Power grid main business decomposition method, device, equipment and medium

Abstract

The invention relates to a power grid main business decomposition method, device and equipment and a medium, and the method comprises the following steps: obtaining power grid business data, extracting business action description phrases, and carrying out the statistics of the execution frequency of the phrases; automatically distributing the extracted service action description phrases to the known first-level main service of the power grid based on semantic similarity; clustering is carried out based on the semantic similarity between the corresponding business action description phrases, the aggregation density and the resource consumption ratio of each group are calculated, effective groups are screened, and second-level and third-level business units of the power grid are generated in sequence; and for each power grid three-level service unit, phrase screening is carried out based on the semantic similarity between the corresponding service action description phrases and the phrase execution frequency proportion, and an underlying service unit is generated. Compared with the prior art, the method has the advantages of being capable of automatically and objectively decomposing the main business of the power grid and the like.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of text processing, in particular to a power grid main business decomposition method, device, equipment and medium. BACKGROUND

[0002] As the core support of national energy security, power grid business has the characteristics of complex business scenarios, numerous function modules, intensive resource investment (funds, manpower, materials), and fragmented execution links. The scientific decomposition of its main business is the prerequisite for realizing the optimization of resource allocation, the quantification of performance evaluation, and the standardization of business processes. At present, the mainstream method of main business decomposition in the power grid industry is still dominated by manual experience. With the rapid development of artificial intelligence, some technologies have also been applied to business decomposition. For example, Chinese patent CN120744108A discloses a processing method and system based on dual intelligent agents of business decomposition and visual presentation. The processing method includes: receiving a natural language request of a power system by a business decomposition intelligent agent, performing semantic analysis on the natural language request through a small language model, performing semantic retrieval based on a power system knowledge graph, realizing knowledge-enhanced intent recognition and task decomposition, and generating a corresponding task instruction chain in a directed acyclic graph structure; sending the task instruction chain to an instruction registration center, dispatching a visual presentation intelligent agent by the instruction registration center, and generating an interactive dashboard and a natural language summary through a large language model. This method realizes the automatic understanding and efficient decomposition of power system business, improves the intelligentization and automation level of data processing and visualization, enhances the task scheduling and management capability of the system, effectively reduces manual intervention, and improves the efficiency of operation and maintenance and the accuracy of data display. However, this method still has the following defects: 1) The business decomposition target deviates from the management core, and the decomposition result can only serve single data query, cannot form a structured business system, and is difficult to match subsequent resource allocation, performance evaluation, and other core management scenarios.

[0003] 2) Only relying on language model semantic analysis and knowledge graph retrieval, the decomposition result is greatly affected by the understanding accuracy of the model, and there are problems of fuzzy level attribution and non-reproducible results.

[0004] 3) Weak ability to dynamically adapt to business changes. If the power grid business is adjusted due to new energy access, power grid upgrading, etc., the knowledge graph and language model need to be updated simultaneously, which is high in adjustment cost and low in efficiency, and it is difficult to quickly respond to business dynamics.

[0005] Therefore, there is an urgent need for a technical solution that can specifically solve the above problems, realize the quantitative decomposition of main business, clear level, strong landing, and dynamic adaptation, and provide technical support for the fine management of power grid business. SUMMARY

[0006] The present application aims to provide a power grid main business decomposition method, device, equipment and medium to overcome the defects of the prior art.

[0007] The object of the present application can be achieved by the following technical solutions: According to a first aspect of the present application, a power grid main business decomposition method is provided, comprising the following steps: Obtaining power grid business data, extracting business action description phrases and counting phrase execution frequency; Based on the known power grid primary main business, the extracted business action description phrases are automatically assigned to the corresponding power grid primary main business based on semantic similarity; For each power grid primary main business, clustering is performed based on the semantic similarity between the corresponding business action description phrases, the aggregation density and resource consumption proportion of each group are calculated, the effective groups are screened, and the power grid secondary business unit is generated; For each power grid secondary business unit, clustering is performed based on the semantic similarity between the corresponding business action description phrases, the aggregation density and resource consumption proportion of each group are calculated, the effective groups are screened, and the power grid tertiary business unit is generated; For each power grid tertiary business unit, phrase screening is performed based on the semantic similarity between the corresponding business action description phrases and the phrase execution frequency proportion, and the bottom-level business unit is generated.

