A digital power grid project multi-dimensional precise fitting system and application thereof

Abstract

The application discloses a kind of digital power grid project multi-dimensional accurate fitting system and application thereof, and its core development item is in multi-source electric construction data stream through and its multiple sub-item data collaborative optimization;Among them, multi-source electric construction data stream is based on the construction of multidimensional data model, and the collaborative optimization of multiple sub-item data corresponds to the algorithmic lateral parameter space with the corresponding data configuration of multidimensional data cube.The data model of the core development of the application and the data tools developed in front and back end provide a new data science tool and method for the digital upgrading of power grid construction associated projects, with basic, innovative and important practical value.

Description

Technical Field

[0001] This invention relates to the field of power grid data technology, and in particular to a multi-dimensional accurate fitting system for digital power grid projects and its application. Background Technology

[0002] Currently, the digital transformation of power grid construction projects is becoming an important trend in the industry and a crucial component of smart grid and digital grid construction. During this transformation, how to efficiently and accurately process and analyze large amounts of multi-source power grid construction data has become a pressing technical challenge.

[0003] Existing technologies have achieved a certain degree of digitalization and intelligence in power grid project management. Examples include using project management software to track construction progress, using geographic information systems (GIS) to manage the geographical information of the power grid, and using SCADA systems to monitor the operating status of power grid equipment. However, these technologies generally rely on traditional data processing methods and simple data models, such as form-based data and related data flows in project management software. These methods often fail to handle the accuracy and efficiency of data analysis when dealing with complex power grid construction project data. For instance, considering the complexity of power grid construction projects and the multi-source and diverse nature of data, a single data model is insufficient to effectively process multi-source data, leading to low efficiency in data integration and analysis. Furthermore, existing technologies lack effective collaborative optimization mechanisms during data processing, failing to achieve efficient collaborative processing of diverse sub-items of data, thus affecting the comprehensiveness and depth of data processing. Moreover, due to the lack of efficient data models and algorithms, existing technologies are often inefficient in analyzing and processing power grid construction project data, failing to meet the requirements of power grid digital transformation in the field of power construction projects. This invention addresses these technological shortcomings. Summary of the Invention

[0004] The technical problem to be solved by this invention is to address the various shortcomings of the existing technology by providing a multi-dimensional accurate fitting system for digital power grid projects and its application. Its core development item lies in the construction of a multi-dimensional data model and its algorithm embedded in the side parameter space, thereby enabling the interconnection of multi-source power construction data flow and the collaborative optimization of multi-sub-item data.

[0005] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows.

[0006] The Digital Power Grid Project Multidimensional Precision Fitting System is a data-driven system designed for the digital and intelligent transformation of power construction projects. Its core development focuses on the integration of multi-source power construction data flows and the collaborative optimization of diverse sub-item data. The integration of multi-source power construction data flows is based on the construction of a multidimensional data model. This model, built upon the collection and integration of multi-source data from power grid construction projects, constructs a multidimensional data cube with global and redundant data dimensions for the integration of multi-source power construction data flows. Furthermore, for the collaborative optimization of diverse sub-item data, a side parameter space with a corresponding data configuration to the multidimensional data cube is constructed, and algorithms are embedded within this parameter space with a focus on improving data efficiency.

[0007] As a preferred embodiment of the present invention, the leading data process of the core development item includes:

[0008] A. For the aforementioned multidimensional data "cube," a multidimensional data model is used to represent and describe the complex data structure of power grid construction projects. The cube is specifically constructed as an orthogonal data structure in a high-dimensional data space. The "cube" in its name does not necessarily correspond to a three-dimensional structure, but rather to a predefined n-dimensional data structure. Specifically, the multidimensional data cube is configured to contain n=5+m dimensions: namely, spatial dimension, time dimension, cost dimension, quality dimension, resource dimension, and other dimensions 1, 2, 3, ... The data content of the first five dimensions is as follows: the spatial dimension constructs and fills data according to the geographical location and construction area division of the power grid construction project; the time dimension constructs data according to the timeline, milestones, and construction progress of the power grid construction project. The data cube is constructed and populated based on the following dimensions: Cost dimension: data is constructed and populated based on the project's budget, actual expenditures, and cost analysis; Quality dimension: data is constructed and populated based on the project's quality standards, testing results, and quality control measures; Resource dimension: data is constructed and populated based on the project's human resources, material resources, and equipment resources; Other dimensions are initially set as blank dimensions to accommodate additional data analysis dimensions added as needed, thus adding sufficient redundant data dimensions on top of the global data dimensions. This multidimensional data cube can directly correspond to an n-order tensor in terms of data paradigm. A specific data implementation path instance is constructed corresponding to the above data architecture, specifically: an n-order tensor (T_proj ∈ R^(d_1 × d_2 × ... × d_n)) is used to represent the multidimensional data cube connecting multiple power construction data streams; each dimension (d_i) of the tensor data corresponds to a specific aspect of the power construction data stream, including the aforementioned time dimension, spatial dimension, cost dimension, quality dimension, resource dimension, and m other redundant dimensions; n = 5 + m.

