Substation planning and site selection method based on artificial intelligence

Abstract

The invention provides a transformer substation planning and site selection method based on artificial intelligence. The method comprises the following steps: (1) designing a real-time topographic feature recognition model and deeply researching an activation function and an optimization strategy; a1, setting a plurality of specific scenes; obtaining a real-time topographic feature recognition model; a2, storing an activation function; a3, establishing an optimization strategy for reducing the complexity and calculation amount of the model; a4, evaluating the performance of the real-time topographic feature recognition model; (2) formulating a diversified training strategy and a data enhancement scheme; (3) performing data processing in combination with unmanned aerial vehicle data, and extracting landform and ground feature category features in a complex terrain environment; (4) identifying and analyzing unknown landform data through the real-time topographic feature identification model; (5) constructing an adaptive scoring model; and (6) providing a substation site selection strategy according to the landform data and the adaptive scoring model system. The invention provides an efficient substation planning and site selection method by using an artificial intelligence technology.

Description

Technical Field

[0001] The present invention relates to the technical field of substation planning and site selection, and in particular to a substation planning and site selection method based on artificial intelligence. Background Art

[0002] At present, the breakthrough progress of drone and artificial intelligence technology has brought a new era to the planning and construction of power transmission and transformation projects. The application of these advanced technologies has not only greatly improved the efficiency of the project, but also opened up new possibilities in terms of safety, precision and economy. They provide unprecedented opportunities for the development of the power industry, especially in improving the accuracy of project planning, accelerating construction progress, and reducing operating costs. In addition, the integrated application of these technologies will also promote the development of the power industry in a more intelligent and automated direction, and provide strong technical support for the efficient and sustainable operation of the power system. Combining drone technology with artificial intelligence, by analyzing multiple data sources, including earthwork volume, distance to the station access road, and the relationship between the proposed site and its surrounding environment, such as resident distribution, agricultural and animal husbandry activities, etc., it is possible to more accurately predict the amount of demolition in the early stage and other related compensation amounts. The image data taken by drones, combined with advanced artificial intelligence algorithms, can quickly analyze key parameters such as line conditions, difficulty, corridors and channels, so as to more deeply evaluate the advantages and disadvantages of each potential site and provide a quantitative analysis basis for site selection.

[0003] However, the current site selection of substations still relies on traditional intelligent algorithms and evaluation systems, and the identification of terrain and landforms in remote areas needs to be improved. At the same time, in the process of planning and optimizing the transmission line route plan, there are often situations such as densely populated houses, numerous crossings, and complex geological conditions. It takes a lot of time to count the types and quantities of trees cut, houses demolished, and crossings. Traditional methods require manual field measurements, or manual classification and measurement of land features based on drone aerial survey images. There are problems such as large workload and low accuracy, which greatly affects the efficiency of line engineering design and engineering quantity accounting.

[0004] Therefore, there is an urgent need to find a more efficient, comprehensive and intelligent site selection solution, improve the accuracy and real-time performance of terrain data, explore efficient integration solutions of drone technology and artificial intelligence recognition algorithms, and adopt an artificial intelligence-based substation planning and site selection method to improve the efficiency and accuracy of site selection. Summary of the invention

[0005] In view of the shortcomings of the prior art, the purpose of the present invention is to provide a substation planning and site selection method based on artificial intelligence, which utilizes artificial intelligence technology to provide an efficient substation planning and site selection method, improve the accuracy and real-time performance of terrain data, and improve the efficiency and accuracy of power grid planning and site selection.

[0006] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0007] The present invention provides a substation planning and site selection method based on artificial intelligence, comprising the following steps:

[0008] (1) Design a real-time terrain feature recognition model and conduct in-depth research on activation functions and optimization strategies; specifically, the following steps are included:

[0009] A1. Set up several specific scenarios; establish a multi-level feature extraction and abstraction method for terrain images, and obtain a real-time terrain feature recognition model;

[0010] A2. Store activation function; Adaptive activation function, specifically: dynamically adjust the activation function according to specific scenarios;

[0011] A3. Establish optimization strategies to reduce model complexity and computational effort;

[0012] A4. Evaluate the performance of the real-time terrain feature recognition model and achieve the preset requirements;

[0013] (2) Develop diversified training strategies and data enhancement schemes;

[0014] (3) Extracting landform and object category features in complex terrain environments after combining drone data with data processing;

[0015] (4) Identify and analyze unknown landform data through real-time terrain feature recognition models;

[0016] (5) Construct an adaptive scoring model;

[0017] (6) Provide substation site selection strategy based on geomorphic data and adaptive scoring model system.