[0008] The counting method of the phrase execution frequency is: In a predetermined statistical period, the task description text in the work order of the power grid business data is extracted, the task description text is converted into a semantic vector, and the work order task description vector is obtained; The business action description phrase is converted into a semantic vector, and the business action description vector is obtained; The semantic similarity between the work order task description vector and each business action description vector is calculated, if the semantic similarity is greater than a predetermined threshold, the matching is successful, the number of work orders successfully matched with each business action description phrase is counted, and the work orders without work records in the matching results are eliminated, and the final number of work orders is taken as the phrase execution frequency of the business action description phrase.

[0009] The calculation method of the resource consumption proportion is: the total resource consumption of a group obtained by clustering under the current level is counted, the ratio of the total resource consumption in the group to the total resource consumption under the corresponding previous level is calculated, and the resource consumption proportion is obtained.

[0010] The aggregation density is calculated as follows: The sum of execution frequencies of phrases within a group obtained from clustering at the current level is calculated; the ratio of the sum of execution frequencies of phrases within the group to the total execution frequency of phrases at the corresponding previous level is calculated to obtain the execution frequency percentage; the absolute value of the difference between the execution frequency percentage and the resource consumption percentage is calculated; 1 minus the absolute value is used to calculate the matching degree; a weighted sum of the execution frequency percentage and the resource consumption percentage is calculated based on preset weights, and divided by the matching degree to obtain the aggregation density.

[0011] The method for generating the power grid secondary business unit or the power grid tertiary business unit is as follows: each valid group corresponds to a power grid secondary business unit or a power grid tertiary business unit. Based on the selected valid groups, the high-frequency core keywords of each valid group are extracted using the TextRank keyword extraction algorithm. The names of each valid group are then named in accordance with power grid industry conventions to obtain the names of the power grid secondary business unit or the power grid tertiary business unit.

[0012] The method further includes: verifying the decomposition results after the business decomposition is completed. Hierarchical attribution verification: Each underlying business unit is uniquely traced back to its corresponding primary power grid business. The semantic similarity between the business action description phrase in the underlying business unit and the primary power grid business is calculated. If the semantic similarity is lower than a preset threshold, the verification fails. Hierarchical coverage verification: Calculate the sum of the resource consumption ratios of the second-level business units and the third-level business units of the power grid. If the sum is less than the preset threshold, the verification fails. Non-overlap verification: For power grid level 2 business units and power grid level 3 business units, calculate the average semantic similarity of the business action description phrases between different business units within the current level. If the semantic similarity is greater than the preset threshold, the verification fails. If the validation fails, adjust the clustering parameters or the screening threshold, and perform the decomposition again.

[0013] The method further includes: using the analytic hierarchy process (AHP) to calculate the weight of each business unit obtained from the decomposition.

[0014] According to a second aspect of the present invention, a power grid main business decomposition apparatus is provided for implementing the method, the apparatus comprising: Data acquisition and phrase extraction module: used to acquire power grid business data, extract business action description phrases, and count the frequency of phrase execution; First-level decomposition module: Based on the known first-level main business of the power grid, it automatically assigns the extracted business action description phrases to the corresponding first-level main business of the power grid based on semantic similarity; The secondary decomposition module: for each power grid primary business, based on the semantic similarity between the corresponding business action description phrases, clustering is performed, the aggregation density and resource consumption proportion of each group are calculated, the effective group is screened, and the power grid secondary business unit is generated; The tertiary decomposition module: for each power grid secondary business unit, based on the semantic similarity between the corresponding business action description phrases, clustering is performed, the aggregation density and resource consumption proportion of each group are calculated, the effective group is screened, and the power grid tertiary business unit is generated; The bottom layer decomposition module: for each power grid tertiary business unit, based on the semantic similarity between the corresponding business action description phrases and the phrase execution frequency proportion, phrase screening is performed, and the bottom layer business unit is generated.

