Electric power review opinion automatic generation method based on artificial intelligence

Abstract

The invention provides an electric power review opinion automatic generation method based on artificial intelligence, and relates to the field of electric power review opinion automatic generation. The electric power review opinion automatic generation method based on artificial intelligence comprises the steps that a, related documents, drawings and project data of electric power engineering are obtained through a data collection unit, data sources comprise local storage, a cloud database or an external management system, the obtained data are analyzed through an electric power engineering semantic analysis module, and an electric power engineering semantic analysis module is obtained; the context semantic information of the electric power engineering terms is analyzed by combining a bidirectional long-short term memory network and an attention mechanism, so that the understanding depth of the electric power engineering terms is improved. By automatically processing the electric power engineering project data, the key information can be rapidly extracted, and the preliminary review opinions are generated based on the deep learning model, so that the review period is greatly shortened, the manual intervention time is reduced, the overall review efficiency is greatly improved, and the powerful analysis and learning capabilities of the deep learning model are utilized.

Description

Technical Field

[0001] The present invention relates to the technical field of automatic generation of power review opinions, and specifically to an automatic generation method of power review opinions based on artificial intelligence. Background Art

[0002] In the field of power engineering technology review, traditional methods highly rely on manual analysis and expert judgment. The review process not only requires reviewers to have profound professional knowledge, rich practical experience and keen insight, but also requires a large amount of time and effort to carefully read project materials, compare relevant technical standards, analyze potential problems, and finally form written review opinions. This method becomes particularly laborious when faced with an increasing number of complex and ever-changing power engineering projects. The traditional review method has the problem of low efficiency. With the rapid development of the power industry, the number of engineering projects has increased sharply, while the number and ability of reviewers have increased relatively limitedly, resulting in an extended review cycle and affecting the overall progress of the project. Strong subjectivity is another significant problem. Since review opinions are mainly based on the personal experience and subjective judgment of reviewers, there may be significant differences in the review results of the same project by different reviewers, which not only affects the fairness and objectivity of the review, but also brings uncertainty to the subsequent implementation of the project. The traditional review method also has the problem of high cost. In order to ensure the review quality, it is often necessary to hire multiple experts for joint review, which not only increases the project cost, but also increases the difficulty of organization and management.

[0003] Currently, the related technologies and products on the market are still in the primary stage and have many deficiencies. For example, the understanding of power engineering professional terms is not deep enough, the ability to analyze complex problems is limited, and the quality of the generated review opinions is uneven. Therefore, it is particularly important and urgent to develop a more mature, stable and efficient automatic generation method of power review opinions based on artificial intelligence. Summary of the Invention

[0004] (I) Technical Problems to be Solved

[0005] Aiming at the deficiencies of the prior art, the present invention provides an automatic generation method of power review opinions based on artificial intelligence, which solves the problems of insufficient understanding of power engineering professional terms, limited ability to analyze complex problems, and uneven quality of the generated review opinions.

[0006] (II) Technical Solutions

[0007] To achieve the above object, the present invention is realized through the following technical solutions: An automatic generation method of power review opinions based on artificial intelligence, including the following steps:

[0008] a. Data collection: Obtain relevant documents, drawings, and project data of the power project through the data collection unit. The data sources include local storage, cloud databases, or external management systems;

[0009] b. Semantic parsing: Parse the acquired data through the power project semantic parsing module. Adopt a combination of bidirectional long short-term memory network (BiLSTM) and attention mechanism to parse the context semantic information of power project terms and improve the understanding depth of power professional terms;

[0010] c. Term recognition: Use the pre-trained BERT model for fine-tuning, and use the term recognition model to automatically recognize professional terms in the power project to ensure the accuracy and precision of term recognition;

[0011] d. Problem analysis: Through the power project problem analysis model, combine machine learning algorithms such as decision trees and random forests to structurally analyze complex problems in the power project and generate a draft review opinion;

[0012] e. Review opinion generation: Generate preliminary review opinions based on the Generative Adversarial Networks (GANs) technology. At the same time, integrate the analysis results of different models and use the following formula to generate the final review opinion:

[0013]

[0014] Where, Q final is the finally generated review opinion, Q i is the review opinion result generated by each model, W i is the weight coefficient of each model, and n is the number of models participating in the review;

[0015] f. Review opinion optimization: Through the review opinion quality optimization module, based on historical review results, expert feedback, and user input, use reinforcement learning algorithms to optimize the generated review opinions, adjust the model weights, and improve the accuracy and consistency of the review opinions;

[0016] g. User interaction and feedback: Display the generated review opinions through the user interface module. Users can review and give feedback on the generated review opinions. The system dynamically adjusts the model parameters according to the feedback to further improve the quality of the review opinions;

[0017] h. Data storage and update: Store the finally generated review opinions in the storage unit, and perform data synchronization and update with the cloud database or external system through the network communication unit to ensure the real-time and consistency of the data;

[0018] I. Automatically match the generated review opinions with relevant regulations and industry standards of the power project to ensure that the generated review opinions meet the compliance requirements.