[0009] B. Further, for the side parameter space and its algorithm embedding, firstly, a set of side parameter spaces with a corresponding data configuration to the multidimensional data cube is constructed, and its data implementation path is the side parameter space database (D_range); specifically, for the multidimensional data cube and its data paradigm instance n-order tensor (T_proj ∈R^(d_1 × d_2 × ... × d_n)), each component (T_(i_1, i_2, ..., i_n)) represents the specific value at the intersection of these dimensions. Then, the parameter space database (D_range) is constructed to store the adjustable numerical range of each component of the tensor ([a_(i_1, i_2, ..., i_n), b_(i_1, i_2,..., ..., ...)) according to the data configuration of the multidimensional data cube. i_n)]);(The adjustable numerical range data source is based on the standardized data of power construction projects, the expert knowledge data or historical data of power construction projects, the project establishment, demand or acceptance standard data of power construction projects, etc., and is preset); Therefore, the construction of the parameter space database (D_range) enables the generation of a multidimensional data cube group, that is, a set of multidimensional data cubes, through the data calling program, thus forming the underlying data group foundation for multidimensional fitting and optimization of digital power grid projects. On the basis of this underlying multidimensional data cube group, according to the specific power construction data processing task requirements, such as reducing costs, improving quality, optimizing resource allocation or accelerating construction progress, the algorithm is constructed and embedded in the parameter space to screen and compare all or part of the possible multidimensional data cube tensors in the multidimensional data cube group, that is, the set of multidimensional data cubes, to obtain the optimal multidimensional data cube corresponding to the current specific power construction data processing task.

[0010] As a preferred technical solution of the present invention, for any specific task of power construction data processing, the algorithm is optimized and embedded based on the parameter space. The basic algorithm is constructed as follows: for each component of the multidimensional data cube tensor, an arbitrary value is retrieved from the numerical range given by the adjacent callable parameter space database to generate many different specific multidimensional data cube tensors. Each multidimensional data cube tensor corresponds to a possible actual execution data model of the power grid construction project. Then, the numerical performance of all possible multidimensional data cubes or n-order tensors in the current data processing requirements is compared, the optimal multidimensional data cube tensor is found, and its corresponding actual execution data model of the power grid construction project is returned.

[0011] As a preferred technical solution of the present invention, based on the aforementioned basic algorithm, a specific data implementation path framework example is constructed: First, each component of (T_proj) is filled with numerical values ​​from (D_range) to generate a series of specific multidimensional data cubes or n-order tensors; each specific tensor (T_proj^(k)) corresponds to a possible actual execution of a power grid construction project; then, an optimization objective function (f_opt(T_proj^(k))) is defined, which evaluates the data and numerical representation of the specific tensor in meeting data processing requirements such as reducing costs, improving quality, optimizing resource allocation, and accelerating construction progress; then, the optimal solution (T_proj^{}) is found among all possible (T_proj^(k)) such that (f_opt(T_proj^{})) takes the maximum or minimum value; the maximum or minimum value depends on the specific optimization objective.

[0012] As a preferred embodiment of the present invention, in the basic algorithm, a global comparison algorithm is used to obtain the optimal solution.

[0013] As a preferred technical solution of the present invention, in the basic algorithm, the following cross-optimization algorithm is further constructed on the basis of the global comparison algorithm to obtain the optimal solution: First, randomly or sequentially select values ​​from (D_range) to fill (T_proj), generating an initial set of (T_proj^(k)). Then, for each (T_proj^(k)), construct (f_opt(T_proj^(k))) to characterize its fitness relative to the data target. Then, select several (T_proj^(k)) with better performance according to the fitness, randomly select two (T_proj^(k)) from them, and perform random or sequential cross-interaction on their components to generate a new (T_proj^(k)). On this basis, and optionally, randomly select the components of (T_proj^(k)) and retrieve new values ​​from (D_range) to replace them to increase diversity. Then, iterative optimization is performed, that is, the above process is repeated until the fitness improvement is less than a pre-set threshold. Correspondingly, the following algorithmic data process framework is constructed: T_proj^* = Optimize(T_proj, D_range, f_opt, select, crossover,mutate) ].