[0015] According to a third aspect of the present application, an electronic device is provided, comprising a memory and a processor, the memory stores a computer program, and the processor executes the program to realize the method.

[0016] According to a fourth aspect of the present application, a computer readable storage medium is provided, which stores a computer program, and the program is executed by a processor to realize the method.

[0017] Compared with the prior art, the present application has the following beneficial effects: (1) The present application is specially designed for hierarchical decomposition of power grid primary business, through progressive decomposition from top to bottom, a four-level business system with strict logic is generated, which directly interfaces with subsequent resource allocation, performance evaluation and other core management scenarios, and provides basis for them.

[0018] (2) The bottom layer business unit obtained by the present application directly corresponds to the non-divisible business action description phrase, and is completely bound with the actual business execution action, realizing seamless connection from business decomposition to actual operation, so that precise matching with technical standards can be realized, and basis support is provided for subsequent transmission and derivation of the contribution of the entire power grid primary business to the implementation benefit through the contribution layer of the technical standard in the bottom layer business.

[0019] (3) The present application makes automatic decision through semantic similarity, aggregation density and resource consumption proportion, all hierarchical divisions are based on objective business data without manual experience intervention, ensuring that the decomposition result is unified and reproducible, and avoiding subjective influence.

[0020] (4) The decomposition process of the present application is completely based on business original data (work order record, resource account, execution frequency, etc.), when the business changes, only the input data needs to be updated, and the decomposition result can be automatically regenerated without manual adjustment of the model or knowledge graph, realizing rapid adaptation to business changes, and greatly reducing the adjustment cost. BRIEF DESCRIPTION OF DRAWINGS

[0021] Figure 1 This is a flowchart of the method of the present invention; Figure 2 This is a schematic diagram of the decomposition results of the present invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0023] Unless otherwise defined, the technical or scientific terms used in this application shall have the ordinary meaning understood by one of ordinary skill in the art to which this application pertains. The terms “a,” “an,” “an,” “the,” and similar words used in this application do not indicate quantity limitation and may indicate singular or plural. The terms “comprising,” “including,” “having,” and any variations thereof used in this application are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or modules (units) is not limited to the listed steps or units, but may also include steps or units not listed, or may include other steps or units inherent to these processes, methods, products, or devices. The terms “connected,” “linked,” “coupled,” and similar words used in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. “Multiple” used in this application refers to two or more. “And / or” describes the relationship between related objects, indicating that three relationships may exist; for example, “A and / or B” can represent: A alone, A and B simultaneously, and B alone. The character " / " generally indicates that the preceding and following objects are in an "or" relationship. The terms "first," "second," and "third" used in this application are merely to distinguish similar objects and do not represent a specific ordering of the objects.

[0024] Example 1 This embodiment provides a method for decomposing the main business of a power grid, such as... Figure 1 As shown, the method includes the following steps: S1: Obtain power grid business data, extract business action description phrases, and count the frequency of phrase execution.

[0025] In this embodiment, the collected power grid business data includes, but is not limited to: unstructured text data, such as employee daily work reports, job description documents, cross-departmental collaboration logs, etc.; and structured business data, such as task work order system records (including task description, completion status, execution time), attendance and working hours system data, budget consumption ledgers, material requisition records, etc.

[0026] The acquired power grid business data is preprocessed before storage. For example, for text data, garbled characters and duplicate content are removed, and the format is unified as UTF-8; for business record data, canceled, rejected, and duplicate work orders are removed.

[0027] Based on the preprocessed business data, the text is split using Jieba word segmentation to extract candidate phrases with the structure of [specific action + business object] (such as "design 220 kV power grid structure", "review power access scheme", etc.). Based on the BERT model fine-tuned for the power grid industry, the candidate phrases are semantically filtered to remove ambiguous expressions and obtain business action description phrases.

[0028] Next, we will count the frequency of phrase usage, which includes the following steps: Within a preset statistical period (e.g., a natural month), the task description text in the work orders of the power grid business data is extracted, and the task description text is converted into a semantic vector to obtain the work order task description vector. The business action description phrases are converted into semantic vectors to obtain the business action description vectors; Calculate the semantic similarity between the work order task description vector and each business action description vector (in this embodiment, all semantic similarities are measured using cosine similarity). If the semantic similarity is greater than a preset threshold, the match is successful. Count the number of work orders that are successfully matched for each business action description phrase, and remove work orders with no working hours recorded in the matching results. Use the final number of work orders as the phrase execution frequency of the business action description phrase.

