Electric power operation and maintenance business management system and method

Abstract

The invention relates to the technical field of power grid dispatching, and discloses a power operation and maintenance service management system and method.The power operation and maintenance service management method comprises the following steps that operation and maintenance information is obtained, and the operation and maintenance information comprises historical monitoring information of all transformers, grade information of the transformers and information of all maintenance personnel; generating first two-dimensional structure data based on historical monitoring information codes of all transformers; inputting the operation and maintenance information into the distribution model, and outputting a distribution strategy; generating No.2 two-dimensional structure data based on the load of each transformer, the skill level of each maintainer and the distribution strategy; and inputting the second two-dimensional structure data into the power grid dispatching model, and outputting a power grid dispatching strategy. According to the invention, factors of evaluation of transformer grades and maintenance time of maintenance personnel are added when a maintenance personnel scheduling distribution strategy for transformer maintenance is carried out, and the conflict that the number of fault transformers is greater than that of the maintenance personnel is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid dispatching, and more specifically, to a power operation and maintenance business management system and method. Background Art

[0002] Transformers are critical equipment in power transmission lines, and their operating status directly impacts the reliability of the entire power supply system. Transformer damage can lead to widespread power supply problems, which, if not addressed promptly, can have a long-term impact on residents' daily lives. Therefore, timely detection of transformer faults, determination of their location and severity, and rapid dispatch of maintenance personnel to the site for repairs are crucial to ensuring stable power grid operation.

[0003] Traditional transformer maintenance personnel scheduling and allocation strategies may take fewer factors into consideration and fail to consider multiple factors. Since transformers are divided not only into fault levels but also into importance levels, there may be transformers that do not have spare transformers and are relatively important, but have minor faults. If left untreated for a long time, this may lead to subsequent larger faults and greater impacts. In addition, the maintenance time of each maintenance personnel is different, so it is necessary to increase the evaluation of the transformer level and the maintenance time of the maintenance personnel.

[0004] There is also an imbalance between the number of faulty transformers and the number of maintenance personnel. Multiple transformers may fail at the same time, but there are not enough maintenance personnel. Therefore, the load of the transformers in the power grid can be dispatched to alleviate the maintenance time of some faulty transformers, so that the number of transformers requiring emergency maintenance is less than the number of maintenance personnel. Summary of the Invention

[0005] The present invention provides an electric power operation and maintenance business management system and method, which solves the technical problems in the related art that the maintenance personnel scheduling and allocation strategy for traditional transformer maintenance does not take multiple factors into consideration and the number of transformers with possible faults is greater than the number of maintenance personnel.

[0006] The present invention provides a method for managing power operation and maintenance services, comprising the following steps:

[0007] Step 100: Acquire operation and maintenance information, which includes historical monitoring information of all transformers, transformer level information, and information of all maintenance personnel;

[0008] In a preferred embodiment, the historical monitoring information of the transformer includes: the transformer number, position, vibration, load, winding temperature, core grounding current, oil level, oil temperature, and fan speed at n consecutive moments before and including the current moment;

[0009] Transformer level information includes: transformer number, location, number of spare transformers, and power supply range;

[0010] The maintenance personnel information includes: maintenance personnel number, location, skill level, and maintenance fault level.

[0011] Step 200: Generate No. 1 two-dimensional structure data based on the historical monitoring information encoding of all transformers. The No. 1 two-dimensional structure data includes No. 1 data matrix and No. 1 relationship matrix. One unit of the No. 1 data matrix represents No. 1 one-dimensional structure data of a transformer. The element in the i-th row and j-th column of the No. 1 relationship matrix represents the association between the transformers represented by the i-th unit and the j-th unit of the No. 1 data matrix. If there is an association, the value of the element in the No. 1 relationship matrix is ​​1, otherwise it is 0.

[0012] The associations that exist between transformers refer to: there are line connections between transformers;

[0013] The one-dimensional structure data No. 1 includes n data items sorted by time, and the t-th data item represents the historical monitoring information of the transformer collected at the t-th moment;

[0014] Step 300: Input the operation and maintenance information into an allocation model, which includes a first intermediate layer, a second intermediate layer, a third intermediate layer, a fourth intermediate layer, a fifth intermediate layer, a feature fusion layer, a first output layer, a second output layer, a third output layer, and a fourth output layer; and output an allocation strategy, according to which maintenance personnel can be allocated to all transformers.

