the attenuation factor between (0,1), τ t represents the probability control factor in the t-th training process; Based on the acceptance probability of inferior solutions, the roulette algorithm is used to select mutant individuals Or the original parameter individual As parameter individuals after collaborative search.

10. The method for monitoring electrical equipment operation based on artificial intelligence according to claim 9, characterized in that: Determining whether the training end condition is met includes: determining whether the current training times is greater than or equal to a preset maximum training times, if so, determining that the training end condition is met, otherwise determining that the training end condition is not met.

Images