CN117131366B

CN117131366B - Transformer maintenance equipment control method and device, electronic equipment and readable medium

Info

Publication number: CN117131366B
Application number: CN202311394518.3A
Authority: CN
Inventors: 张秀丽; 冯东; 王成章; 刘振圻; 谈卉; 张剑鹏; 焦士杰
Original assignee: State Grid Information and Telecommunication Co Ltd; Beijing Guodiantong Network Technology Co Ltd
Current assignee: State Grid Information and Telecommunication Co Ltd; Beijing Guodiantong Network Technology Co Ltd
Priority date: 2023-10-26
Filing date: 2023-10-26
Publication date: 2024-02-06
Anticipated expiration: 2043-10-26
Also published as: CN117131366A

Abstract

The embodiment of the disclosure discloses a transformer maintenance equipment control method, a device, an electronic equipment and a readable medium. One embodiment of the method comprises the following steps: acquiring audio information; preprocessing the audio information; generating a linear characteristic parameter information set according to the preprocessed audio information set; generating a nonlinear characteristic parameter information set according to the preprocessed audio information set; generating a fusion characteristic parameter information set according to the linear characteristic parameter information set and the nonlinear characteristic parameter information set; generating a key fusion characteristic parameter information set according to the fusion characteristic parameter information set; generating transformer fault prediction information according to the key fusion characteristic parameter information set and the transformer fault prediction model; and controlling the transformer maintenance equipment to perform maintenance operation on the target transformer in response to determining that the transformer has faults according to the transformer fault prediction information. According to the embodiment, the power outage time can be shortened, and the waste of maintenance resources of the transformer is reduced.

Description

Transformer maintenance equipment control method and device, electronic equipment and readable medium

Technical Field

Embodiments of the present disclosure relate to the field of computer technology, and in particular, to a method and apparatus for controlling a transformer maintenance device, an electronic device, and a readable medium.

Background

As an electric power device, a transformer is used, and its operation conditions directly affect various aspects of factory production and people's life. Currently, in the maintenance operation of a transformer, the following methods are generally adopted: the method comprises the steps of extracting characteristics of collected transformer audio information (the extracted characteristics generally have higher recognition rate in an audio environment with high signal to noise ratio), predicting the extracted characteristic information (all characteristic information) through a general transformer fault prediction model (such as a convolutional neural network), and carrying out manual maintenance operation on the transformer according to a prediction result.

However, the inventors found that when the maintenance operation is performed on the transformer in the above manner, there are often the following technical problems:

firstly, in the process of carrying out fault prediction on the transformer, the identification rate of the extracted characteristics is higher only in an audio environment with high signal-to-noise ratio, so that the identification accuracy of audio information in the environment with low signal-to-noise ratio is lower, the accuracy of the fault prediction on the transformer is lower, the error rate of transformer maintenance is higher, and further, the power outage time is longer and the waste of transformer maintenance resources is caused.

Secondly, in the process of carrying out fault prediction on the transformer, fault prediction is carried out on all the characteristic information, so that the characteristic dimension is higher, the complexity of model prediction is higher, the complexity of CPU processing is higher, the processing time is longer, and further the waste of CPU processor resources is caused.

Thirdly, in the process of predicting faults of the transformer, a single-mode (audio information) prediction result is adopted as a prediction result, so that the accuracy of predicting faults of the transformer is low, the error rate of maintaining the transformer is high, and further, the power outage time is long and the transformer maintenance resource is wasted.

The above information disclosed in this background section is only for enhancement of understanding of the background of the inventive concept and, therefore, may contain information that does not form the prior art that is already known to those of ordinary skill in the art in this country.

Disclosure of Invention

The disclosure is in part intended to introduce concepts in a simplified form that are further described below in the detailed description. The disclosure is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter.

Some embodiments of the present disclosure propose a transformer maintenance device control method, apparatus, electronic device, and computer readable medium to solve one or more of the technical problems mentioned in the background section above.

In a first aspect, some embodiments of the present disclosure provide a method of controlling a transformer maintenance device, the method comprising: acquiring audio information of a target transformer; preprocessing the audio information to generate a preprocessed audio information set; generating a linear characteristic parameter information set according to the preprocessed audio information set; generating a nonlinear characteristic parameter information set according to the preprocessed audio information set; generating a fusion characteristic parameter information set according to the linear characteristic parameter information set and the nonlinear characteristic parameter information set; generating a key fusion characteristic parameter information set according to the fusion characteristic parameter information set; generating transformer fault prediction information according to the key fusion characteristic parameter information set and a pre-trained transformer fault prediction model, wherein the transformer fault prediction information comprises: fault type and probability of failure; and controlling associated transformer maintenance equipment to perform maintenance operation on the target transformer in response to determining that the transformer is faulty according to the transformer fault prediction information.

In a second aspect, some embodiments of the present disclosure provide a transformer maintenance equipment control apparatus, the apparatus comprising: an acquisition unit configured to acquire audio information of a target transformer; a preprocessing unit configured to preprocess the above-mentioned audio information to generate a preprocessed audio information set; the first generation unit is configured to generate a linear characteristic parameter information set according to the preprocessed audio information set; a second generating unit configured to generate a nonlinear characteristic parameter information set according to the preprocessed audio information set; a third generation unit configured to generate a fusion feature parameter information set according to the linear feature parameter information set and the nonlinear feature parameter information set; a fourth generation unit configured to generate a key fusion feature parameter information set according to the fusion feature parameter information set; a fifth generating unit configured to generate transformer fault prediction information according to the key fusion characteristic parameter information set and a pre-trained transformer fault prediction model, where the transformer fault prediction information includes: fault type and probability of failure; and a control unit configured to control an associated transformer maintenance device to perform maintenance operations on the target transformer in response to determining that the transformer fault prediction information characterizes the transformer as faulty.

In a third aspect, some embodiments of the present disclosure provide an electronic device comprising: one or more processors; a storage device having one or more programs stored thereon, which when executed by one or more processors causes the one or more processors to implement the method described in any of the implementations of the first aspect above.

In a fourth aspect, some embodiments of the present disclosure provide a computer readable medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the method described in any of the implementations of the first aspect above.