[0029] S2, based on the known primary business of the power grid, automatically assigns the extracted business action description phrases to the corresponding primary business of the power grid based on semantic similarity.

[0030] In this embodiment, the primary business of the power grid is known, including five primary businesses: power grid planning, engineering construction, dispatching and operation, operation and maintenance, and marketing services.

[0031] The five primary business names, described in text form, are converted into 768-dimensional semantic vectors. Then, the semantic similarity between the business action description vector and the corresponding semantic vector of each primary business name is calculated. The maximum value among the five semantic similarities calculated for each business action description vector is taken, and it is determined whether this maximum value is greater than a preset threshold of 0.6. If so, it is assigned to the primary business corresponding to the maximum value; if the maximum value is less than 0.6, it is marked as a phrase to be classified. After all classifications are completed, the average semantic similarity between all phrases to be classified and the already classified phrases under each primary business is calculated. The primary business corresponding to the maximum semantic similarity is taken as the final assignment result for the phrases to be classified.

[0032] S3, for each primary business of the power grid, clusters are performed based on the semantic similarity between the corresponding business action description phrases, calculates the aggregation density and resource consumption ratio of each group, selects effective groups, and generates secondary business units of the power grid.

[0033] This embodiment takes power grid planning as an example to illustrate in detail the process of decomposing and generating secondary business units of the power grid.

[0034] First, all descriptive phrases for business actions under the primary business of power grid planning are obtained. Hierarchical clustering is then used, with a semantic similarity of 0.7 or greater as the merging threshold, to cluster highly semantically related phrases into several groups. For example, "designing a 220 kV power grid structure" and "calculating main grid simulation data" have a semantic similarity of 0.82 and are clustered into group 1; "compiling a distribution network feasibility study report" and "optimizing the distribution network structure" have a semantic similarity of 0.78 and are clustered into group 2.

[0035] After clustering, the execution frequency of each phrase within a group is summed. The ratio of this sum to the total execution frequency of the corresponding phrases at the next higher level is then calculated to obtain the execution frequency percentage. For example, group 1, obtained from clustering under power grid planning, contains 5 business action description phrases. The sum of the execution frequencies of these 5 phrases is calculated and divided by the sum of the execution frequencies of all business action description phrases in power grid planning to obtain the execution frequency percentage for group 1.

[0036] Secondly, calculate the resource consumption ratio: Sum the total resource consumption within a specific group obtained from clustering at the current level, and calculate the ratio of this total to the total resource consumption at the corresponding level above. For example, if group 1, obtained from clustering under power grid planning, contains 5 business action description phrases, calculate the total resource consumption corresponding to these 5 phrases, and divide it by the total resource consumption of all business action description phrases in power grid planning to obtain the resource consumption ratio for group 1. The resource consumption corresponding to each phrase can be calculated during preprocessing using the work order ID, then matched to the corresponding phrase.

[0037] Therefore, the aggregation density can be calculated as follows: calculate the absolute value of the difference between the execution frequency ratio and the resource consumption ratio, subtract the absolute value from 1 to calculate the matching degree; calculate the weighted sum of the execution frequency ratio and the resource consumption ratio based on the preset weight, and divide by the matching degree to obtain the aggregation density.

[0038] When both the resource consumption ratio and aggregation density are greater than or equal to their respective thresholds, the group is determined to be valid, and other groups that do not meet the conditions are removed.

[0039] Each valid group corresponds to a secondary business unit of the power grid. Based on the selected valid groups, the high-frequency core keywords of each valid group are extracted using the TextRank keyword extraction algorithm. Combined with the power grid industry conventions, each valid group is named to obtain the name of the secondary business unit of the power grid.

[0040] S4. For each secondary business unit of the power grid, clustering is performed based on the semantic similarity between the corresponding business action description phrases. The aggregation density and resource consumption ratio of each group are calculated, and effective groups are selected to generate the tertiary business unit of the power grid.