[0015] The first intermediate layer is used to input the first two-dimensional structure data and output the first intermediate representation data; the second intermediate layer is used to input the first intermediate representation data and output the second intermediate representation data to the feature fusion layer and the first output layer; the first output layer outputs the fault data representing all transformers; the third intermediate layer is used to input the level information of the transformer and output the third intermediate representation data to the feature fusion layer and the second output layer; the second output layer outputs the importance level of the transformer; the fourth intermediate layer is used to input the information of the maintenance personnel and output the fourth intermediate representation data to the feature fusion layer and the third output layer, and the third output layer outputs the fault repair time of the maintenance personnel; the feature fusion layer outputs the first fused intermediate representation data to the fourth output layer, and the fourth output layer outputs the allocation strategy;

[0016] In a preferred embodiment, the calculation formula of the first intermediate layer is as follows:

[0017] ;

[0018] ;

[0019] ;

[0020] ;

[0021] in, 、 、 represents the first, second and third weight parameters, 、 、 represents the first, second and third bias parameters, represents the dot product, 、 and They represent the first, second and third intermediate states of the vth unit of the data matrix No. 1 in the two-dimensional structure data No. 1, respectively. Represents the tth data item of the one-dimensional structure data corresponding to the vth unit of the data matrix No. 1 in the two-dimensional structure data No. 1, and They represent the t-th and t-1-th first intermediate representation data of the v-th unit of the data matrix No. 1 in the two-dimensional structure data No. 1, respectively. n≥t≥1, n represents the total number of data items of the one-dimensional structure data. When t=1 , tanh is the hyperbolic tangent function, Represents the sigmoid function.

[0022] In a preferred embodiment, the calculation formula of the second intermediate layer is as follows:

[0023] ;

[0024] ;

[0025] ;

[0026] in The second intermediate representation data of the vth unit of the data matrix No. 1, is the set of cells of data matrix No. 1 that are associated with the v-th cell of data matrix No. 1, and Respectively represent the nth first intermediate representation data of the vth unit and the uth unit of the data matrix No. 1 in the two-dimensional structure data No. 1, is a learnable scalar parameter, represents the first two-dimensional recognition weight parameter, and MLP represents the multi-layer perceptron.

[0027] In a preferred embodiment, the first and second intermediate layers are pre-trained independently, and the loss function of the pre-training is the mean square error loss function, which is formulated as follows:

[0028] ;

[0029] in represents the loss value, represents the total number of training samples, Indicates the The actual fault data of the transformer, The output of the second output layer Fault data of a transformer.

[0030] In a preferred embodiment, the calculation formula of the third intermediate layer is as follows:

[0031] ;

[0032] in, Indicates the data encoded into the transformer's level information, represents the first convolution function, Represents third intermediate representation data.

[0033] In a preferred embodiment, the calculation formula of the fourth intermediate layer is as follows:

[0034] ;

[0035] in, The data representing the maintenance personnel's information is encoded. represents the second convolution function, Indicates the fourth intermediate representation data.

[0036] In a preferred embodiment, the calculation formula of the feature fusion layer is as follows:

[0037] ;

[0038] in, represents the concatenation function, represents the sum weight matrix, represents the summation bias parameter, The second intermediate representation data of the vth unit of the data matrix No. 1, represents the third intermediate representation data, represents the fourth intermediate representation data, Represents the first fused intermediate representation data.

[0039] In a preferred embodiment, the calculation formula of the fifth intermediate layer is as follows:

[0040] ;

[0041] in The fifth intermediate representation data of the b-th unit of the second data matrix, is the set of cells of the data matrix No. 2 that are associated with the b-th cell of the data matrix No. 2, and Respectively represent the load or skill level of the independent object represented by the bth and cth units of the second data matrix, is a learnable scalar parameter, represents the second-dimensional recognition weight parameter, and MLP represents a multi-layer perceptron.

[0042] In a preferred embodiment, the third intermediate layer is independently pre-trained. During pre-training, the third intermediate layer is connected to the second output layer. The output of the second output layer represents the importance level of the transformer. The loss function of the pre-training of the third intermediate layer is a mean square error loss function, and the formula is as follows:

[0043] ;

[0044] in represents the loss value, represents the total number of training samples, Indicates the The true importance level of each transformer, The output of the second output layer The important level of a transformer.