[0018] Furthermore, in step A1, the specific method of obtaining the real-time terrain feature recognition model includes the following steps:

[0019] Acquire high-resolution image data through drones, and perform correction, denoising and image enhancement on the image data;

[0020] Artificial intelligence algorithms identify and classify surface features while providing topographic information;

[0021] The data is classified and segmented to further extract landform features, and feature detection engineering and image segmentation technology are used to achieve accurate identification of land objects.

[0022] Furthermore, the optimization strategy of A3 in step (1) includes using an optimizer to adjust the learning rate and regularization parameters to ensure that the model converges quickly during training and does not fall into a local optimum.

[0023] Furthermore, the optimization strategy also includes pruning the model, quantizing parameters and reducing model parameters.

[0024] Furthermore, the step (2) specifically includes the following steps:

[0025] Use a variety of training strategies to build multiple models or learn different model weights based on step (1);

[0026] Use data augmentation methods to expand the dataset during training, including geometric transformation, color transformation, and noise injection of training samples;

[0027] Use these trained models and expanded datasets to train models;

[0028] Through experimental verification and performance evaluation, the model was selected as the real-time terrain feature recognition model after confirming the effectiveness of diversified training strategies and data enhancement schemes in improving model stability and generalization performance.

[0029] Furthermore, the geometric transformation includes rotation and scaling.

[0030] Furthermore, the experimental verification is to implement a pilot project on the site selection of no less than 2 substations with voltage levels of 110kV, 220kV or 500kV, obtain drone data and input it into the model verification, and select the model that meets the standards.

[0031] Furthermore, the experimental verification is specifically carried out through cross-validation.

[0032] Furthermore, the method for establishing the adaptive scoring model is as follows: formulating decision factors for substation site selection, establishing a multi-index evaluation model, and constructing a scoring system to score each potential site selection location; the potential site selection location is a number of substation site selection addresses obtained after analyzing geomorphic data;

[0033] Analyze and mine a large amount of historical data to determine various scoring criteria. The decision-making factors include topography, land use policy, electricity demand and supply, and ecological environment, and establish a correlation model between decision-making factors;

[0034] Clarify the target layer, criterion layer and basic layer of the substation site selection problem, and establish initial weights with the help of expert experience and domain knowledge; use artificial intelligence algorithms to enable the adaptive scoring model to continuously iterate and learn, and dynamically adjust the scoring criteria and weights based on real-time data and feedback information.

[0035] Compared with the prior art, the present invention has the following beneficial effects:

[0036] (1) Since high-accuracy identification of complex geomorphic features directly affects the decision-making of substation site selection, the present invention designs a new artificial intelligence model and conducts in-depth research on connection modes, activation functions and optimization strategies to ensure that the model has real-time and high accuracy, and then adopts a variety of training strategies and data enhancement methods to improve the stability and generalization performance of the model. In addition, drone data is combined to explore the extraction of geomorphic and land feature categories and identify unknown geomorphic data, which can meet the needs of fast and accurate identification in complex terrain environments and provide more comprehensive and accurate geographic information support for substation site selection.

[0037] (2) The present invention uses artificial intelligence algorithms to identify and classify surface features, such as buildings, vegetation, water areas, etc., and provides terrain information. Secondly, the data is classified and segmented to further extract geomorphic features, and feature detection engineering and image segmentation technology are used to accurately identify objects. In order to achieve rapid identification, it is necessary to combine real-time processing technology to improve the recognition speed of landforms and object categories by optimizing algorithms and models. Finally, a model is established and verified, cross-validation and performance evaluation are performed, providing reliable technical support for solving the problem of extracting landform and object category features in complex terrain environments.