The above embodiments of the present disclosure have the following advantageous effects: by the control method of the transformer maintenance equipment, the power outage time can be shortened, and the waste of transformer maintenance resources can be reduced. Specifically, the reasons for the longer power outage time and the waste of transformer maintenance resources are as follows: in the process of predicting faults of the transformer, the extracted characteristics are higher in recognition rate only in an audio environment with high signal-to-noise ratio, so that the recognition accuracy of audio information in the environment with low signal-to-noise ratio is lower, the accuracy of predicting faults of the transformer is lower, the error rate of maintaining the transformer is higher, and further, the power outage time is longer and the maintenance resource of the transformer is wasted. Based on this, the transformer maintenance device control method of some embodiments of the present disclosure first acquires the audio information of the target transformer. Thus, the audio information of the target transformer can be obtained, and the method can be used for fault identification of the target transformer. The audio information is then preprocessed to generate a set of preprocessed audio information. Thus, a set of preprocessed audio information is obtained, which can be used for feature extraction of audio information. And then, generating a linear characteristic parameter information set according to the preprocessed audio information set. Therefore, the linear characteristic parameter information set can be obtained, and the method can be used for improving the fault identification accuracy of the transformer in an environment with high signal-to-noise ratio. And then, generating a nonlinear characteristic parameter information set according to the preprocessed audio information set. Thus, a nonlinear characteristic parameter information set can be obtained. Therefore, the method can be used for improving the fault recognition accuracy of the transformer in an environment with low signal-to-noise ratio. And secondly, generating a fusion characteristic parameter information set according to the linear characteristic parameter information set and the nonlinear characteristic parameter information set. Therefore, the fusion characteristic parameter information set can be obtained, the sensitivity of the environment to the audio information can be weakened, and the characteristic characterization capability of the transformer is improved. And then, generating a key fusion characteristic parameter information set according to the fusion characteristic parameter information set. Thus, a key fusion characteristic parameter information set can be obtained. Therefore, the dimension of the characteristics of the transformer can be reduced, and the complexity of information processing is reduced. And then, generating transformer fault prediction information according to the key fusion characteristic parameter information set and a pre-trained transformer fault prediction model. The transformer fault prediction information may include: fault type and probability of failure. Thus, transformer fault prediction information can be obtained, and can be used to determine whether a transformer has a fault. Finally, in response to determining that the transformer fault prediction information indicates that the transformer is faulty, controlling associated transformer maintenance equipment to perform maintenance operations on the target transformer. Thereby, maintenance operations can be performed on the faulty transformer. Also, since the recognition accuracy in the audio environment with a high signal-to-noise ratio can be improved by generating the linear characteristic parameter information set, and since the failure recognition accuracy in the audio environment with a low signal-to-noise ratio can be improved by generating the nonlinear characteristic parameter information set. And meanwhile, the linear characteristic parameter information set and the nonlinear characteristic parameter information set are subjected to characteristic fusion, so that the accuracy of transformer fault identification can be improved on the whole, the error rate of transformer maintenance is reduced, the power outage time is shortened, and the waste of transformer maintenance resources is reduced.

Drawings

The above and other features, advantages, and aspects of embodiments of the present disclosure will become more apparent by reference to the following detailed description when taken in conjunction with the accompanying drawings. The same or similar reference numbers will be used throughout the drawings to refer to the same or like elements. It should be understood that the figures are schematic and that elements and components are not necessarily drawn to scale.

FIG. 1 is a flow chart of some embodiments of a method of controlling a transformer maintenance device according to the present disclosure;

FIG. 2 is a schematic structural view of some embodiments of a transformer maintenance equipment control device according to the present disclosure;

fig. 3 is a schematic structural diagram of an electronic device suitable for use in implementing some embodiments of the present disclosure.

Detailed Description

Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While certain embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete. It should be understood that the drawings and embodiments of the present disclosure are for illustration purposes only and are not intended to limit the scope of the present disclosure.

It should be noted that, for convenience of description, only the portions related to the present invention are shown in the drawings. Embodiments of the present disclosure and features of embodiments may be combined with each other without conflict.

It should be noted that the terms "first," "second," and the like in this disclosure are merely used to distinguish between different devices, modules, or units and are not used to define an order or interdependence of functions performed by the devices, modules, or units.

It should be noted that references to "one", "a plurality" and "a plurality" in this disclosure are intended to be illustrative rather than limiting, and those of ordinary skill in the art will appreciate that "one or more" is intended to be understood as "one or more" unless the context clearly indicates otherwise.

The names of messages or information interacted between the various devices in the embodiments of the present disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

The present disclosure will be described in detail below with reference to the accompanying drawings in conjunction with embodiments.

Fig. 1 illustrates a flow 100 of some embodiments of a transformer maintenance equipment control method according to the present disclosure. The control method of the transformer maintenance equipment comprises the following steps:

Step 101, obtaining audio information of a target transformer.

In some embodiments, an executing body (e.g., a computing device) of the transformer maintenance device control method may acquire audio information of the target transformer from the associated audio information acquisition device through a wired connection or a wireless connection. The associated audio information collecting device may be a device capable of collecting audio information. For example, the associated audio information gathering device may be a microphone. The target transformer may be any transformer for which a transformer failure prediction is required. The target transformer is not particularly limited herein. The audio information may be audio information collected by the associated audio information collection device. Specifically, the audio information may be a sound signal when the target transformer is operating. It should be noted that the wireless connection may include, but is not limited to, 3G/4G connections, wiFi connections, bluetooth connections, wiMAX connections, zigbee connections, UWB (ultra wideband) connections, and other now known or later developed wireless connection means.

Step 102, preprocessing the audio information to generate a preprocessed audio information set.

In some embodiments, the executing body may preprocess the audio information to generate a preprocessed audio information set.

In some optional implementations of some embodiments, the executing entity may preprocess the audio information to generate a preprocessed audio information set by:

and the first step, pre-emphasis processing is carried out on the audio information, and the audio information after the pre-emphasis processing is obtained to be used as pre-emphasis audio information. In practice, the executing body may perform pre-emphasis processing on the audio information through a first-order high-pass filter, so as to obtain the audio information after the pre-emphasis processing as pre-emphasis audio information. Thereby, the high frequency part of the voice signal can be improved, so that the spectrum of the voice signal becomes flat.

And secondly, carrying out framing treatment on the pre-emphasis audio information to obtain the pre-emphasis audio information after framing treatment as a framing audio information set. In practice, the executing body may perform frame division processing on the pre-emphasis audio information with a preset frame length threshold as a frame length and with a preset frame shift threshold as a frame shift, so as to obtain the pre-emphasis audio information after frame division processing as a frame division audio information set. The preset frame shift threshold is smaller than the preset frame length threshold, and the frame shift is the displacement of the next frame to the previous frame. The preset frame shift threshold may be a preset frame shift threshold. For example, the preset frame shift threshold may be 100. The preset frame length threshold may be a preset frame length threshold. For example, the preset frame length threshold may be 200.

And thirdly, windowing the framing audio information set to obtain the framing audio information set after windowing as a windowing audio information set. In practice, the executing body may determine, as the windowed audio information set, a product of a preset window function and each of the framed audio information included in the framed audio information set. Wherein, the preset window function can be expressed by the following formula:

。

wherein,indicating a preset frame length. Here, the preset frame length may be 200./>Representing the signal sequence number in the framed audio information included in the set of framed audio information. />Representing a preset window function.

And step four, determining the windowed audio information set as a preprocessed audio information set.

And 103, generating a linear characteristic parameter information set according to the preprocessed audio information set.

In some embodiments, the executing entity may generate a set of linear feature parameter information according to the set of pre-processed audio information.

In some optional implementations of some embodiments, according to the preprocessing the audio information set, the executing entity may generate the linear feature parameter information set by:

The first step is to perform frequency domain transformation processing on the preprocessed audio information set to generate a spectrum information set. The frequency domain transform process may include, but is not limited to: discrete time fourier transform, discrete fourier transform, and fast fourier transform. Here, the above-described frequency domain transform process may be a fast fourier transform. In practice, the execution body may perform a fast fourier transform process on the preprocessed audio information set to generate a preprocessed audio information set after the fast fourier transform process as the spectrum information set.

And a second step of determining, for each spectrum information in the spectrum information set, a square of a modulus of the spectrum information as power spectrum information.

And thirdly, generating a linear energy information set according to the determined power spectrum information. In practice, the execution body may perform filtering processing on each determined power spectrum information by using a mel filter, so as to obtain each power spectrum information after the filtering processing as the linear energy information set.

And step four, performing logarithmic transformation processing on each linear energy information included in the linear energy information set to generate a logarithmic linear energy information set.