[0045] In a preferred embodiment, the fourth intermediate layer is independently pre-trained. During pre-training, the fourth intermediate layer is connected to the third output layer. The output of the third output layer represents the importance level of all transformers. The loss function of the pre-training of the fourth intermediate layer is a mean square error loss function, and the formula is as follows:

[0046] ;

[0047] in represents the loss value, represents the total number of training samples, Indicates the The actual fault repair time of each maintenance personnel, The output of the third output layer The maintenance time of a maintenance personnel.

[0048] Step 400: Generate second two-dimensional structure data based on the load of each transformer, the skill level of each maintenance personnel, and the allocation strategy. The structure data includes a second data matrix and a second relationship matrix. Each cell of the second data matrix represents information about an independent object, including transformers, maintenance personnel, and allocation strategies. A cell of the second data matrix only contains information about the independent object it represents. The element in the i-th row and j-th column of the second relationship matrix represents the association between the i-th cell of the second data matrix and the independent object represented by the j-th cell. If an association exists, the value of the element in the second relationship matrix is ​​1, otherwise it is 0.

[0049] The relationship between maintenance personnel and transformers refers to: whether the maintenance personnel and transformers are matched;

[0050] The association between the allocation strategy and transformers and maintenance personnel means that: the allocation strategy includes transformers and maintenance personnel;

[0051] Step 500, input the second two-dimensional structure data into the power grid dispatching model, the power grid dispatching model includes a fifth intermediate layer and a fifth output layer, the fifth intermediate layer is used to input the second two-dimensional structure data and output the fifth intermediate representation data, the fifth output layer inputs the fifth intermediate representation data and outputs the power grid dispatching strategy.

[0052] The beneficial effects of the present invention are as follows: when scheduling and allocating maintenance personnel for transformer maintenance, factors such as the evaluation of transformer grades and the evaluation of maintenance personnel's maintenance time are taken into consideration; when the number of faulty transformers is greater than the number of maintenance personnel, the load of the transformers in the power grid is scheduled so that the number of transformers requiring emergency maintenance is less than the number of maintenance personnel. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 It is a flow chart of a power operation and maintenance business management method of the present invention. DETAILED DESCRIPTION

[0054] The subject matter described herein will now be discussed with reference to example embodiments. It should be understood that these embodiments are discussed solely to enable those skilled in the art to better understand and implement the subject matter described herein, and that the functions and arrangements of the elements discussed may be varied without departing from the scope of this specification. Various examples may omit, substitute, or add various processes or components as needed. Furthermore, features described in some examples may be combined in other examples.

[0055] At least one embodiment of the present invention discloses a method for managing power operation and maintenance services, such as Figure 1 Shown, including:

[0056] Step 100: Acquire operation and maintenance information, which includes historical monitoring information of all transformers, transformer level information, and information of all maintenance personnel;

[0057] The transformer's historical monitoring information includes: the transformer's number, location, vibration, load, winding temperature, core grounding current, oil level, oil temperature, and fan speed for n consecutive moments before and including the current moment;

[0058] Transformer level information includes: transformer number, location, number of spare transformers, and power supply range;

[0059] The maintenance personnel information includes: maintenance personnel number, location, skill level, and maintenance fault level.

[0060] Step 200: Generate No. 1 two-dimensional structure data based on the historical monitoring information encoding of all transformers. The No. 1 two-dimensional structure data includes No. 1 data matrix and No. 1 relationship matrix. One unit of the No. 1 data matrix represents No. 1 one-dimensional structure data of a transformer. The element in the i-th row and j-th column of the No. 1 relationship matrix represents the association between the transformers represented by the i-th unit and the j-th unit of the No. 1 data matrix. If there is an association, the value of the element in the No. 1 relationship matrix is ​​1, otherwise it is 0.

[0061] The associations that exist between transformers refer to: there are line connections between transformers;

[0062] The one-dimensional structure data No. 1 includes n data items sorted by time, and the t-th data item represents the historical monitoring information of the transformer collected at the t-th moment;

[0063] Step 300: Input the operation and maintenance information into an allocation model, which includes a first intermediate layer, a second intermediate layer, a third intermediate layer, a fourth intermediate layer, a fifth intermediate layer, a feature fusion layer, a first output layer, a second output layer, a third output layer, and a fourth output layer; and output an allocation strategy, according to which maintenance personnel can be allocated to all transformers.