[0014] As a preferred technical solution of the present invention, considering the low computational efficiency caused by the basic algorithm directly exhaustively processing the entire parameter space (D_range), a new data processing algorithm is constructed as follows. The new algorithm generates tensors based on the combination number formula. First, it identifies several dimensions that have the greatest impact on power construction projects, such as no more than 5 dimensions. Then, it generates specific multidimensional data cubes and their corresponding tensors in the parameter space of these key dimension combinations. The identification of key dimension combinations includes the following data integration: through historical data analysis and / or theoretical derivation and / or manual or machine screening according to set data standards, a predetermined data paradigm is obtained to identify several dimensions that have the greatest impact on power grid construction projects for the current data analysis task, such as cost and time, resources and quality. Based on the identified key dimension combinations, the parameter space (D_range) is reduced, retaining the parameter range of the selected dimensions. On this basis, the combination number formula is applied. First, for combination number generation, the combination number formula (C(n, k) = n! / (k!(nk)!)) is used to calculate the number of possible combinations within a given parameter range for each key dimension combination, where (n) The number of discrete values ​​within the parameter range of this dimension is given by (k), where (k) is the number of parameters selected in the combination. Next, a series of specific multidimensional data cube tensors (T_proj^(k)) are generated based on the combination number formula. Each (T_proj^(k)) reflects a possible actual implementation of a power grid construction project. Then, a formalized optimization objective function (f_opt(T_proj^(k))) is defined to evaluate the performance of each specific tensor in meeting project requirements such as lowest cost, highest quality, and shortest time. Based on this, algorithm iteration is performed. By comparing the (f_opt) values ​​of all (T_proj^(k)), the optimal (T_proj^*) is selected. During the algorithm iteration process, the key dimension combinations and parameter space are adjusted as needed, iterating until the optimal tensor that satisfies the optimization objective is found. Correspondingly, the following algorithmic data process framework is constructed: [T_proj^* = Optimize(Combination(D_range^reduced), f_opt)] The efficiency of the algorithm is improved by using an optimization algorithm based on the combination formula and the combination of key dimensions and its embedding in the parameter space.

[0015] As a preferred technical solution of the present invention, the specific parameters of the optimization function can be adapted and adjusted according to the key objectives of the power grid construction project or the current data processing task; conventionally, they include cost optimization parameters, schedule acceleration parameters, quality improvement parameters, resource efficiency allocation parameters, and other parameters.

[0016] As a preferred technical solution of the present invention, the data development items of the fitting system also include diversified back-end data processing tools. Based on the multidimensional data "cube" data configuration, tensor analysis tools such as data task-oriented tensor decomposition tools are introduced to construct diversified digital optimization processes for power grid projects.

[0017] As a preferred technical solution of the present invention, the data development items of the fitting system also include front-end data processing and transformation, especially for unstructured data in the source data of power grid projects, performing feature processing and labeling processing corresponding to the data dimensions in the multidimensional data "cube".

[0018] The present invention also includes the application of the above-mentioned digital power grid project multidimensional accurate fitting system in power grid construction planning and optimization.

[0019] The beneficial effects of adopting the above technical solution are as follows: This invention develops a multi-dimensional accurate fitting system for digital power grid projects and its application, addressing the shortcomings of existing technologies. The core of this invention lies in the construction of a multi-dimensional data model and its embedded algorithm in a side parameter space, realizing the integration of multi-source power construction data streams and the collaborative optimization of multi-item data. Specifically, regarding the construction of the multi-dimensional data model: This research effectively integrates multi-source data from power grid construction projects by constructing a multi-dimensional data cube, including multiple dimensions such as space, time, cost, quality, and resources, as well as other data analysis dimensions added as needed, providing a comprehensive representation and description of the complex data structure of power grid construction projects. Regarding the side parameter space and algorithm embedding: This invention constructs a side parameter space with a corresponding data configuration to the multi-dimensional data cube and embeds algorithms in the parameter space to achieve collaborative optimization of multi-item data oriented towards improving data efficiency, thereby improving the accuracy and efficiency of data processing. Furthermore, this research also performs accurate fitting and optimization of the multi-dimensional data cube, effectively improving the efficiency of data analysis for power grid construction projects and providing strong technical support for the digital transformation of the power grid. Detailed Implementation

[0020] In the following description of embodiments, specific details such as particular system structures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. It should be understood that, as used in this specification and the appended claims, the term "comprising" indicates the presence of a described feature, integral, step, operation, element, and / or component, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or collections thereof. References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a particular feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification, do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof all mean "including but not limited to," unless otherwise specifically emphasized.

[0021] Example 1: Technical Framework of Multidimensional Precise Fitting System for Digital Power Grid Projects

[0022] The initial data processing for the multi-dimensional precision fitting system for digital power grid projects involves data collection and integration. This includes gathering various data from power grid construction projects, such as engineering design data, construction progress data, quality control data, cost management data, supply chain management data, environmental monitoring data, and equipment operation data. Based on our previous experience and preliminary research for this development, the data source format is categorized by data volume and accuracy. Multiple sources correspond to a larger data volume, but may reduce data quality and accuracy. The solution is to use several mainstream data sources while discarding small-scale, low-value, and non-standardized data sources, followed by appropriate data standardization and organization; that is, a technical solution combining mainstream multi-source data with data organization. Building upon data collection and organization, the multi-dimensional precision fitting system for digital power grid projects is data-driven, transforming power construction projects into digital and intelligent systems. It facilitates the integration of multi-source power construction data flows and the collaborative optimization of its diverse sub-items.