And fifthly, performing time domain transformation on the logarithmic linear energy information set to obtain the logarithmic linear energy information set after the time domain transformation as a linear characteristic parameter information set. In practice, the execution body may perform discrete cosine transform processing on the log linear energy information set, to obtain a log linear energy information set after discrete cosine transform processing as a linear characteristic parameter information set.

And 104, generating a nonlinear characteristic parameter information set according to the preprocessed audio information set.

In some embodiments, the executing entity may generate a nonlinear feature parameter information set according to the preprocessed audio information set.

In some optional implementations of some embodiments, according to the preprocessing the audio information set, the executing entity may generate the nonlinear feature parameter information set by:

first, a nonlinear energy information set is generated according to the determined power spectrum information. In practice, the executing body may perform filtering processing on each determined power spectrum information by using a gamma-pass filter, so as to obtain each power spectrum information after the filtering processing as a non-linear energy information set.

And a second step of performing logarithmic transformation processing on each piece of nonlinear energy information included in the nonlinear energy information set to generate a logarithmic nonlinear energy information set.

And thirdly, performing time domain transformation processing on the logarithmic non-linear energy information set to obtain the logarithmic non-linear energy information set after the time domain transformation processing as a non-linear characteristic parameter information set. In practice, the execution body may perform discrete cosine transform processing on the log-nonlinear energy information set, to obtain a log-nonlinear energy information set after discrete cosine transform processing as a nonlinear characteristic parameter information set.

And 105, generating a fusion characteristic parameter information set according to the linear characteristic parameter information set and the nonlinear characteristic parameter information set.

In some embodiments, the execution body may generate a fused set of feature parameter information according to the set of linear feature parameter information and the set of nonlinear feature parameter information. The linear characteristic parameter information in the linear characteristic parameter information set comprises an energy parameter and a linear characteristic parameter set. The nonlinear characteristic parameter information in the nonlinear characteristic parameter information set comprises an energy parameter and a nonlinear characteristic parameter set. Here, the energy parameter may be a volume of the audio information.

In some optional implementations of some embodiments, the executing entity may generate the fused set of characteristic parameter information according to the set of linear characteristic parameter information and the set of nonlinear characteristic parameter information by:

and the first step is to delete the energy parameters included in the linear characteristic parameter information set to obtain each piece of linear characteristic parameter information after deletion as a linear parameter information set. The linear parameter information in the linear parameter information set corresponds to the linear characteristic parameters in the linear characteristic parameter set one by one.

And a second step of deleting the energy parameters included in the nonlinear characteristic parameter information set to obtain each piece of nonlinear characteristic parameter information after the deleting process as a nonlinear parameter information set. The nonlinear parameter information in the nonlinear parameter information set corresponds to the nonlinear characteristic parameters in the nonlinear characteristic parameter set one by one.

And thirdly, combining the linear parameter information set and the nonlinear parameter information set to generate a fusion characteristic parameter information set. In practice, the execution body may perform a series combination process on the linear parameter information set and the nonlinear parameter information set, so as to obtain a linear parameter information set and the nonlinear parameter information set after the series combination process as a fusion characteristic parameter information set.

And 106, generating a key fusion characteristic parameter information set according to the fusion characteristic parameter information set.

In some embodiments, the executing entity may generate the key fusion feature parameter information set according to the fusion feature parameter information set.

In some optional implementations of some embodiments, according to the fused feature parameter information set, the executing entity may generate the key fused feature parameter information set by:

and determining each linear characteristic parameter with the same characteristic dimension in the linear parameter information set as a linear characteristic parameter set with the same dimension.

And a second step of determining, for each of the determined identical-dimensional linear feature parameter sets, the average value of the identical-dimensional linear feature parameters included in the identical-dimensional linear feature parameter set as identical-dimensional linear average value information.

And thirdly, determining the determined linear average information with the same dimension as a linear average information set with the same dimension.

And fourthly, determining each nonlinear characteristic parameter with the same characteristic dimension in the nonlinear parameter information set as a nonlinear characteristic parameter set with the same dimension.

And fifthly, for each of the determined same-dimension nonlinear feature parameter sets, determining the mean value of the same-dimension nonlinear feature parameters included in the same-dimension nonlinear feature parameter set as the same-dimension nonlinear mean value information.

And sixthly, determining the determined nonlinear mean value information with the same dimension as a nonlinear mean value information set with the same dimension. The same-dimension nonlinear mean information in the same-dimension nonlinear mean information set corresponds to the same-dimension linear mean information in the same-dimension linear mean information set one by one.

And seventh, determining each fusion characteristic parameter with the same characteristic dimension in the fusion characteristic parameter information set as a fusion characteristic parameter set with the same dimension.

And eighth, determining the number of the fusion characteristic parameters included in the fusion characteristic parameter set with the same dimension as the number of the fusion characteristic parameters.

And ninth, for each identical-dimension fusion feature parameter set in the determined identical-dimension fusion feature parameter sets, determining the average value of each identical-dimension fusion feature parameter included in the identical-dimension fusion feature parameter set as identical-dimension fusion average value information.

And tenth, determining the determined fusion mean information of each same dimension as a fusion mean information set of the same dimension. The same dimension fusion mean value information in the same dimension fusion mean value information set corresponds to the same dimension nonlinear mean value information in the same dimension nonlinear mean value information set one by one. The same dimension fusion mean value information in the same dimension fusion mean value information set corresponds to the same dimension linear mean value information in the same dimension linear mean value information set one by one.

And eleventh step, generating a linear inter-class variance information set according to the same-dimension fusion mean information set and the same-dimension linear mean information set. In practice, first, for each of the same-dimension fusion mean information in the same-dimension fusion mean information set, the execution body may determine, as the first linear inter-class variance information, a square of a difference between the same-dimension fusion mean information and the same-dimension linear mean information corresponding to the same-dimension fusion mean information in the same-dimension linear mean information set. Then, a product of the first preset threshold and the first linear inter-class variance information may be determined as the linear inter-class variance information. Finally, each of the determined linear inter-class variance information may be determined as a set of linear inter-class variance information. The first preset threshold may be a preset threshold. Here, the first preset threshold may be 1/2.

And twelfth, generating a nonlinear inter-class variance information set according to the same-dimension fusion mean information set and the same-dimension nonlinear mean information set. In practice, first, for each of the same-dimension fusion mean information in the same-dimension fusion mean information set, the execution body may determine, as the first nonlinear inter-class variance information, a square of a difference between the same-dimension fusion mean information and the same-dimension nonlinear mean information corresponding to the same-dimension fusion mean information in the same-dimension nonlinear mean information set. Then, a product of the first preset threshold and the first nonlinear inter-class variance information may be determined as nonlinear inter-class variance information. Finally, each of the determined nonlinear inter-class variance information may be determined as a nonlinear inter-class variance information set.

And thirteenth, generating a linear intra-class variance information set according to the determined linear characteristic parameter sets with the same dimensions and the linear average information set with the same dimensions. In practice, the execution subject may generate the linear intra-class variance information set by:

a first step of, for each of the determined respective same-dimensional linear feature parameter sets, performing the steps of:

And a first sub-step of determining the same-dimension linear mean information corresponding to the same-dimension linear feature parameter set in the same-dimension linear mean information set as target same-dimension linear mean information.

And a second sub-step of determining, for each of the same-dimension linear feature parameters in the same-dimension linear feature parameter set, a square of a difference between the same-dimension linear feature parameter and the target same-dimension linear mean information as first linear intra-class variance information.

And a third sub-step of determining the sum of the determined respective first linear intra-class variance information as linear intra-class variance information.