[0038] (3) Traditional scoring models lack adaptability to complex substation site selection conditions and are difficult to effectively respond to the needs of different regional environments. Therefore, new technical means and methods are needed to overcome these limitations and improve the scientificity and accuracy of the evaluation. By deeply exploring the decision-making factors of substation site selection, using advanced weight allocation technology, and combining artificial intelligence technology to carefully consider the scientificity of scoring criteria and weight settings, an adaptive scoring model is constructed. At the same time, the differences in site selection conditions in cities and suburbs are analyzed to ensure that the evaluation system is suitable for site selection needs in different regional environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0040] Figure 1 The following is a flowchart of an embodiment of the present invention. DETAILED DESCRIPTION

[0041] The embodiments of the present disclosure are described in detail below with reference to the accompanying drawings.

[0042] The following describes the embodiments of the present invention through specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present disclosure.

[0043] like Figure 1 As shown, an embodiment of the present invention provides a substation planning and site selection method based on artificial intelligence, comprising the following steps:

[0044] (1) Design a real-time terrain feature recognition model and conduct in-depth research on activation functions and optimization strategies; specifically, the following steps are included:

[0045] A1. Set several specific scenarios; establish a multi-level feature extraction and abstraction method for terrain images, and obtain a real-time terrain feature recognition model; the specific method for obtaining the real-time terrain feature recognition model includes the following steps:

[0046] Acquire high-resolution image data through drones, and perform correction, denoising and image enhancement on the image data;

[0047] Artificial intelligence algorithms identify and classify surface features while providing topographic information;

[0048] The data is classified and segmented to further extract landform features, and feature detection engineering and image segmentation technology are used to achieve accurate identification of land objects.

[0049] A2. Store activation function; Adaptive activation function, specifically: dynamically adjust the activation function according to specific scenarios;

[0050] A3. Establish optimization strategies to reduce model complexity and computational complexity; optimization strategies include using optimizers, adjusting learning rates and regularization parameters, and ensuring that the model converges quickly during training and does not fall into local optimality. Optimization strategies also include pruning models, quantizing parameters, and reducing model parameters.

[0051] A4. Evaluate the performance of the real-time terrain feature recognition model and achieve the preset requirements;

[0052] (2) Formulate diversified training strategies and data enhancement schemes; Step (2) specifically includes the following steps:

[0053] Use a variety of training strategies to build multiple models or learn different model weights based on step (1);

[0054] The data augmentation method is used to expand the data set during the training process, including geometric transformation, color transformation, and noise injection of training samples; geometric transformation includes rotation and scaling.

[0055] Use these trained models and expanded datasets to train models;

[0056] Through experimental verification and performance evaluation, the model was selected as the real-time terrain feature recognition model after confirming the effectiveness of diversified training strategies and data enhancement schemes in improving model stability and generalization performance. The experimental verification was specifically carried out through cross-validation. The experimental verification was to conduct a pilot project on the site selection of no less than 2 substations with voltage levels of 110kV, 220kV or 500kV, obtain drone data, and input it into the model for verification, and select a model that meets the standards.

[0057] (3) Extracting landform and object category features in complex terrain environments after combining drone data with data processing;

[0058] (4) Identify and analyze unknown landform data through real-time terrain feature recognition models;

[0059] (5) constructing an adaptive scoring model; the method for establishing the adaptive scoring model is: formulating decision factors for substation site selection, establishing a multi-index evaluation model, and constructing a scoring system to score each potential site location; the substation site is the substation site address obtained after analyzing the geomorphic data;

[0060] Through the analysis and mining of a large amount of historical data, various scoring criteria are determined. The decision-making factors include factors such as topography, land use policy, power demand and supply, and ecological environment, and a correlation model between decision-making factors is established. The target layer, criterion layer, and basic layer of the substation site selection problem are clarified, and the initial weights are established with the help of expert experience and domain knowledge to ensure the scientificity and comprehensiveness of the scoring criteria. Artificial intelligence algorithms are used to enable the adaptive scoring model to continuously iterate and learn, and dynamically adjust the scoring criteria and weights based on real-time data and feedback information, so that the scoring system is closer to the actual situation and improves prediction accuracy and adaptability.