[0023] The core development item of building a multidimensional accurate fitting system for digital power grid projects lies in the construction of a multidimensional data model and its embedded side parameter space.

[0024] In addition to this core development item, related to the following are: Supporting Development ①, which is front-end supporting development; its core is the data format characterization processing of the power grid project source data corresponding to the multidimensional data model, and the labeling processing corresponding to the data dimensions in the multidimensional data "cube". Supporting Development ②, which is back-end supporting development; based on the multidimensional data model, diverse data processing tools such as data task-oriented tensor decomposition tools can be further developed to construct diversified digital optimization methods for power grid projects.

[0025] Example 2: Front-end Supporting Development

[0026] Before constructing the multidimensional data model and its multidimensional data "cube" for the power grid project, it is essential to process and transform the source data of the power grid project. In particular, unstructured data in the source data of the power grid project needs to be characterized, especially the data dimensions in the multidimensional data "cube" constructed according to the subsequent plan, which require corresponding dimension labeling. The technical development overview for this part is as follows.

[0027] Customized data processing is performed on unstructured data from power grid construction projects, such as engineering design documents, construction logs, and quality reports.

[0028] The basic components include: Text cleaning: In addition to basic cleaning steps, customized cleaning rules are developed for terms and symbols specific to power grid construction; the data paradigm is: [CleanedText_proj = f_clean_proj(RawText)]. Token segmentation and standardization: Token segmentation is performed using a professional vocabulary database for the power grid construction field, employing domain-specific word form restoration rules; the data paradigm is: [Tokens_proj = f_tokenize_proj(CleanedText_proj)]. Specific entity identification: Key entities in power grid construction projects are identified, such as equipment names and geographical locations; the data paradigm is: [Entities_proj = f_entity_proj(Tokens_proj)].

[0029] Then, we consider using advanced NLP techniques such as the Transformer model to extract features from the context of power grid construction projects; the data paradigm is: [Features_proj = f_transformer_proj(Entities_proj)]. Simultaneously, we use deep learning models, especially those pre-trained for power grid construction projects, to understand the meaning of text in specific contexts; the data paradigm is: [Context_proj = f_context_proj(Features_proj)]. Then, based on deep learning relation extraction techniques, we identify the relationships between entities in power grid construction projects, such as equipment and fault types, construction progress and influencing factors, etc.; the data paradigm is: [Relations_proj = f_relation_proj(Context_proj)].

[0030] It is worth noting that the goal of data processing in these steps is to support the construction of a multidimensional data cube. Therefore, it is necessary to ensure that the data processing process can identify and extract multiple dimensions of information related to the power grid construction project, including time, space, cost, quality, resources, etc., as well as data space reserved for any new analytical dimensions that may be added in the subsequent data processing of the power grid project.

[0031] Furthermore, on the one hand, in the data preprocessing and feature extraction stages, it is not only necessary to clean and standardize the data, but also to label the data to clarify which dimension each data point belongs to; for example, the date and timestamp in the construction log are labeled as the time dimension, the location information is labeled as the spatial dimension, the cost data is labeled as the cost dimension, and so on. On the other hand, when using the Transformer model and other models to extract features for each dimension, attention should be paid to fine-tuning the existing model in order to identify and extract features related to different dimensions.

[0032] Furthermore, considering the unstructured and complex nature of power grid project data, a genetic algorithm is introduced into this front-end development to improve the efficiency of identifying and processing unstructured data. Here, to adapt to the unstructured and complex nature of power grid project data, a novel atypical genetic algorithm is proposed.

[0033] The key point is to embed pre-organized data tables (which are pre-built to contain known entities and relationships in the power grid construction project). That is, in the initialization phase of the genetic algorithm, the pre-organized data tables are embedded into the algorithm as the initial population. Since these tables contain known entities and relationships in the power grid construction project, the efficiency of natural mutation of the algorithm is greatly improved.

[0034] The second key point is the deep integration of combination and permutation; that is, in the process of natural variation and heredity, mathematical knowledge of combination and permutation is used to generate new individuals; for example, when searching for the optimal structured representation, the algorithm will explore the combination of different entities and relationships and their permutation order in the multidimensional data cube, so as to find the representation method that best reflects the complexity of the power grid construction project.