And a second step of determining each of the determined second linear intra-class variance information as a set of linear intra-class variance information.

And fourteenth step, generating a nonlinear intra-class variance information set according to the determined nonlinear characteristic parameter sets with the same dimensions and the nonlinear mean value information set with the same dimensions. In practice, the execution subject may generate the nonlinear intra-class variance information set by:

a first step of, for each of the determined respective same-dimension nonlinear feature parameter sets, performing the steps of:

And a first sub-step of determining the same-dimension nonlinear mean information corresponding to the same-dimension nonlinear characteristic parameter set in the same-dimension nonlinear mean information set as second target same-dimension nonlinear mean information.

And a second sub-step of determining, for each of the same-dimension nonlinear feature parameters in the same-dimension nonlinear feature parameter set, a square of a difference between the same-dimension nonlinear feature parameter and the second target same-dimension nonlinear mean information as first nonlinear intra-class variance information.

And a third sub-step of determining the sum of the determined respective first nonlinear intraclass variance information as nonlinear intraclass variance information.

And a second step of determining each of the determined nonlinear intraclass variance information as a nonlinear intraclass variance information set.

Fifteenth, determining an inter-class variance information set according to the linear inter-class variance information set and the nonlinear inter-class variance information set. In practice, first, for each piece of linear inter-class variance information in the linear inter-class variance information set, the execution body may determine nonlinear inter-class variance information corresponding to the linear inter-class variance information in the nonlinear inter-class variance information set as target nonlinear inter-class variance information. Then, the sum of the linear inter-class variance information and the target nonlinear inter-class variance information may be determined as inter-class variance information. Finally, the determined individual inter-class variance information may be determined as an inter-class variance information set.

Sixteenth, generating a class variance information set according to the linear class variance information set, the nonlinear class variance information set and the fusion characteristic parameter quantity. In practice, first, for each piece of the linear-class variance information in the set of linear-class variance information, the execution body may determine nonlinear-class variance information corresponding to the linear-class variance information in the set of nonlinear-class variance information as target nonlinear-class variance information. Then, the sum of the linear intra-class variance information and the target nonlinear intra-class variance information may be determined as first intra-class variance information, and then, a ratio of the first intra-class variance information to the number of fusion feature parameters may be determined as intra-class variance information. Finally, the determined individual intra-class variance information may be determined as an intra-class variance information set.

Seventeenth, generating a characteristic value information set according to the inter-class variance information set and the intra-class variance information set. Wherein, the characteristic value information in the characteristic value information set characterizes the contribution degree of the characteristic value information corresponding to the characteristic. In practice, first, for each of the inter-class variance information sets, the execution body may determine, as the target intra-class variance information, intra-class variance information corresponding to the inter-class variance information in the intra-class variance information set. Then, a ratio of the inter-class variance information to the target intra-class variance information may be determined as feature value information. Finally, the determined individual characteristic value information may be determined as a characteristic value information set.

Eighteenth, determining each piece of characteristic value information meeting the key characteristic condition in the characteristic value information set as a key characteristic value information set. The key feature condition may be that feature value information is greater than or equal to a preset feature threshold. The preset feature threshold may be a preset threshold. For example, the preset feature threshold may be 0.3.

Nineteenth, determining a key fusion characteristic parameter information set according to the key characteristic value information set and the fusion characteristic parameter information set. In practice, first, for each piece of fused feature parameter information in the fused feature parameter information set, the execution body may determine each feature parameter corresponding to the key feature value information set in the fused feature parameter information as key fused feature parameter information. The determined respective key fusion characteristic parameter information may then be determined as a set of key fusion characteristic parameter information.

The first to nineteenth steps and related matters serve as an invention point of the embodiments of the present disclosure, and the second technical problem mentioned in the background art is solved, where in the process of performing fault prediction on the transformer, fault prediction is performed on all feature information, so that feature dimensions are higher, and thus complexity of model prediction is higher, complexity of CPU processing is higher, processing time is longer, and further, resource waste of CPU processor is caused. Factors that cause the CPU to waste processing resources are often as follows: in the process of carrying out fault prediction on the transformer, fault prediction is carried out on all the characteristic information, so that the characteristic dimension is higher, the complexity of model prediction is higher, the complexity of CPU processing is higher, the processing time is longer, and further the waste of CPU processor resources is caused. If the above factors are solved, the effect of reducing CPU resource waste can be achieved. To achieve this effect, first, each linear feature parameter having the same feature dimension in the above-described linear parameter information set is determined as the same-dimension linear feature parameter group. Therefore, the same-dimension linear characteristic parameter set for representing the single-dimension characteristic information under the condition of higher signal-to-noise ratio can be obtained. Then, for each of the determined identical-dimensional linear feature parameter sets, the average value of the identical-dimensional linear feature parameters included in the identical-dimensional linear feature parameter set is determined as identical-dimensional linear average value information. And determining the determined linear average information of each same dimension as a linear average information set of the same dimension. Thus, the same-dimension linear mean information set for representing the single-dimension mean information can be obtained. And then, determining each nonlinear characteristic parameter with the same characteristic dimension in the nonlinear parameter information set as a nonlinear characteristic parameter set with the same dimension. Therefore, the same dimension nonlinear characteristic parameter set for representing the single dimension characteristic information under the condition of low signal-to-noise ratio can be obtained. As for each of the determined identical-dimension nonlinear feature parameter sets, the average value of each of the identical-dimension nonlinear feature parameters included in the identical-dimension nonlinear feature parameter set is determined as identical-dimension nonlinear average value information. And then, determining the determined nonlinear mean value information with the same dimension as a nonlinear mean value information set with the same dimension. The same-dimension nonlinear mean information in the same-dimension nonlinear mean information set corresponds to the same-dimension linear mean information in the same-dimension linear mean information set. Therefore, the nonlinear mean value information set with the same dimension and used for representing the mean value information with the single dimension can be obtained. And then, determining each fusion characteristic parameter with the same characteristic dimension in the fusion characteristic parameter information set as a fusion characteristic parameter set with the same dimension. Therefore, the same dimension fusion characteristic parameter set for representing the single dimension characteristic information under various environmental conditions can be obtained. And secondly, determining the number of the fusion characteristic parameters included in the fusion characteristic parameter set with the same dimension as the number of the fusion characteristic parameters. Thus, a fused feature parameter number characterizing the total feature number can be obtained. And then, for each of the determined identical-dimension fusion feature parameter sets, determining the average value of the identical-dimension fusion feature parameters included in the identical-dimension fusion feature parameter set as identical-dimension fusion average value information. And determining the determined fusion mean value information of each same dimension as a fusion mean value information set of the same dimension. The same dimension fusion mean information in the same dimension fusion mean information set corresponds to the same dimension nonlinear mean information in the same dimension nonlinear mean information set, and the same dimension fusion mean information in the same dimension fusion mean information set corresponds to the same dimension linear mean information in the same dimension linear mean information set. Therefore, the same dimension fusion mean value information set for representing the single dimension characteristic mean value information can be obtained. And then generating a linear inter-class variance information set according to the same-dimension fusion mean information set and the same-dimension linear mean information set. Therefore, the linear inter-class variance information set for representing the inter-class feature discrete degree under the condition of higher signal-to-noise ratio can be obtained. And then generating a nonlinear inter-class variance information set according to the same-dimension fusion mean information set and the same-dimension nonlinear mean information set. Therefore, the nonlinear inter-class variance information set for representing the inter-class feature discrete degree under the condition of low signal-to-noise ratio can be obtained. And generating a linear intra-class variance information set according to the determined linear characteristic parameter sets with the same dimensions and the linear mean information set with the same dimensions. Therefore, the linear inter-class variance information set for representing the characteristic discrete degree in the class under the condition of higher signal-to-noise ratio can be obtained. And then, generating a nonlinear intra-class variance information set according to the determined nonlinear characteristic parameter sets with the same dimensions and the nonlinear mean information set with the same dimensions. Therefore, the nonlinear inter-class variance information set for representing the characteristic discrete degree in the class under the condition of low signal-to-noise ratio can be obtained. And determining an inter-class variance information set according to the linear inter-class variance information set and the nonlinear inter-class variance information set. Thus, an inter-class variance information set can be obtained, so that the importance of the feature can be determined from the viewpoint of class. And then, generating a class inner variance information set according to the linear class inner variance information set and the nonlinear class inner variance information set. Thus, a set of intra-class variance information can be obtained, so that the importance of the feature can be determined from the perspective of the intra-class distance. And then, generating a characteristic value information set according to the inter-class variance information set and the intra-class variance information set. Wherein, the characteristic value information in the characteristic value information set characterizes the contribution degree of the characteristic value information corresponding to the characteristic. Thus, a feature value information set characterizing the feature contribution can be obtained. And finally, determining each characteristic value information meeting the key characteristic condition in the characteristic value information set as a key characteristic value information set. And determining a key fusion characteristic parameter information set according to the key characteristic value information set and the fusion characteristic parameter information set. Therefore, the key fusion characteristic parameter information set can be obtained, and the key fusion characteristic parameter information set can be used for carrying out dimension reduction processing on the characteristics. Also, since the inter-class distance size of the feature can be determined by generating the inter-class variance information set, the intra-class distance size of the feature can be determined by generating the intra-class variance information set. And the contribution degree of the features can be determined by determining the feature value information set, and the features are screened according to the contribution degree, so that the feature dimension can be reduced, the complexity of model prediction is reduced, the complexity of CPU processing is reduced, the processing time is shortened to be longer, and the waste of CPU processor resources is reduced.