[0064] The first intermediate layer is used to input the first two-dimensional structure data and output the first intermediate representation data; the second intermediate layer is used to input the first intermediate representation data and output the second intermediate representation data to the feature fusion layer and the first output layer; the first output layer outputs the fault data representing all transformers; the third intermediate layer is used to input the level information of the transformer and output the third intermediate representation data to the feature fusion layer and the second output layer; the second output layer outputs the importance level of the transformer; the fourth intermediate layer is used to input the information of the maintenance personnel and output the fourth intermediate representation data to the feature fusion layer and the third output layer, and the third output layer outputs the fault repair time of the maintenance personnel; the feature fusion layer outputs the first fused intermediate representation data to the fourth output layer, and the fourth output layer outputs the allocation strategy;

[0065] In one embodiment of the present invention, the fault levels are divided into no fault, minor fault, general fault, serious fault, and extremely serious fault, and the corresponding maintenance personnel's skill levels are also divided into five skill levels corresponding to the fault levels.

[0066] In one embodiment of the present invention, the calculation formula of the first intermediate layer is as follows:

[0067] ;

[0068] ;

[0069] ;

[0070] ;

[0071] in, 、 、 represents the first, second and third weight parameters, 、 、 represents the first, second and third bias parameters, represents the dot product, 、 and They represent the first, second and third intermediate states of the vth unit of the data matrix No. 1 in the two-dimensional structure data No. 1, respectively. Represents the tth data item of the one-dimensional structure data corresponding to the vth unit of the data matrix No. 1 in the two-dimensional structure data No. 1, and They represent the t-th and t-1-th first intermediate representation data of the v-th unit of the data matrix No. 1 in the two-dimensional structure data No. 1, respectively. n≥t≥1, n represents the total number of data items of the one-dimensional structure data. When t=1 , tanh is the hyperbolic tangent function, Represents the sigmoid function.

[0072] In one embodiment of the present invention, the calculation formula of the second intermediate layer is as follows:

[0073] ;

[0074] ;

[0075] ;

[0076] in The second intermediate representation data of the vth unit of the data matrix No. 1, is the set of cells of data matrix No. 1 that are associated with the v-th cell of data matrix No. 1, and Respectively represent the nth first intermediate representation data of the vth unit and the uth unit of the data matrix No. 1 in the two-dimensional structure data No. 1, is a learnable scalar parameter, represents the first two-dimensional recognition weight parameter, and MLP represents the multi-layer perceptron.

[0077] In one embodiment of the present invention, the first and second intermediate layers are pre-trained independently. During pre-training, the second intermediate layer is connected to the first output layer. The calculation formula of the first output layer is as follows:

[0078] ;

[0079] in Represents a vector, where the ath component contains three sub-vectors, where the first, second, and third sub-vectors of the ath component represent the fault level, fault time, and fault type of the ath transformer, respectively. represents the second intermediate representation data of the vth unit of the data matrix No. 1, M1 represents the set of all units of the data matrix No. 1, is the fourth weight parameter, is the fourth bias parameter, Represents the sigmoid function.

[0080] The loss function of pre-training is the mean square error loss function, and the formula is as follows:

[0081] ;

[0082] in represents the loss value, represents the total number of training samples, Indicates the The actual fault data of the transformer, The output of the second output layer Fault data of a transformer.

[0083] In one embodiment of the present invention, the calculation formula of the third intermediate layer is as follows:

[0084] ;

[0085] in, Indicates the data encoded into the transformer's level information, represents the first convolution function, Represents third intermediate representation data.

[0086] In one embodiment of the present invention, the calculation formula of the fourth intermediate layer is as follows:

[0087] ;

[0088] in, The data representing the maintenance personnel's information is encoded. represents the second convolution function, Indicates the fourth intermediate representation data.

[0089] In one embodiment of the present invention, the calculation formula of the feature fusion layer is as follows:

[0090] ;

[0091] in, represents the concatenation function, represents the sum weight matrix, represents the summation bias parameter, The second intermediate representation data of the vth unit of the data matrix No. 1, represents the third intermediate representation data, represents the fourth intermediate representation data, Represents the first fused intermediate representation data.

[0092] In one embodiment of the present invention, the calculation formula of the fourth output layer is as follows:

[0093] ;

[0094] in, Represents the first fused intermediate representation data, represents the third convolution function, Represents a matrix, where the element in the i-th row and j-th column of the matrix represents the assignment relationship between the i-th maintenance personnel and the j-th transformer. If the element value is greater than 0.5, it means that the i-th maintenance personnel is matched with the j-th transformer, otherwise it means they are not matched.