[0035] The improved atypical genetic algorithm proposed here can not only optimize the structured representation of power grid project data, but also provide fundamental support for the construction of multidimensional data cubes. It also has important fundamental value for the realization of multidimensional accurate fitting systems for digital power grid projects.

[0036] Example 3, Core Development Item: Construction of the Embedded Side Parameter Space for Multidimensional Data Models and Algorithms

[0037] For the core development item, namely the construction of a multi-dimensional data model and its embedded side parameter space, the main technical approach is as follows: to conduct multi-source power construction data flow integration and collaborative optimization of multi-item data. The multi-source power construction data flow integration is based on a multi-dimensional data model. This model, built upon the collection and integration of multi-source data from power grid construction projects, constructs a multi-dimensional data cube with global data dimensions and redundant data dimensions for the integration of multi-source power construction data flows. For the collaborative optimization of multi-item data, a side parameter space with a corresponding data configuration to the multi-dimensional data cube is constructed, and algorithms are embedded within this parameter space with a focus on improving data efficiency.

[0038] The data process corresponding to the above-mentioned dominant technical route is as follows:

[0039] Step A: For the multidimensional data "cube," a multidimensional data model is used to represent and describe the complex data structure of power grid construction projects. The cube is specifically constructed as an orthogonal data structure in a high-dimensional data space. The "cube" in its name does not necessarily correspond to a three-dimensional structure, but rather to a predefined n-dimensional data structure. Specifically, the multidimensional data cube is set to contain n=5+m dimensions: namely, spatial dimension, time dimension, cost dimension, quality dimension, resource dimension, and other dimensions 1, 2, 3, ... The data content of the first five dimensions is as follows: the spatial dimension is constructed and filled with data according to the geographical location and construction area division of the power grid construction project; the time dimension is constructed according to the timeline and mileage of the power grid construction project. The data cube is constructed and populated based on the following dimensions: monument and construction progress; cost dimension based on the project's budget, actual expenditures, and cost analysis; quality dimension based on the project's quality standards, testing results, and quality control measures; resource dimension based on the project's human resources, material resources, and equipment resources; other dimensions are initially set as blank dimensions to accommodate additional data analysis dimensions added as needed, thus providing sufficient redundant data dimensions on top of the global data dimensions. This multidimensional data cube directly corresponds to an n-order tensor in terms of data paradigm. A specific data implementation path instance is constructed to correspond to the above data architecture, specifically:

[0040] A multidimensional data cube representing the interconnected multi-source power construction data flow is represented by an n-order tensor (T_proj ∈ R^(d_1 × d_2 × ... × d_n)); each dimension (d_i) of the tensor data corresponds to a specific aspect of the power construction data flow, including the aforementioned time dimension, spatial dimension, cost dimension, quality dimension, resource dimension, and m other redundant dimensions; n=5+m;

[0041] Step B

[0042] B-0. For the side parameter space and its algorithm embedding, firstly, a set of side parameter spaces with a corresponding data configuration to the multidimensional data cube is constructed. The data implementation path is the side parameter space database (D_range). Specifically, for the multidimensional data cube and its data paradigm instance n-order tensor (T_proj ∈ R^(d_1 ×d_2 × ... × d_n)) constructed above, each component (T_(i_1, i_2, ..., i_n)) represents the specific value at the intersection of these dimensions. Then, the parameter space database (D_range) is constructed to store the adjustable value range of each component of the tensor ([a_(i_1, i_2, ..., i_n), b_(i_1, i_2, ..., i_n)]) according to the data configuration of the multidimensional data cube. (The data source of the adjustable value range is based on the standardized data of the power construction project, the expert knowledge data or historical data of the power construction project, the project establishment, requirements or acceptance standard data of the power construction project, etc., and is pre-set).

[0043] Therefore, the construction of the parameter space database (D_range) enables the generation of a multidimensional data cube group, i.e., a set of multidimensional data cubes, through data calling programs. This constitutes the underlying data group foundation for multidimensional fitting and optimization of digital power grid projects. Based on this underlying multidimensional data cube group, according to the specific power construction data processing task requirements, such as reducing costs, improving quality, optimizing resource allocation, or accelerating construction progress, algorithms are constructed and embedded in the parameter space to filter and compare all or some possible multidimensional data cube tensors in the multidimensional data cube group, i.e., the set of multidimensional data cubes, to obtain the optimal multidimensional data cube corresponding to the current specific power construction data processing task.