And step 107, generating transformer fault prediction information according to the key fusion characteristic parameter information set and the pre-trained transformer fault prediction model.

In some embodiments, the executing entity may generate the transformer fault prediction information according to the key fusion feature parameter information set and a pre-trained transformer fault prediction model. The transformer fault prediction information may include: fault type and probability of failure.

Optionally, the above execution body may further execute the following steps:

first, acquiring a fault state information set of the target transformer. Wherein, the fault state information in the fault state information set may include, but is not limited to, any of the following: high voltage transformer, low voltage transformer, transformer current, upper layer oil temperature and winding temperature.

And secondly, carrying out normalization processing on each piece of fault state information included in the fault state information set, and obtaining each piece of fault state information after normalization processing as a fault state information vector. Wherein, the normalization process can include but is not limited to: maximum and minimum value normalization processing, zero mean normalization processing, logarithmic function normalization processing and arctangent function normalization processing. Here, the normalization process described above may be zero-mean normalization process. In practice, the execution body may perform zero-mean normalization processing on each fault state information included in the fault state information set, so as to obtain each fault state information after zero-mean normalization processing as a fault state information vector.

And thirdly, inputting the fault state information vector into an embedded layer of a pre-trained transformer fault state prediction model to obtain a low-dimensional fault state information vector. The transformer fault state prediction model further comprises a convolution network layer, a bidirectional circulation network layer, an attention mechanism layer and a classification layer. The transformer fault state prediction model may be a network model with a fault state information vector as input and a transformer fault prediction result as output. The convolutional network layer may be a network layer that converts input data into high-dimensional feature vectors by a convolutional kernel. The bi-directional cyclic network layer may be a network layer that converts input data having time sequential characteristics into a vector having time cumulative characteristics. The attention mechanism layer may be a network layer that converts an input vector into a vector with attention weights. The classification layer may be a network layer that classifies input data by a Softmax multiple classification function.

And step four, inputting the low-dimensional fault state information vector into the convolutional network layer to obtain a fault state convolutional vector. Here, the convolutional network layer may be a network layer that converts the low-dimensional fault state information into a fault state convolutional vector through a convolutional kernel.

And fifthly, inputting the fault state convolution vector into the bidirectional cyclic network layer to obtain a time characteristic fault state vector. Here, the bidirectional torus network layer may be a network layer that converts a fault state convolution vector having a time sequence feature into a time feature fault state vector having a time accumulation feature.

And sixthly, inputting the time characteristic fault state vector into the attention mechanism layer to obtain an updated fault state vector. Here, the attention mechanism layer may be a network layer that converts the time-feature fault state vector into an updated fault state vector with attention weights.

And seventhly, inputting the updated fault state vector into the classification layer to obtain a transformer fault prediction result. Wherein, the above-mentioned transformer trouble prediction result still includes: transformer fault type and transformer fault probability. The classification layer may be a network layer that classifies the updated fault state vector by a Softmax multiple classification function.

And eighth step, in response to determining that the fault type included in the transformer fault prediction information is the same as the transformer fault type included in the transformer fault prediction result, generating a transformer update fault probability according to the fault probability and the transformer fault probability.

And ninth, determining an updated transformer fault prediction result according to the transformer updated fault probability and the transformer fault type. In practice, the execution body may combine the transformer update failure probability and the transformer failure type to generate an updated transformer failure prediction result. Here, the combination may be splicing.

And tenth, determining the updated transformer fault prediction result as transformer fault prediction information so as to update the transformer fault prediction information.

In some optional implementations of some embodiments, the executing entity may generate the transformer update failure probability according to the failure probability and the transformer failure probability by:

and a first step of determining first prediction error information according to a preset threshold value and the fault probability. The first prediction error information is error information obtained by predicting the audio information of the target transformer. The preset threshold may be a preset threshold. Here, the preset threshold may be 1. In practice, the execution body may determine the absolute value of the difference between the preset threshold and the failure probability as the first prediction error information.

And a second step of determining second prediction error information according to the preset threshold value and the transformer fault probability. The second prediction error information is error information predicted by a fault state information set of the target transformer. In practice, the execution body may determine the absolute value of the difference between the preset threshold and the transformer failure probability as the second prediction error information.

And a third step of determining the reciprocal of the first prediction error information as first reciprocal error information.

A fourth step of determining the reciprocal of the second prediction error information as second reciprocal error information;

and a fifth step of determining the sum of the first error count information and the second error count information as error count information.

And a sixth step of determining a ratio of the first inverse error information to the inverse error information as a first prediction weight.

And seventh, determining the ratio of the second inverse error information to the inverse error information as a second prediction weight.

Eighth step, determining a first prediction probability according to the fault probability and the first prediction weight. In practice, the execution body may determine a product of the failure probability and the first prediction weight as a first prediction probability.

And a ninth step of determining a second prediction probability according to the transformer failure probability and the second prediction weight. In practice, the execution body may determine a product of the transformer failure probability and the second prediction weight as a second prediction probability.

And a tenth step of determining the sum of the first prediction probability and the second prediction probability as a transformer update failure probability.