[0041] The implementation process of this step is the same as that of step S3. The only difference is the level of the business action description phrase on which it is based and the threshold for filtering valid groups. In this step, the filtering threshold must be lower than the threshold in step S3.

[0042] S5, for each third-level business unit of the power grid, selects phrases based on the semantic similarity between the corresponding business action description phrases and the proportion of phrase execution frequency, and generates the underlying business unit.

[0043] In this step, for each third-level business unit of the power grid, the semantic similarity between the business action description phrases under it is calculated, and phrases with an average semantic similarity greater than a preset threshold are retained. Then, among the remaining phrases, phrases with an execution frequency ratio greater than a preset ratio are retained. Finally, each remaining business action description phrase corresponds to a bottom-level business unit.

[0044] The final decomposition result architecture is as follows Figure 2 As shown, each decomposed underlying business unit can be matched with technical standards to perform subsequent contribution calculations and business benefit assessments.

[0045] Tables 1 and 2 show examples of the decomposition results of power grid planning and engineering construction in the primary business.

[0046] Table 1 Table 2 As can be seen from Tables 1 and 2, the present invention can effectively decompose the primary business of the power grid at each level, and the decomposition results are reliable and practical.

[0047] Example 2 This embodiment, based on embodiment 1, provides a method for verifying the decomposition results after the business decomposition is completed, in order to verify the validity of the decomposition results. Specifically, it includes the following steps: Step 1) Hierarchical Attribution Verification: Each bottom-level business unit is uniquely traced back to its corresponding primary power grid business. The average semantic similarity between the business action description phrase in the bottom-level business unit and the primary power grid business is calculated. If the average semantic similarity is lower than the preset threshold, it indicates that the bottom-level business unit should not belong to the corresponding primary power grid business, and the verification fails. Step 2) Hierarchical Coverage Verification: Calculate the sum of the resource consumption ratios of the second-level business units and the third-level business units of the power grid. If it is less than the preset threshold, it indicates that during the subdivision process, many phrases failed to meet the threshold verification requirements and were not retained as valid groups. Therefore, the clustering results cannot accurately reflect the essence of the subdivision, and the verification fails. Step 3) No overlap verification: For the power grid secondary business unit and the power grid tertiary business unit, calculate the average semantic similarity of the business action description phrases between different business units in the current level. If the semantic similarity is greater than the preset threshold, it indicates that the similarity between different business units is too high and there is overlap. They may need to be merged, so the verification fails. Step 4) If the verification fails, adjust the clustering parameters or the threshold for filtering valid groups, and decompose again.

[0048] Example 3 This embodiment, based on Embodiment 1, provides a method for calculating the weight of each business unit obtained from the analytic hierarchy process (AHP). By assigning scores to each business unit pairwise, it obtains expert opinions on the importance of each unit. Based on this, through consistency assessment of expert scores and comprehensive calculation, the final weights of the business units are obtained, thereby evaluating the importance of each business unit to the overall benefits of the final implementation. The main steps are as follows: Step 1) Establish a hierarchical structure and hierarchize the evaluation indicators.

[0049] When applying Analytic Hierarchy Process (AHP) to analyze decision problems, the first step is to organize and hierarchically structure the problem, constructing a hierarchical model. In this model, complex problems are decomposed into constituent elements, which are then organized into several levels according to their attributes and relationships. Elements at higher levels act as criteria, governing the relevant elements at lower levels.

[0050] Step 2) Construct the comparison judgment matrix.

[0051] To quantify decision-making and form a numerical judgment matrix, a metric needs to be determined when comparing two factors, as shown in Table 3.

[0052] Table 3 Step 3) Solve for the eigenvectors.

[0053] Multiply and raise the nth root of each row of the numerical judgment matrix P: , in, This represents the element in row i and column j of P. The number of rows / columns of the matrix. As an intermediate variable; Calculate weights : , Summing the elements in each column of P: , Calculate characteristic quantities Value: .

[0054] Step 4) Consistency check.

[0055] Perform a consistency test. If CR = CI / RI ≤ 0.1, the consistency test is passed, where RI is the average random consistency index. .