[0044] B-1. For any specific task in power grid construction data processing, algorithm optimization and embedding are performed based on parameter space. The basic algorithm is as follows: For each component of the multidimensional data cube tensor, arbitrary values ​​are retrieved from the numerical range given in the adjacent callable parameter space database to generate many different specific multidimensional data cube tensors. Each multidimensional data cube tensor corresponds to a possible actual execution data model of a power grid construction project. Then, the numerical performance of all possible multidimensional data cubes or n-order tensors in the current data processing requirements is compared, the optimal multidimensional data cube tensor is found, and its corresponding actual execution data model of the power grid construction project is returned. Based on the above basic algorithm, a specific data implementation path framework example can be constructed: First, retrieve the numerical values ​​from (D_range) to fill each component of (T_proj), generating a series of specific multidimensional data cubes or n-order tensors; each specific tensor (T_proj^(k)) corresponds to a possible actual execution of a power grid construction project; then, define an optimization objective function (f_opt(T_proj^(k))), which evaluates the data and numerical representation of the specific tensor in meeting data processing requirements such as reducing costs, improving quality, optimizing resource allocation, and accelerating construction progress; then, find the optimal solution (T_proj^{}) among all possible (T_proj^(k)) such that (f_opt(T_proj^{})) reaches its maximum or minimum value; the maximum or minimum value depends on the specific optimization objective.

[0045] B-1-1. For obtaining the optimal solution, a global comparison algorithm can generally be used.

[0046] B-1-2. To improve data processing efficiency—especially when the data cube has a high dimensionality and a large data scale—improving data processing efficiency is crucial for the accurate fitting and optimization of power grid projects. Therefore, based on the global comparison algorithm, the following cross-optimization algorithm is further constructed:

[0047] First, randomly or sequentially select values ​​from (D_range) to fill (T_proj), generating an initial set of (T_proj^(k)). Then, for each (T_proj^(k)), construct (f_opt(T_proj^(k))) to characterize its fitness relative to the data target. Based on the fitness, select several well-performing (T_proj^(k)), randomly select two (T_proj^(k)) from them, and randomly or sequentially cross over their components to generate new (T_proj^(k)). Further, and optionally, randomly select components of (T_proj^(k)) and replace them with new values ​​from (D_range) to increase diversity. Then, iterative optimization is performed, repeating the above process until the fitness improvement is less than a pre-set threshold. Correspondingly, the following algorithmic data process framework is constructed: [T_proj^* =Optimize(T_proj, D_range, f_opt, select, crossover, mutate).

[0048] B-2. Furthermore, considering the low computational efficiency caused by the basic algorithm's exhaustive processing of the entire parameter space (D_range), a new data processing algorithm is constructed. This new algorithm generates tensors based on the combination number formula. First, it identifies several dimensions (e.g., no more than five) that have the greatest impact on power grid construction projects. Then, it generates specific multidimensional data cubes and their corresponding tensors in the parameter space of these key dimension combinations. The identification of key dimension combinations includes the following data integration: obtaining predetermined data paradigms through historical data analysis and / or theoretical derivation and / or manual or machine screening according to set data standards. This identifies several dimension combinations that have the greatest impact on power grid construction projects for the current data analysis task, such as cost and time, resources and quality. Based on the identified key dimension combinations, the parameter space (D_range) is reduced, retaining the parameter range of the selected dimensions. This guides the application of the combination number formula. First, for combination number generation, the combination number formula is used for each key dimension combination: (C(n, k) = n! / (k!(nk)!)), to calculate the possible number of combinations within a given parameter range, where (n) The number of discrete values ​​within the parameter range of this dimension is given by (k), which is the number of parameters selected in the combination. Next, a series of specific multidimensional data cube tensors (T_proj^(k)) are generated based on the combination formula, each (T_proj^(k)) reflecting a possible actual implementation of a power grid construction project. Then, a formalized optimization objective function (f_opt(T_proj^(k))) is set to evaluate the performance of each specific tensor in meeting project requirements such as lowest cost, highest quality, and shortest time. Based on this, algorithm iteration is performed, comparing the (f_opt) values ​​of all (T_proj^(k)) to select the optimal (T_proj^*). During the algorithm iteration process, the key dimension combinations and parameter space are adjusted as needed, iterating until the optimal tensor that satisfies the optimization objective is found. Correspondingly, the following algorithmic data process framework is constructed: [T_proj^* = Optimize(Combination(D_range^reduced), f_opt)];

[0049] This optimization algorithm, based on the combination formula and key dimension combination, and its embedding in the parameter space, improves the efficiency of the algorithm. The specific parameters of the objective function can be adapted and adjusted according to the key objectives of the power grid construction project or the current data processing task. Conventional parameters include cost optimization parameters, schedule acceleration parameters, quality improvement parameters, resource efficiency allocation parameters, and other parameters.