The first to tenth steps and their related matters are taken as an invention point of the embodiments of the present disclosure, and the third technical problem mentioned in the background art is solved, in the process of performing fault prediction on a transformer, a prediction result of a single mode (audio information) is adopted as a prediction result, which results in lower accuracy of fault prediction of the transformer, higher error rate of transformer maintenance, and further longer power outage time and waste of transformer maintenance resources. Factors that lead to longer power outage and waste of transformer maintenance resources are often as follows: in the process of predicting faults of the transformer, a single-mode (audio information) prediction result is adopted as a prediction result, so that the accuracy of the transformer fault prediction is low, the error rate of transformer maintenance is high, and further, the power outage time is long and the transformer maintenance resource is wasted. If the above factors are solved, the effects of shortening the power outage time and reducing the waste of maintenance resources of the transformer can be achieved. To achieve this effect, first, a set of fault state information of the above-described target transformer is acquired. Wherein, the fault state information in the fault state information set may include, but is not limited to, any of the following: high voltage transformer, low voltage transformer, transformer current, upper layer oil temperature and winding temperature. Thus, a set of fault state information may be obtained, which may be used to determine the operating state of the target transformer from the perspective of the sensor data information. And then, carrying out normalization processing on each piece of fault state information included in the fault state information set to obtain each piece of fault state information after normalization processing as a fault state information vector. Thus, a fault state information vector can be obtained, so that the dimensional influence among indexes can be reduced, and the adverse influence of an abnormal sample can be reduced. And then, inputting the fault state information vector into an embedded layer of a pre-trained transformer fault state prediction model to obtain a low-dimensional fault state information vector. The transformer fault state prediction model may further include a convolutional network layer, a bidirectional cyclic network layer, an attention mechanism layer and a classification layer. Therefore, the low-dimensional fault state information vector can be obtained, and the method can be used for converting the high-dimensional sparse vector into the low-dimensional dense vector to realize dimension reduction. And then, inputting the low-dimensional fault state information vector into the convolution network layer to obtain a fault state convolution vector. Thus, a fault state convolution vector can be obtained, which can be used for feature extraction. And then, inputting the fault state convolution vector into the bidirectional circulating network layer to obtain a time characteristic fault state vector. Thus, a time-feature fault state vector having a time-cumulative feature can be obtained. Along with the time feature fault state vector is input into the attention mechanism layer, an updated fault state vector is obtained. Thus, an updated fault state vector with attention weighting characteristics can be obtained. And then, inputting the updated fault state vector into the classification layer to obtain a transformer fault prediction result. The transformer fault prediction result may include: transformer fault type and transformer fault probability. Thus, a transformer failure prediction result can be obtained, and thus can be used to determine whether a target transformer has a failure. And secondly, generating a transformer update fault probability according to the fault probability and the transformer fault probability in response to determining that the fault type included in the transformer fault prediction information is the same as the transformer fault type included in the transformer fault prediction result. Thus, a transformer update failure probability can be obtained, and thus can be used to determine whether a target transformer has a failure at the same time from the perspective of the transformer's audio signal and transformer sensor data. And then, determining an updated transformer fault prediction result according to the transformer updated fault probability and the transformer fault type. Therefore, the updated transformer fault prediction result with higher accuracy of the characterization prediction result can be obtained. And then determining the updated transformer fault prediction result as transformer fault prediction information so as to update the transformer fault prediction information. And then, determining first prediction error information according to a preset threshold value and the fault probability, wherein the first prediction error information is error information obtained through prediction of the audio information of the target transformer. Thus, the first prediction error information can be obtained, so that the prediction error can be determined from the perspective of the audio information. And determining second prediction error information according to the preset threshold value and the fault probability of the transformer. The second prediction error information is error information predicted by a fault state information set of the target transformer. Thereby, the second prediction error information can be obtained, so that the predicted error can be determined from the perspective of the sensor data. Then, the inverse of the first prediction error information is determined as first inverse error information. Thus, the first error reciprocal information can be obtained. Next, the reciprocal of the second prediction error information is determined as second reciprocal error information. Thus, the second error reciprocal information can be obtained. And determining the sum of the first error reciprocal information and the second error reciprocal information as error reciprocal information. Thus, inverse error information characterizing the audio signal and the sensor data information can be obtained. Then, a ratio of the first reciprocal error information to the reciprocal error information is determined as a first prediction weight. Thereby, a first prediction weight can be obtained, which can be used for determining weight information determined by the audio signal. And then, determining the ratio of the second error reciprocal information to the error reciprocal information as a second prediction weight. Thereby, a second predictive weight can be obtained, which can be used for determining weight information determined by the sensor data. Then, a first prediction probability is determined according to the fault probability and the first prediction weight. Thus, a first prediction probability characterizing the prediction of the audio signal having a weight influence may be obtained. Then, a second prediction probability is determined according to the transformer failure probability and the second prediction weight. Thus, a second prediction probability of the sensor data prediction with a weight influence can be obtained. And finally, determining the sum of the first prediction probability and the second prediction probability as the transformer update fault probability. Thus, the transformer update failure probability can be obtained. The method has the advantages that the accuracy of the transformer fault prediction can be improved by combining the fault result predicted by the audio signal with the two prediction modes of the fault result predicted by the sensor data, and the accuracy of the transformer fault prediction can be further improved by determining the respective prediction weights according to the error information predicted by the two prediction modes, so that the error rate of transformer maintenance is reduced, the power outage time is shortened, and the waste of transformer maintenance resources is reduced.

And step 108, controlling the associated transformer maintenance equipment to perform maintenance operation on the target transformer in response to determining that the transformer has faults according to the transformer fault prediction information.

In some embodiments, the executive may control an associated transformer maintenance device to perform a maintenance operation on the target transformer in response to determining that the transformer fault prediction information characterizes the transformer as faulty. The above-mentioned associated transformer maintenance device may be a device capable of performing fault maintenance on the transformer. For example, the above-described associated transformer maintenance devices may be smart robotic arms and smart robots. The transformer fault prediction information indicates that the transformer has a fault, which can be understood that the fault probability included in the transformer fault prediction information is greater than a preset fault threshold. Here, the preset failure threshold value may be a preset failure threshold value. For example, the preset failure threshold may be 0.75. In practice, the above-described executing entity may control the associated transformer maintenance device to perform a maintenance operation on the target transformer in response to determining that the transformer fault prediction information characterizes the transformer as faulty.

Optionally, the executing body may further control the associated sound playing device to play the fault-free prompting information of the transformer in response to determining that the transformer fault prediction information characterizes that the transformer is fault-free. Wherein the associated sound playing device may be a device capable of playing sound. For example, the above-described associated sound playback devices may include, but are not limited to: power amplifier, audio amplifier, multimedia console. The fault-free prompt information of the transformer can be information for prompting that the transformer has no fault and can be normally used. For example, the above-mentioned fault-free prompt message of the transformer may be "the transformer is fault-free temporarily and can be used normally".

With further reference to fig. 2, as an implementation of the method shown in the above figures, the present disclosure provides some embodiments of a transformer maintenance device control apparatus, which apparatus embodiments correspond to those method embodiments shown in fig. 1, and which apparatus is particularly applicable in various electronic devices.