[0056] Table 4 shows the weight generation results of the secondary business units under the power grid planning in one embodiment. The weight of the primary main business is evenly distributed, that is, the weight of each primary main business is 20%. The weights of the other business units can also be generated one by one by referring to the above process. This embodiment will not be described in detail here.

[0057] Table 4 Example 4 This embodiment provides a power grid main business decomposition device for implementing the method described in Embodiment 1. The device includes: Data acquisition and phrase extraction module: used to acquire power grid business data, extract business action description phrases, and count the frequency of phrase execution; First-level decomposition module: Based on the known first-level main business of the power grid, it automatically assigns the extracted business action description phrases to the corresponding first-level main business of the power grid based on semantic similarity; Second-level decomposition module: For each primary business of the power grid, clustering is performed based on the semantic similarity between the corresponding business action description phrases, the aggregation density and resource consumption ratio of each group are calculated, effective groups are selected, and secondary business units of the power grid are generated. The three-level decomposition module: For each power grid secondary business unit, clustering is performed based on the semantic similarity between the corresponding business action description phrases, the aggregation density and resource consumption ratio of each group are calculated, effective groups are selected, and power grid tertiary business units are generated. The underlying decomposition module is used to filter phrases based on the semantic similarity between the corresponding business action description phrases and the proportion of phrase execution frequency for each power grid level 3 business unit, and generate the underlying business unit.

[0058] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the described module can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0059] Example 5 The electronic device of this invention includes a central processing unit (CPU), which can perform various appropriate actions and processes according to computer program instructions stored in read-only memory (ROM) or loaded from a storage unit into random access memory (RAM). The RAM may also store various programs and data required for device operation. The CPU, ROM, and RAM are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.

[0060] Multiple components in the device are connected to the I / O interface, including: input units such as keyboards and mice; output units such as various types of displays and speakers; storage units such as disks and optical discs; and communication units such as network interface cards (NICs), modems, and wireless transceivers. The communication unit allows the device to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0061] The processing unit executes the various methods and processes described above, such as methods S1 to S5. For example, in some embodiments, methods S1 to S5 may be implemented as computer software programs tangibly contained in a machine-readable medium, such as a storage unit. In some embodiments, part or all of the computer program may be loaded and / or installed on the device via ROM and / or a communication unit. When the computer program is loaded into RAM and executed by the CPU, one or more steps of methods S1 to S5 described above may be performed. Alternatively, in other embodiments, the CPU may be configured to execute methods S1 to S5 by any other suitable means (e.g., by means of firmware).

[0062] The functions described above in this document can be performed, at least in part, by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: Field Programmable Gate Arrays (FPGAs), Application-Specific Integrated Circuits (ASICs), Application Standard Products (ASSPs), System-on-Chip (SoCs), Complex Programmable Logic Devices (CPLDs), and so on.

[0063] The program code used to implement the methods of the present invention can be written in any combination of one or more programming languages. This program code can be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code can be executed entirely on the machine, partially on the machine, as a standalone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0064] In the context of this invention, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0065] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for decomposing the main business of a power grid, characterized in that, The method includes the following steps: Acquire power grid business data, extract business action description phrases, and count the frequency of phrase execution; Based on the known primary business of the power grid, the extracted business action description phrases are automatically assigned to the corresponding primary business of the power grid based on semantic similarity; For each primary business of the power grid, clustering is performed based on the semantic similarity between the corresponding business action description phrases. The aggregation density and resource consumption ratio of each group are calculated, and effective groups are selected to generate secondary business units of the power grid. For each secondary business unit of the power grid, clustering is performed based on the semantic similarity between the corresponding business action description phrases. The aggregation density and resource consumption ratio of each group are calculated, and effective groups are selected to generate the tertiary business unit of the power grid. For each third-level business unit of the power grid, phrases are filtered based on the semantic similarity between the corresponding business action description phrases and the proportion of phrase execution frequency to generate the underlying business unit.