[0050] Example 4: Backend Support Development

[0051] Based on the constructed multidimensional data "cube" of power grid projects, data analysis tools are introduced to build a diversified digital optimization process for power grid projects. For example, tensor decomposition is performed based on specific power grid project tasks. The multidimensional data cube we construct corresponds to a tensor (T), where each dimension represents the space, time, cost, quality, and resources of the power grid project (the blank m dimensions are subsequently filled in). Then, based on this, CP decomposition or Tucker decomposition is introduced to decompose the tensor (T) into the product of multiple smaller tensors and matrices. This extracts the core features of power grid construction project data while reducing data complexity. Taking the implementation progress, construction and material spatial relationships, and cost data of a power grid project as an example, its data implementation framework can be characterized as follows: The tensor (T) is decomposed into factor matrices (A), (B), and (C), where (A), (B), and (C) represent the features of the spatial, temporal, and cost dimensions, respectively; the objective of tensor decomposition is to minimize the reconstruction error between the decomposed tensors: [ \min_{A,B,C} | T - [[A,B,C]] |_F^2 ]; where (| \cdot |_F) represents the Frobenius norm, and ([[A,B,C]]) represents the result of tensor decomposition.

[0052] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0053] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A multi-dimensional accurate fitting system for digital power grid projects, characterized by: This fitting system is data-driven, aiming to transform power construction projects into digital and intelligent systems. Its core development items include the integration of multi-source power construction data flows and the collaborative optimization of multi-item data. The integration of multi-source power construction data flows is based on a multi-dimensional data model. This multi-dimensional data model is constructed based on the collection and integration of multi-source data from power grid construction projects. It constructs a multi-dimensional data cube with global data dimensions and redundant data dimensions for the integration of multi-source power construction data flows. For the collaborative optimization of multi-item data, a side parameter space with a corresponding data configuration to the multi-dimensional data cube is constructed. The algorithm is embedded in the parameter space with the goal of improving data efficiency. The leading data processes include: A. For the aforementioned multidimensional data cube, it is adapted to the complex data structure of power grid construction projects and is represented and described using a multidimensional data model. The cube is specifically constructed as an orthogonal data structure in a high-dimensional data space. The "cube" in its name does not necessarily correspond to a three-dimensional structure, but rather to a predefined n-dimensional data structure. Specifically, the multidimensional data cube is configured to contain n=5+m dimensions: namely, spatial dimension, time dimension, cost dimension, quality dimension, resource dimension, and other dimensions 1, 2, 3, ..., m. The data content of the first five dimensions is as follows: the spatial dimension is constructed and filled with data based on the geographical location and construction area division of the power grid construction project; the time dimension is constructed and filled with data based on the timeline, milestones, and construction progress of the power grid construction project; the cost dimension is constructed and filled with data based on the budget, actual expenditures, and cost analysis of the power grid construction project. The process involves constructing and populating the data. For the quality dimension, data points are constructed and populated based on the power construction project's quality standards, testing results, and quality control measures. For the resource dimension, data points are constructed and populated based on the power construction's human resources, material resources, and equipment resources. Other dimensions are initially set as blank dimensions to accommodate subsequent data analysis dimensions added as needed, thus adding sufficient redundant data dimensions on top of the global data dimensions. This constructed multidimensional data cube directly corresponds to an n-order tensor in terms of data paradigm. A specific data implementation path instance is constructed corresponding to the above data architecture: an n-order tensor represents the multidimensional data cube connecting multiple power construction data flows. Each dimension of the tensor data corresponds to a specific aspect of the power construction data flow, including the aforementioned time dimension, spatial dimension, cost dimension, quality dimension, resource dimension, and m other redundant dimensions; n = 5 + m. B. For the flanking parameter space and its algorithm embedding, firstly, a set of flanking parameter spaces with a corresponding data configuration to the multidimensional data cube is constructed, and its data implementation path is the flanking parameter space database; specifically, for the multidimensional data cube and its data paradigm instance n-order tensor constructed above, each component represents the specific value at the intersection of these dimensions, and the parameter space database is constructed to store the adjustable value range of each component of the tensor according to the data configuration of the multidimensional data cube; the data source for the adjustable value range is based on the standardized data of power construction projects, expert knowledge data or historical data of power construction projects, power construction project initiation, requirements or verification. Standard data is pre-defined; thus, the construction of the parameter space database enables the generation of a multidimensional data cube group, i.e., a set of multidimensional data cubes, through the data calling program. This constitutes the underlying data group foundation for multidimensional fitting and optimization of digital power grid projects. Based on this underlying multidimensional data cube group, according to the specific power construction data processing task requirements, all or some possible multidimensional data cube tensors in the multidimensional data cube group, i.e., the set of multidimensional data cubes, are screened and compared by constructing and embedding algorithms in the parameter space to obtain the optimal multidimensional data cube corresponding to the current specific power construction data processing task.