[0061] To verify the effectiveness of the adaptive scoring model, we compared it with the review results of experienced experts. Through large-scale data verification, we analyzed the performance of the model in different scenarios, evaluated the advantages of the adaptive model, and continuously optimized and improved the accuracy and stability of the model.

[0062] (6) Provide substation site selection strategy based on geomorphic data and adaptive scoring model system.

[0063] The above is a detailed introduction to a substation planning and site selection method based on artificial intelligence provided by the present invention. This article uses specific examples to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A substation planning and site selection method based on artificial intelligence, characterized in that: The following steps are involved: (1) Design a real-time terrain feature recognition model and conduct in-depth research on activation functions and optimization strategies; The specific steps include: A1. Set up several specific scenarios; establish a multi-level feature extraction and abstraction method for terrain images, and obtain a real-time terrain feature recognition model; A2. Store activation function; Adaptive activation function, specifically: dynamically adjust the activation function according to specific scenarios; A3. Establish optimization strategies to reduce model complexity and computational effort; A4. Evaluate the performance of the real-time terrain feature recognition model and achieve the preset requirements; (2) Develop diversified training strategies and data enhancement schemes; (3) Extracting landform and object category features in complex terrain environments after combining drone data with data processing; (4) Identify and analyze unknown landform data through real-time terrain feature recognition models; (5) Construct an adaptive scoring model; (6) Provide substation site selection strategy based on geomorphic data and adaptive scoring model system.

2. The method for substation planning and site selection based on artificial intelligence according to claim 1 is characterized in that: In step A1, the specific method of obtaining the real-time terrain feature recognition model includes the following steps: Acquire high-resolution image data through drones, and perform correction, denoising and image enhancement on the image data; Artificial intelligence algorithms identify and classify surface features while providing topographic information; The data is classified and segmented to further extract landform features, and feature detection engineering and image segmentation technology are used to achieve accurate identification of land objects.

3. The method for substation planning and site selection based on artificial intelligence according to claim 1 is characterized in that: The optimization strategy of A3 in step (1) includes using an optimizer to adjust the learning rate and regularization parameters to ensure that the model converges quickly during training and does not fall into a local optimum.

4. The method for substation planning and site selection based on artificial intelligence according to claim 1 is characterized in that: The optimization strategy also includes pruning the model, quantizing parameters, and reducing model parameters.

5. The method for substation planning and site selection based on artificial intelligence according to claim 1 is characterized in that: The step (2) specifically comprises the following steps: Use a variety of training strategies to build multiple models or learn different model weights based on step (1); Use data augmentation methods to expand the dataset during training, including geometric transformation, color transformation, and noise injection of training samples; Use these trained models and expanded datasets to train models; Through experimental verification and performance evaluation, the model was selected as the real-time terrain feature recognition model after confirming the effectiveness of diversified training strategies and data enhancement schemes in improving model stability and generalization performance.

6. The method for substation planning and site selection based on artificial intelligence according to claim 1, characterized in that: The geometric transformation includes rotation and scaling.

7. The method for substation planning and site selection based on artificial intelligence according to claim 1 is characterized in that: The experimental verification is to carry out a pilot project on the site selection of no less than 2 substations with voltage levels of 110kV, 220kV or 500kV, obtain drone data, input it into the model verification, and select a model that meets the standards.

8. The method for substation planning and site selection based on artificial intelligence according to claim 1, characterized in that: The experimental verification is specifically carried out through cross validation.

9. The method for substation planning and site selection based on artificial intelligence according to claim 1, characterized in that: The method for establishing the adaptive scoring model is as follows: formulating decision factors for substation site selection, establishing a multi-index evaluation model, and constructing a scoring system to score each potential site selection location; the potential site selection location is a number of substation site selection addresses obtained after analyzing geomorphic data; Analyze and mine a large amount of historical data to determine various scoring criteria. The decision-making factors include topography, land use policy, electricity demand and supply, and ecological environment, and establish a correlation model between decision-making factors; Clarify the target layer, criterion layer and foundation layer of the substation site selection problem, and establish initial weights with the help of expert experience and domain knowledge; The use of artificial intelligence algorithms enables the adaptive scoring model to continuously iterate and learn, and dynamically adjust the scoring criteria and weights based on real-time data and feedback information.

Images