[0095] Step 400: Generate second two-dimensional structure data based on the load of each transformer, the skill level of each maintenance personnel, and the allocation strategy. The structure data includes a second data matrix and a second relationship matrix. Each cell of the second data matrix represents information about an independent object, including transformers, maintenance personnel, and allocation strategies. A cell of the second data matrix only contains information about the independent object it represents. The element in the i-th row and j-th column of the second relationship matrix represents the association between the i-th cell of the second data matrix and the independent object represented by the j-th cell. If an association exists, the value of the element in the second relationship matrix is ​​1, otherwise it is 0.

[0096] The relationship between maintenance personnel and transformers refers to: whether the maintenance personnel and transformers are matched;

[0097] The association between the allocation strategy and transformers and maintenance personnel means that: the allocation strategy includes transformers and maintenance personnel;

[0098] The information about the transformer includes the transformer load, the information about the maintenance personnel includes the skill level of the maintenance personnel, and the information about the allocation strategy includes the allocation strategy.

[0099] Step 500, input the second two-dimensional structure data into the power grid dispatching model, the power grid dispatching model includes a fifth intermediate layer and a fifth output layer, the fifth intermediate layer is used to input the second two-dimensional structure data and output the fifth intermediate representation data, the fifth output layer inputs the fifth intermediate representation data and outputs the power grid dispatching strategy.

[0100] In one embodiment of the present invention, the calculation formula of the fifth intermediate layer is as follows:

[0101] ;

[0102] in The fifth intermediate representation data of the b-th unit of the second data matrix, is the set of cells of the data matrix No. 2 that are associated with the b-th cell of the data matrix No. 2, and Respectively represent the load or skill level of the independent object represented by the bth and cth units of the second data matrix, is a learnable scalar parameter, represents the second-dimensional recognition weight parameter, and MLP represents a multi-layer perceptron.

[0103] In one embodiment of the present invention, the calculation formula of the fifth output layer is as follows:

[0104] ;

[0105] in Represents the output of a vector, each vector represents the probability of an action matrix, and the element in the i-th row and j-th column of the action matrix represents the load transfer amount from the i-th transformer to the j-th transformer. represents the fifth intermediate representation data of the bth unit of the second data matrix, M2 represents a set of units of the second data matrix related to the transformer load transfer amount, is the fifth weight parameter, is the fourth bias parameter, Represents the sigmoid function.

[0106] In one embodiment of the present invention, the element in the i-th row and j-th column of the action matrix represents the load transfer amount from the i-th transformer to the j-th transformer, wherein the load transfer amount is discretized and an action in the action space represents an action matrix formed by a free combination of the load transfer amounts of all transformers to the remaining connected transformers.

[0107] In one embodiment of the present invention, the third intermediate layer is independently pre-trained. During pre-training, the third intermediate layer is connected to the second output layer. The output of the second output layer represents the importance level of the transformer. The loss function of the pre-training of the third intermediate layer is a mean square error loss function, and the formula is as follows:

[0108] ;

[0109] in represents the loss value, represents the total number of training samples, Indicates the The true importance level of each transformer, The output of the second output layer The important level of a transformer.

[0110] In one embodiment of the present invention, the fourth intermediate layer is independently pre-trained. During pre-training, the fourth intermediate layer is connected to the third output layer. The output of the third output layer represents the importance level of all transformers. The loss function of the pre-training of the fourth intermediate layer is a mean square error loss function, and the formula is as follows:

[0111] ;

[0112] in represents the loss value, represents the total number of training samples, Indicates the The actual fault repair time of each maintenance personnel, The output of the third output layer The maintenance time of a maintenance personnel.

[0113] In one embodiment of the present invention, the step of training the distribution model includes:

[0114] Step 101: Build a target network and copy the target network distribution model;

[0115] Step 102: Randomly obtain an experience from the experience pool. The experience includes the state at time e. , the action performed at time e , the state at time e+1 , perform the action Rewards received ;

[0116] The states at time e and e+1 represent the operation and maintenance information at time e and time e+1 input to the allocation model, respectively;

[0117] Then calculate the state The corresponding target network output The probability value of ;

[0118] ;

[0119] express The maximum probability value output by the target network;

[0120] Step 103, according to and Calculate the loss value loss and update the distribution model;

[0121] ;

[0122] express The output of the next distribution model corresponds to The probability value of

[0123] is the discount factor, a value between 0 and 1;

[0124] Step 104: Update the target network every fixed number of network updates so that its parameters are the same as those of the current allocation model;

[0125] Step 105: The process ends when the number of collected experiences reaches a set threshold or the allocation model converges.