[0019] Preferably, in the semantic parsing step, a technology combining a bidirectional long short-term memory network and an attention mechanism is adopted to improve the parsing ability for complex power engineering terms and contexts. The term recognition step is fine-tuned through a pre-trained BERT model, which can adaptively adjust the model parameters according to the background data of different power engineering projects to ensure the accuracy of term recognition and context consistency.

[0020] Preferably, the review opinion generation step can generate diverse and innovative review opinions through generative adversarial network technology, improving the quality of review results.

[0021] Preferably, the weight coefficient W in the formula in the review opinion generation step i is dynamically adjusted according to historical review opinions and model performance to optimize the finally generated review opinions. The review opinion optimization step uses a reinforcement learning algorithm to continuously optimize the parameters of the review opinion generation model by using user feedback information, improving the quality of review opinions.

[0022] Preferably, the user interface module allows users to batch review and edit review opinions for multiple projects, and automatically optimizes the review opinion generation process according to user feedback. The compliance check step can compare with power engineering industry standards and regulations in real time and automatically prompt non-compliant content in review opinions.

[0023] Preferably, the review opinion generation step can automatically adjust the content of the generated review opinions by comparing historical similar project data to make it more in line with the actual situation of the project. The system can automatically generate relevant project reports based on the generated review opinions and provide further analysis and planning suggestions.

[0024] Preferably, the system can adaptively adjust the format and content of the generated opinions according to the review criteria input by the user to ensure that it meets the user's needs.

[0025] Preferably, the term recognition model can automatically adjust the term recognition strategy according to the type and scale of power engineering projects to ensure the accuracy of recognition.

[0026] (III) Beneficial effects

[0027] The present invention provides an automatic generation method for power review opinions based on artificial intelligence, having the following beneficial effects:

[0028] By automating the processing of power engineering project documents, key information can be quickly extracted, and preliminary review opinions can be generated based on deep learning models. This method significantly shortens the review cycle, reduces the time of manual intervention, and greatly improves the overall review efficiency. By leveraging the powerful analysis and learning capabilities of deep learning models, it simulates the review process of experts, avoids the influence of human factors on the review results, and ensures the accuracy and objectivity of the opinions. In addition, by continuously optimizing and iterating the model, the system can continuously enhance its ability to identify and analyze complex problems, ensuring high-quality output of review opinions.

[0029] By reducing the dependence on traditional manual reviewers, the present invention effectively reduces the human resource cost. At the same time, the improvement of review efficiency also shortens the project cycle, thereby indirectly reducing the decision-making risk and time cost. The present invention not only supports automated review but also allows manual intervention and personalized customization. Reviewers can adjust and supplement the automatically generated review opinions according to the actual project needs to meet the special requirements of different projects. Brief Description of the Drawings

[0030] Figure 1 It is a schematic diagram of the internal structure of the present invention. Detailed Embodiments

[0031] The technical solutions in the embodiments of the present invention are clearly and completely described. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0032] The embodiments of the present invention provide an automatic generation method for power review opinions based on artificial intelligence, including,

[0033] a. Data collection: Obtain power engineering-related documents, drawings, and project data through a data collection unit. The data sources include local storage, cloud databases, or external management systems.

[0034] b. Semantic parsing: Parse the acquired data through a power engineering semantic parsing module. Adopt a combination of bidirectional long short-term memory network (BiLSTM) and attention mechanism to parse the context semantic information of power engineering terms, and improve the understanding depth of power professional terms. In the semantic parsing step, a technology combining bidirectional long short-term memory network and attention mechanism is adopted to improve the parsing ability for complex power engineering terms and context.

[0035] b. Term recognition: Fine-tune using a pre-trained BERT model. Use the term recognition model to automatically identify professional terms in power engineering to ensure the precision and accuracy of term recognition. The term recognition step is fine-tuned through a pre-trained BERT model, which can adaptively adjust model parameters according to the background data of different power engineering projects to ensure the accuracy and context consistency of term recognition. The term recognition model can automatically adjust the term recognition strategy according to the type and scale of power engineering projects to ensure the accuracy of recognition.

[0036] d. Problem analysis: Through the power engineering problem analysis model, combine machine learning algorithms such as decision trees and random forests to conduct a structured analysis of complex problems in power engineering and generate a draft review opinion.