2. The method for decomposing the main business of a power grid according to claim 1, characterized in that, The method for calculating the frequency of execution of the phrase is as follows: Within a preset statistical period, the task description text in the work orders of power grid business data is extracted, and the task description text is converted into a semantic vector to obtain the work order task description vector. The business action description phrases are converted into semantic vectors to obtain the business action description vectors; Calculate the semantic similarity between the work order task description vector and each business action description vector. If the semantic similarity is greater than a preset threshold, the match is successful. Count the number of work orders that are successfully matched for each business action description phrase, and remove work orders with no working hours recorded in the matching results. The final number of work orders is taken as the phrase execution frequency of the business action description phrase.

3. The method for decomposing the main business of a power grid according to claim 1, characterized in that, The method for calculating the resource consumption ratio is as follows: calculate the total resource consumption within a certain group obtained by clustering at the current level, and calculate the ratio of the total resource consumption within the group to the total resource consumption at the corresponding previous level to obtain the resource consumption ratio.

4. The method for decomposing the main business of a power grid according to claim 1, characterized in that, The method for calculating the aggregation density is as follows: sum the execution frequencies of phrases within a certain group obtained by clustering at the current level, calculate the ratio of the sum of execution frequencies of phrases within the group to the total execution frequency of the corresponding phrases at the previous level, and obtain the execution frequency ratio. The absolute value of the difference between the execution frequency ratio and the resource consumption ratio is calculated, and the matching degree is obtained by subtracting the absolute value from 1. The aggregation density is obtained by calculating the weighted sum of the execution frequency ratio and the resource consumption ratio based on the preset weights and dividing by the matching degree.

5. The method for decomposing the main business of a power grid according to claim 1, characterized in that, The method for generating the power grid secondary business unit or the power grid tertiary business unit is as follows: each valid group corresponds to a power grid secondary business unit or a power grid tertiary business unit. Based on the selected valid groups, the high-frequency core keywords of each valid group are extracted using the TextRank keyword extraction algorithm. The names of each valid group are then named in accordance with power grid industry conventions to obtain the names of the power grid secondary business unit or the power grid tertiary business unit.

6. The method for decomposing the main business of a power grid according to claim 1, characterized in that, The method further includes: verifying the decomposition results after the business decomposition is completed. Hierarchical attribution verification: Each underlying business unit is uniquely traced back to its corresponding primary power grid business. The semantic similarity between the business action description phrase in the underlying business unit and the primary power grid business is calculated. If the semantic similarity is lower than a preset threshold, the verification fails. Hierarchical coverage verification: Calculate the sum of the resource consumption ratios of the second-level business units and the third-level business units of the power grid. If the sum is less than the preset threshold, the verification fails. Non-overlap verification: For power grid level 2 business units and power grid level 3 business units, calculate the average semantic similarity of the business action description phrases between different business units within the current level. If the semantic similarity is greater than the preset threshold, the verification fails. If the validation fails, adjust the clustering parameters or the screening threshold, and perform the decomposition again.

7. The method for decomposing the main business of a power grid according to claim 1, characterized in that, The method further includes: using the analytic hierarchy process (AHP) to calculate the weight of each business unit obtained from the decomposition.

8. A device for decomposing the main business of a power grid, characterized in that, For implementing the method as described in any one of claims 1 to 7, the apparatus comprises: Data acquisition and phrase extraction module: used to acquire power grid business data, extract business action description phrases, and count the frequency of phrase execution; First-level decomposition module: Based on the known first-level main business of the power grid, it automatically assigns the extracted business action description phrases to the corresponding first-level main business of the power grid based on semantic similarity; Second-level decomposition module: For each primary business of the power grid, clustering is performed based on the semantic similarity between the corresponding business action description phrases, the aggregation density and resource consumption ratio of each group are calculated, effective groups are selected, and secondary business units of the power grid are generated. The three-level decomposition module: For each power grid secondary business unit, clustering is performed based on the semantic similarity between the corresponding business action description phrases, the aggregation density and resource consumption ratio of each group are calculated, effective groups are selected, and power grid tertiary business units are generated. The underlying decomposition module is used to filter phrases based on the semantic similarity between the corresponding business action description phrases and the proportion of phrase execution frequency for each power grid level 3 business unit, and generate the underlying business unit.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the program, it implements the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1 to 7.

Images