2. The multi-dimensional accurate fitting system for digital power grid projects according to claim 1, characterized in that: For any specific task in power grid construction data processing, the algorithm is optimized and embedded based on the parameter space. The basic algorithm is as follows: for each component of the multidimensional data cube tensor, an arbitrary value is retrieved from the numerical range given in the adjacent callable parameter space database to generate many different specific multidimensional data cube tensors. Each multidimensional data cube tensor corresponds to a possible actual execution data model of the power grid construction project. Then, the numerical performance of all possible multidimensional data cubes or n-order tensors in the current data processing requirements is compared, the optimal multidimensional data cube tensor is found, and its corresponding actual execution data model of the power grid construction project is returned.

3. The multi-dimensional accurate fitting system for digital power grid projects according to claim 2, characterized in that: Based on the aforementioned fundamental algorithm, a specific data implementation path framework example is constructed: First, numerical values ​​are retrieved from the parameter space database to fill each component of the n-order tensor, generating a series of specific multidimensional data cubes or n-order tensors; each specific tensor corresponds to a possible actual execution of a power grid construction project; then, an optimization objective function is defined, which evaluates the numerical performance of the specific tensor in meeting data processing requirements; then, the optimal solution is found among all possible specific tensors, such that the optimization objective function reaches its maximum or minimum value; the maximum or minimum value depends on the specific optimization objective.

4. The multi-dimensional accurate fitting system for digital power grid projects according to claim 2, characterized in that: In the basic algorithm, a global comparison algorithm is used to obtain the optimal solution.

5. The multi-dimensional accurate fitting system for digital power grid projects according to claim 2, characterized in that: In the basic algorithm, the optimal solution is obtained by further constructing the following cross-optimization algorithm based on the global comparison algorithm: First, randomly or sequentially select values ​​from the parameter space database to fill the n-order tensor, generating an initial tensor set. Then, for each tensor in the set, construct an optimization function to characterize its fitness relative to the data target. Then, select the top few tensors with better performance based on fitness, randomly select two tensors from them, and randomly or sequentially cross their components to generate a new tensor as a candidate optimal tensor. Based on this, randomly select the components of the candidate optimal tensor and retrieve new values ​​from the parameter space database to replace them to increase diversity. Then, perform iterative optimization, that is, repeat the above process until the fitness improvement is less than a pre-set threshold.

6. The multi-dimensional accurate fitting system for digital power grid projects according to claim 1, characterized in that: A novel data processing algorithm is constructed, which generates tensors based on the combination number formula. First, it identifies several dimensional combinations that have the greatest impact on power grid construction projects. Then, it generates specific multidimensional data cubes and their corresponding tensors in the parameter space of these key dimensional combinations. The identification of key dimensional combinations involves the following data integration: obtaining predetermined data paradigms through historical data analysis and / or theoretical derivation and / or manual or machine screening according to set data standards. This identifies several dimensional combinations that have the greatest impact on power grid construction projects for the current data analysis task. Based on the identified key dimensional combinations, the parameter space is reduced, retaining the parameter range of the selected dimensions. This leads to the application of the combination number formula. First, for combination number generation, the combination number formula is used for each key dimensional combination to calculate the number of possible combinations within a given parameter range. The combination number formula is: C(n, k) = n! / (k!(nk)!), where n is the number of discrete values ​​within the parameter range of this dimension, and k is the number of parameters selected in the combination. Next, a series of specific multidimensional data cube tensors are generated based on the combination number formula. Each multidimensional data cube tensor reflects a possible actual implementation of a power grid construction project. Then, a formalized optimization function is defined to evaluate the performance of each multidimensional data cube tensor in meeting the requirements of the power grid construction project. Based on this, algorithm iteration is performed. By comparing the scores of all multidimensional data cube tensors under the optimization function, the optimal multidimensional data cube tensor is selected. During the algorithm iteration process, the key dimension combinations and parameter space are adjusted as needed, iteratively optimizing until the optimal tensor that satisfies the optimization objective is found. This optimization algorithm based on the combination number formula and key dimension combinations, and its embedding in the parameter space, improves algorithm efficiency. The specific parameters of the optimization function can be adapted and adjusted according to the key objectives of the power grid construction project or the current data processing tasks; these specific parameters include cost optimization parameters, schedule acceleration parameters, quality improvement parameters, and resource efficiency allocation parameters.

7. The multi-dimensional accurate fitting system for digital power grid projects according to claim 1, characterized in that: The data development items for this fitting system also include diversified backend data processing tools. Based on the multidimensional data cube data configuration, tensor analysis tools are introduced to construct a diversified digital optimization process for power grid projects.

8. The multi-dimensional accurate fitting system for digital power grid projects according to claim 1, characterized in that: The data development items of the fitting system also include front-end data processing and transformation, which involves characterization of unstructured data in the power grid project source data and labeling of the data dimensions corresponding to the data dimensions in the multidimensional data cube.

9. The application of the digital power grid project multidimensional accurate fitting system according to any one of claims 1-8 in power grid construction planning and optimization.