[0126] In one embodiment of the present invention, the experience is derived from the record of historical action execution.

[0127] In one embodiment of the present invention, the reward The calculation formula is as follows:

[0128] ;

[0129] in represents the repair time of the fault from time e to time e+1, Indicates the current standard value of fault repair time. Indicates the importance level of the transformer, Indicates the standard value of importance level, Indicates the difference between the maintenance personnel's skill level and the fault level.

[0130] In one embodiment of the present invention, the steps of training the power grid dispatch model include:

[0131] Step 201, initializing the parameters of the power grid dispatch model;

[0132] Step 202: Observe the transformer load information at time f , the strategy executed at time f , transformer load information at time f+1 , execution strategy Rewards received ;

[0133] Step 203, then calculate the policy error:

[0134] ;

[0135] in represents the policy error at time f, represents the discount factor, , Represents the power grid dispatch model input The maximum probability value in the first output vector output when , Represents the power grid dispatch model input The first output vector outputted when the strategy The probability value of

[0136] ;

[0137] in represents the frequency of the power grid, represents the target frequency, Expressed as frequency deviation, Indicates reward.

[0138] Step 204: Update the grid dispatch model. The updated formula is as follows:

[0139] ;

[0140] , represents the step size of deep learning, Indicates delivery update;

[0141] Step 205 , iterate steps 202 - 204 until the grid dispatching model converges or the number of iterations reaches a set value, the default value of which is 100.

[0142] The above describes an embodiment of the present invention, but this embodiment is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Ordinary technicians in this field can also make more forms of equivalent embodiments based on the inspiration of this embodiment, all of which are protected by this embodiment.

Claims

1. A method for managing power operation and maintenance services, characterized in that: The following steps are involved: Step 100, obtaining operation and maintenance information, which includes historical monitoring information of all transformers, transformer level information and information of all maintenance personnel; Step 200, based on the historical monitoring information encoding of all transformers, generate No. 1 two-dimensional structure data, the No. 1 two-dimensional structure data includes No. 1 data matrix and No. 1 relationship matrix, one unit of the No. 1 data matrix represents No. 1 one-dimensional structure data of a transformer; the element of the No. ith row and the No. ith column of the No. 1 relationship matrix represents the association between the transformers represented by the No. ith unit and the No. ith unit of the No. 1 data matrix, if there is an association, the element value of the No. 1 relationship matrix is ​​1, otherwise it is 0; The associations that exist between transformers refer to: there are line connections between transformers; The one-dimensional structure data No. 1 includes n data items sorted by time, and the t-th data item represents the historical monitoring information of the transformer collected at the t-th moment; Step 300, inputting the operation and maintenance information into the allocation model, the allocation model comprising a first intermediate layer, a second intermediate layer, a third intermediate layer, a fourth intermediate layer, a fifth intermediate layer, a feature fusion layer, a first output layer, a second output layer, a third output layer and a fourth output layer; The first intermediate layer is used to input the first two-dimensional structure data and output the first intermediate representation data; The second intermediate layer is used to input the first intermediate representation data, and output the second intermediate representation data to the feature fusion layer and the first output layer; the first output layer outputs the fault data representing all transformers; the third intermediate layer is used to input the level information of the transformer, and output the third intermediate representation data to the feature fusion layer and the second output layer; the second output layer outputs the importance level of the transformer; The fourth intermediate layer is used to input the maintenance personnel's information, output the fourth intermediate representation data to the feature fusion layer and the third output layer, and the third output layer outputs the maintenance personnel's fault repair time; The feature fusion layer outputs the first fused intermediate representation data to the fourth output layer, and the fourth output layer outputs the allocation strategy, according to which maintenance personnel can be allocated to all transformers.