As shown in fig. 2, the transformer maintenance equipment control device 200 of some embodiments includes: an acquisition unit 201, a preprocessing unit 202, a first generation unit 203, a second generation unit 204, a third generation unit 205, a fourth generation unit 206, a fifth generation unit 207, and a control unit 208. Wherein the acquisition unit 201 is configured to acquire audio information of the target transformer; the preprocessing unit 202 is configured to preprocess the above-mentioned audio information to generate a preprocessed audio information set; the first generating unit 203 is configured to generate a linear characteristic parameter information set according to the above-described preprocessed audio information set; the second generating unit 204 is configured to generate a nonlinear feature parameter information set according to the above-described preprocessed audio information set; the third generating unit 205 is configured to generate a fused characteristic parameter information set according to the above-mentioned linear characteristic parameter information set and the above-mentioned nonlinear characteristic parameter information set; the fourth generating unit 206 is configured to generate a key fusion feature parameter information set according to the fusion feature parameter information set; the fifth generating unit 207 is configured to generate transformer fault prediction information according to the above-mentioned key fusion characteristic parameter information set and a pre-trained transformer fault prediction model, where the above-mentioned transformer fault prediction information includes: fault type and probability of failure; the control unit 208 is configured to control the associated transformer maintenance equipment to perform maintenance operations on the target transformer in response to determining that the transformer fault prediction information characterizes a transformer fault.

It will be appreciated that the elements described in the apparatus 200 correspond to the various steps in the method described with reference to fig. 1. Thus, the operations, features and resulting benefits described above for the method are equally applicable to the apparatus 200 and the units contained therein, and are not described in detail herein.

Referring now to fig. 3, a schematic diagram of an electronic device 300 (e.g., a computing device) suitable for use in implementing some embodiments of the present disclosure is shown. The electronic devices in some embodiments of the present disclosure may include, but are not limited to, mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), car terminals (e.g., car navigation terminals), and the like, as well as stationary terminals such as digital TVs, desktop computers, and the like. The electronic device shown in fig. 3 is merely an example and should not impose any limitations on the functionality and scope of use of embodiments of the present disclosure.

As shown in fig. 3, the electronic device 300 may include a processing means 301 (e.g., a central processing unit, a graphics processor, etc.) that may perform various suitable actions and processes in accordance with a program stored in a Read Only Memory (ROM) 302 or a program loaded from a storage means 308 into a Random Access Memory (RAM) 303. In the RAM 303, various programs and data required for the operation of the electronic apparatus 300 are also stored. The processing device 301, the ROM 302, and the RAM 303 are connected to each other via a bus 304. An input/output (I/O) interface 305 is also connected to bus 304.

In general, the following devices may be connected to the I/O interface 305: input devices 306 including, for example, a touch screen, touchpad, keyboard, mouse, camera, microphone, accelerometer, gyroscope, etc.; an output device 307 including, for example, a Liquid Crystal Display (LCD), a speaker, a vibrator, and the like; storage 308 including, for example, magnetic tape, hard disk, etc.; and communication means 309. The communication means 309 may allow the electronic device 300 to communicate with other devices wirelessly or by wire to exchange data. While fig. 3 shows an electronic device 300 having various means, it is to be understood that not all of the illustrated means are required to be implemented or provided. More or fewer devices may be implemented or provided instead. Each block shown in fig. 3 may represent one device or a plurality of devices as needed.

In particular, according to some embodiments of the present disclosure, the processes described above with reference to flowcharts may be implemented as computer software programs. For example, some embodiments of the present disclosure include a computer program product comprising a computer program embodied on a computer readable medium, the computer program comprising program code for performing the method shown in the flow chart. In such embodiments, the computer program may be downloaded and installed from a network via communications device 309, or from storage device 308, or from ROM 302. The above-described functions defined in the methods of some embodiments of the present disclosure are performed when the computer program is executed by the processing means 301.

It should be noted that, the computer readable medium described in some embodiments of the present disclosure may be a computer readable signal medium or a computer readable storage medium, or any combination of the two. The computer readable storage medium can be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples of the computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In some embodiments of the present disclosure, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In some embodiments of the present disclosure, however, the computer-readable signal medium may comprise a data signal propagated in baseband or as part of a carrier wave, with the computer-readable program code embodied therein. Such a propagated data signal may take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination of the foregoing. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: electrical wires, fiber optic cables, RF (radio frequency), and the like, or any suitable combination of the foregoing.

In some implementations, the clients, servers may communicate using any currently known or future developed network protocol, such as HTTP (Hyper Text Transfer Protocol ), and may be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), the internet (e.g., the internet), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future developed networks.

The computer readable medium may be contained in the electronic device; or may exist alone without being incorporated into the electronic device. The computer readable medium carries one or more programs which, when executed by the electronic device, cause the electronic device to: acquiring audio information of a target transformer; preprocessing the audio information to generate a preprocessed audio information set; generating a linear characteristic parameter information set according to the preprocessed audio information set; generating a nonlinear characteristic parameter information set according to the preprocessed audio information set; generating a fusion characteristic parameter information set according to the linear characteristic parameter information set and the nonlinear characteristic parameter information set; generating a key fusion characteristic parameter information set according to the fusion characteristic parameter information set; generating transformer fault prediction information according to the key fusion characteristic parameter information set and a pre-trained transformer fault prediction model, wherein the transformer fault prediction information comprises: fault type and probability of failure; and controlling associated transformer maintenance equipment to perform maintenance operation on the target transformer in response to determining that the transformer is faulty according to the transformer fault prediction information.

Computer program code for carrying out operations for some embodiments of the present disclosure may be written in one or more programming languages, including an object oriented programming language such as Java, smalltalk, C ++ and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computer (for example, through the Internet using an Internet service provider).

The flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

The units described in some embodiments of the present disclosure may be implemented by means of software, or may be implemented by means of hardware. The described units may also be provided in a processor, for example, described as: a processor includes an acquisition unit, a preprocessing unit, a first generation unit, a second generation unit, a third generation unit, a fourth generation unit, a fifth generation unit, and a control unit. The names of these units do not constitute limitations on the unit itself in some cases, and the acquisition unit may also be described as "a unit that acquires audio information of a target transformer", for example.

The functions described above herein may be performed, at least in part, by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: a Field Programmable Gate Array (FPGA), an Application Specific Integrated Circuit (ASIC), an Application Specific Standard Product (ASSP), a system on a chip (SOC), a Complex Programmable Logic Device (CPLD), and the like.

The foregoing description is only of the preferred embodiments of the present disclosure and description of the principles of the technology being employed. It will be appreciated by those skilled in the art that the scope of the invention in the embodiments of the present disclosure is not limited to the specific combination of the above technical features, but encompasses other technical features formed by any combination of the above technical features or their equivalents without departing from the spirit of the invention. Such as the above-described features, are mutually substituted with (but not limited to) the features having similar functions disclosed in the embodiments of the present disclosure.

Claims

1. A method of controlling a transformer maintenance device, comprising:

acquiring audio information of a target transformer;

preprocessing the audio information to generate a preprocessed audio information set;

generating a linear characteristic parameter information set according to the preprocessed audio information set;

generating a nonlinear characteristic parameter information set according to the preprocessed audio information set;

generating a fusion characteristic parameter information set according to the linear characteristic parameter information set and the nonlinear characteristic parameter information set, wherein the linear characteristic parameter information in the linear characteristic parameter information set comprises an energy parameter and a linear characteristic parameter set, the nonlinear characteristic parameter information in the nonlinear characteristic parameter information set comprises an energy parameter and a nonlinear characteristic parameter set, and generating the fusion characteristic parameter information set according to the linear characteristic parameter information set and the nonlinear characteristic parameter information set comprises the following steps:

deleting energy parameters included in the linear characteristic parameter information set to obtain linear characteristic parameter information after deleting as a linear parameter information set, wherein the linear parameter information in the linear parameter information set corresponds to the linear characteristic parameters in the linear characteristic parameter set;

Deleting energy parameters included in the nonlinear characteristic parameter information set to obtain nonlinear characteristic parameter information after deleting as a nonlinear parameter information set, wherein the nonlinear parameter information in the nonlinear parameter information set corresponds to the nonlinear characteristic parameters in the nonlinear characteristic parameter set;

combining the linear parameter information set and the nonlinear parameter information set to generate a fusion characteristic parameter information set;

generating a key fusion characteristic parameter information set according to the fusion characteristic parameter information set, wherein the generating the key fusion characteristic parameter information set according to the fusion characteristic parameter information set comprises:

determining each linear characteristic parameter with the same characteristic dimension in the linear parameter information set as a linear characteristic parameter set with the same dimension;

for each of the determined same-dimension linear feature parameter sets, determining the mean value of each same-dimension linear feature parameter included in the same-dimension linear feature parameter set as same-dimension linear mean value information;

Determining the determined linear mean information of each same dimension as a linear mean information set of the same dimension;

determining each nonlinear characteristic parameter with the same characteristic dimension in the nonlinear parameter information set as a nonlinear characteristic parameter set with the same dimension;

for each of the determined same-dimension nonlinear feature parameter sets, determining the mean value of each same-dimension nonlinear feature parameter included in the same-dimension nonlinear feature parameter set as same-dimension nonlinear mean value information;

determining the determined nonlinear mean information with the same dimension as a nonlinear mean information set with the same dimension, wherein the nonlinear mean information with the same dimension in the nonlinear mean information set with the same dimension corresponds to the linear mean information with the same dimension in the linear mean information set with the same dimension;

determining each fusion characteristic parameter with the same characteristic dimension in the fusion characteristic parameter information set as a fusion characteristic parameter set with the same dimension;

determining the number of the fusion characteristic parameters included in the fusion characteristic parameter set with the same dimension as the number of the fusion characteristic parameters;

For each identical dimension fusion characteristic parameter set in the determined identical dimension fusion characteristic parameter sets, determining the average value of each identical dimension fusion characteristic parameter included in the identical dimension fusion characteristic parameter set as identical dimension fusion average value information;

determining the determined same-dimension fusion mean information as a same-dimension fusion mean information set, wherein the same-dimension fusion mean information in the same-dimension fusion mean information set corresponds to the same-dimension nonlinear mean information in the same-dimension nonlinear mean information set, and the same-dimension fusion mean information in the same-dimension fusion mean information set corresponds to the same-dimension linear mean information in the same-dimension linear mean information set;

generating a linear inter-class variance information set according to the same-dimension fusion mean information set and the same-dimension linear mean information set;

generating a nonlinear inter-class variance information set according to the same-dimension fusion mean information set and the same-dimension nonlinear mean information set;

generating a linear intra-class variance information set according to the determined linear characteristic parameter sets with the same dimension and the linear average information set with the same dimension;

Generating a nonlinear intra-class variance information set according to the determined nonlinear characteristic parameter sets with the same dimensions and the nonlinear mean information set with the same dimensions;

determining an inter-class variance information set according to the linear inter-class variance information set and the nonlinear inter-class variance information set;

generating an intra-class variance information set according to the linear intra-class variance information set and the nonlinear intra-class variance information set;

generating a characteristic value information set according to the inter-class variance information set and the intra-class variance information set, wherein characteristic value information in the characteristic value information set characterizes the contribution degree of the characteristic corresponding to the characteristic value information;

determining each characteristic value information meeting the key characteristic condition in the characteristic value information set as a key characteristic value information set;

determining a key fusion characteristic parameter information set according to the key characteristic value information set and the fusion characteristic parameter information set;

generating transformer fault prediction information according to the key fusion characteristic parameter information set and a pre-trained transformer fault prediction model, wherein the transformer fault prediction information comprises: fault type and probability of failure;

And controlling associated transformer maintenance equipment to perform maintenance operations on the target transformer in response to determining that the transformer fault prediction information indicates that the transformer is faulty.

2. The method of claim 1, wherein the method further comprises:

and controlling the associated sound playing equipment to play the fault-free prompt information of the transformer in response to the fact that the transformer fault prediction information represents the transformer to be fault-free.

3. The method of claim 1, wherein the preprocessing the audio information to generate a preprocessed set of audio information comprises:

pre-emphasis processing is carried out on the audio information, and the audio information after the pre-emphasis processing is obtained to be used as pre-emphasis audio information;

carrying out framing treatment on the pre-emphasis audio information to obtain pre-emphasis audio information subjected to framing treatment as a framing audio information set;

windowing is carried out on the framing audio information set, and the framing audio information set after windowing is obtained and used as a windowing audio information set;

the windowed audio information set is determined as a preprocessed audio information set.

4. The method of claim 1, wherein the generating a set of linear feature parameter information from the set of pre-processed audio information comprises:

Performing frequency domain transformation processing on the preprocessed audio information set to generate a frequency spectrum information set;

for each spectral information in the set of spectral information, determining a square of a modulus of the spectral information as power spectral information;

generating a linear energy information set according to the determined power spectrum information;

performing logarithmic transformation processing on each linear energy information included in the linear energy information set to generate a logarithmic linear energy information set;

and performing time domain transformation processing on the logarithmic linear energy information set to obtain the logarithmic linear energy information set after the time domain transformation processing as a linear characteristic parameter information set.

5. The method of claim 4, wherein the generating a set of nonlinear feature parameter information from the set of pre-processed audio information comprises:

generating a nonlinear energy information set according to the determined power spectrum information;

performing logarithmic transformation processing on each nonlinear energy information included in the nonlinear energy information set to generate a logarithmic nonlinear energy information set;

and performing time domain transformation processing on the logarithmic non-linear energy information set to obtain the logarithmic non-linear energy information set after the time domain transformation processing as a non-linear characteristic parameter information set.

6. A transformer maintenance equipment control device, comprising:

an acquisition unit configured to acquire audio information of a target transformer;

a preprocessing unit configured to preprocess the audio information to generate a preprocessed audio information set;

a first generating unit configured to generate a linear characteristic parameter information set according to the preprocessed audio information set;

a second generating unit configured to generate a nonlinear characteristic parameter information set according to the preprocessed audio information set;

a third generating unit configured to generate a fused feature parameter information set according to the linear feature parameter information set and the nonlinear feature parameter information set, wherein the linear feature parameter information in the linear feature parameter information set includes an energy parameter and a linear feature parameter set, the nonlinear feature parameter information in the nonlinear feature parameter information set includes an energy parameter and a nonlinear feature parameter set, and the generating the fused feature parameter information set according to the linear feature parameter information set and the nonlinear feature parameter information set includes:

a fourth generating unit, configured to generate a key fusion feature parameter information set according to the fusion feature parameter information set, where the generating a key fusion feature parameter information set according to the fusion feature parameter information set includes:

a fifth generating unit configured to generate transformer fault prediction information according to the key fusion characteristic parameter information set and a pre-trained transformer fault prediction model, wherein the transformer fault prediction information includes: fault type and probability of failure;

And a control unit configured to control an associated transformer maintenance device to perform a maintenance operation on the target transformer in response to determining that the transformer fault prediction information characterizes a transformer as faulty.

7. An electronic device, comprising:

one or more processors;

a storage device having one or more programs stored thereon,

when executed by the one or more processors, causes the one or more processors to implement the method of any of claims 1-5.

8. A computer readable medium having stored thereon a computer program, wherein the computer program, when executed by a processor, implements the method of any of claims 1-5.