2. A method for managing electric power operation and maintenance services according to claim 1, characterized in that: Also includes: Step 400, based on the load of each transformer, the skill level of each maintenance personnel and the allocation strategy, generate No. 2 two-dimensional structure data, including No. 2 data matrix and No. 2 relationship matrix, one unit of the No. 2 data matrix represents the information of an independent object, the independent object includes a transformer, a maintenance personnel and an allocation strategy, and one unit of the No. 2 data matrix only contains the information of the independent object it represents; the element of the i-th row and j-th column of the No. 2 relationship matrix represents the association between the i-th unit of the No. 2 data matrix and the independent object represented by the j-th unit, if there is an association, the value of the element of the No. 2 relationship matrix is ​​1, otherwise it is 0; The relationship between maintenance personnel and transformers refers to: whether the maintenance personnel and transformers are matched; The association between the allocation strategy and transformers and maintenance personnel refers to: the allocation strategy includes transformers and maintenance personnel; Step 500, input the second two-dimensional structure data into the power grid dispatching model, the power grid dispatching model includes a fifth intermediate layer and a fifth output layer, the fifth intermediate layer is used to input the second two-dimensional structure data and output the fifth intermediate representation data, the fifth output layer inputs the fifth intermediate representation data and outputs the power grid dispatching strategy.

3. A method for managing electric power operation and maintenance services according to claim 1, characterized in that: The calculation formula for the first intermediate layer is as follows: ； ； ； ； in, , , represents the first, second, and third weight parameters, , , represents the first, second and third bias parameters, represents the dot product, , and They represent the first, second and third intermediate states of the vth unit of the data matrix No. 1 in the two-dimensional structure data No. 1, respectively. represents the tth data item of the one-dimensional structure data corresponding to the vth unit of the data matrix No. 1 in the two-dimensional structure data No. 1, and They represent the t-th and t-1-th first intermediate representation data of the v-th unit of the data matrix No. 1 in the two-dimensional structure data No. 1, respectively. n≥t≥1, n represents the total number of data items of the one-dimensional structure data. When t=1 , tanh is the hyperbolic tangent function, Represents the sigmoid function.

4. A method for managing electric power operation and maintenance services according to claim 1, characterized in that: The calculation formula for the second intermediate layer is as follows: ； ； ； in The second intermediate representation data of the vth unit of the data matrix No. 1, is the set of cells of data matrix No. 1 that are associated with the v-th cell of data matrix No. 1, and They respectively represent the nth first intermediate representation data of the vth unit and the uth unit of the data matrix No. 1 in the two-dimensional structure data No. 1, is a learnable scalar parameter, represents the first two-dimensional recognition weight parameter, and MLP represents a multi-layer perceptron.

5. A method for managing electric power operation and maintenance services according to claim 1, characterized in that: The first and second intermediate layers are pre-trained independently, and the loss function of the pre-training is the mean square error loss function, and the formula is as follows: ； in represents the loss value, represents the total number of training samples, Indicates The actual fault data of the transformer, The output of the second output layer Transformer fault data.

6. A method for managing electric power operation and maintenance services according to claim 1, characterized in that: The calculation formula for the third intermediate layer is as follows: ； in, Represents the data encoded with the transformer's level information, represents the first convolution function, Represents third intermediate representation data.

7. A method for managing electric power operation and maintenance services according to claim 1, characterized in that: The calculation formula for the fourth intermediate layer is as follows: ； in, Data representing maintenance personnel's information encoded into, represents the second convolution function, Represents fourth intermediate representation data.

8. A method for managing electric power operation and maintenance services according to claim 1, characterized in that: The calculation formula of the feature fusion layer is as follows: ； in, represents the concatenation function, represents the sum weight matrix, represents the summation bias parameter, The second intermediate representation data of the vth unit of the data matrix No. 1, represents the third intermediate representation data, represents the fourth intermediate representation data, Represents the first fused intermediate representation data.

9. A method for managing electric power operation and maintenance services according to claim 2, characterized in that: The calculation formula for the fifth intermediate layer is as follows: ； in The fifth intermediate representation data of the b-th unit of the second data matrix, is the set of cells of data matrix No. 2 that are associated with the b-th cell of data matrix No. 2, and They represent the load or skill level of the independent object represented by the bth and cth cells of the second data matrix, respectively. is a learnable scalar parameter, represents the second two-dimensional recognition weight parameter, and MLP represents a multi-layer perceptron.

10. A power operation and maintenance business management system, characterized in that: It is used to execute an electric power operation and maintenance business management method as described in any one of claims 1-9.

Images