[0037] e. Review opinion generation: Generate preliminary review opinions based on Generative Adversarial Networks (GANs) technology. At the same time, integrate the analysis results of different models and use the following formula to generate the final review opinion:

[0038]

[0039] where, Q final is the finally generated review opinion, Q i is the review opinion result generated by each model, W i is the weight coefficient of each model, n is the number of models participating in the review. The review opinion generation step can generate diverse and innovative review opinions through Generative Adversarial Networks technology, improving the quality of review results. The weight coefficient W i in the formula in the review opinion generation step is dynamically adjusted according to historical review opinions and model performance to optimize the finally generated review opinion.

[0040] f. Review opinion optimization: Through the review opinion quality optimization module, based on historical review results, expert feedback, and user input, use reinforcement learning algorithms to optimize the generated review opinions, adjust model weights, and improve the accuracy and consistency of review opinions. The review opinion optimization step uses reinforcement learning algorithms to continuously optimize the parameters of the review opinion generation model using user feedback information, improving the quality of review opinions. The review opinion generation step can automatically adjust the content of the generated review opinion by comparing historical similar project data to make it more in line with the actual situation of the project. The system can automatically generate relevant project reports based on the generated review opinions and provide further analysis and planning suggestions.

[0041] g. User Interaction and Feedback: The generated review opinions are presented through the user interface module. Users can review and provide feedback on the generated review opinions. The system dynamically adjusts the model parameters based on the feedback to further improve the quality of the review opinions. The user interface module allows users to batch review and edit the review opinions of multiple projects and automatically optimizes the review opinion generation process according to user feedback.

[0042] h. Data Storage and Update: The finally generated review opinions are stored in the storage unit, and data synchronization and update are performed with the cloud database or external system through the network communication unit to ensure the real-time nature and consistency of the data. The compliance check step can compare with the power engineering industry standards and regulations in real time and automatically prompt the non-compliant content in the review opinions.

[0043] I. Automatically match the generated review opinions with the regulations and industry standards related to power engineering to ensure that the generated review opinions meet the compliance requirements. The system can adaptively adjust the format and content of the generated opinions according to the review criteria input by the user to ensure that they meet the user's needs.

[0044] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. The method for automatically generating power review opinions based on artificial intelligence is characterized by: The following steps are involved: a. Data collection: Obtain power engineering related documents, drawings and project data through data collection units. Data sources include local storage, cloud databases or external management systems; b. Semantic parsing: The acquired data is parsed through the power engineering semantic parsing module, and the contextual semantic information of power engineering terms is parsed by combining the bidirectional long short-term memory network (BiLSTM) with the attention mechanism to improve the depth of understanding of power professional terms; c. Terminology recognition: fine-tune the pre-trained BERT model and use the terminology recognition model to automatically recognize professional terms in power engineering to ensure the precision and accuracy of terminology recognition; d. Problem analysis: Through the power engineering problem analysis model, combined with machine learning algorithms such as decision trees and random forests, the complex problems of power engineering are structured and analyzed to generate corresponding draft review opinions; e. Review opinion generation: Generate preliminary review opinions based on generative adversarial networks (GANs) technology, integrate the analysis results of different models, and use the following formula to generate the final review opinions: Among them, Q final Q is the final review opinion. i The review results generated for each model, W i is the weight coefficient of each model, and n is the number of models participating in the review; f. Review opinion optimization: Through the review opinion quality optimization module, based on historical review results, expert feedback and user input, the generated review opinions are optimized using reinforcement learning algorithms, and the model weights are adjusted to improve the accuracy and consistency of the review opinions; g. User interaction and feedback: The generated review opinions are displayed through the user interface module. Users can review and provide feedback on the generated review opinions. The system dynamically adjusts the model parameters based on the feedback to further improve the quality of the review opinions. h. Data storage and update: The final review opinions are stored in the storage unit, and the data is synchronized and updated with the cloud database or external system through the network communication unit to ensure the real-time and consistency of the data; I. Automatically match the generated review opinions with the laws and regulations and industry standards related to power engineering to ensure that the generated review opinions meet the compliance requirements.

2. The method for automatically generating power review opinions based on artificial intelligence according to claim 1 is characterized by: The semantic parsing step adopts a technology combining a bidirectional long short-term memory network and an attention mechanism to improve the parsing capability of complex power engineering terms and contexts. The term recognition step is fine-tuned through a pre-trained BERT model, which can adaptively adjust model parameters according to the background data of different power engineering projects to ensure the accuracy of term recognition and context consistency.

3. The method for automatically generating power review opinions based on artificial intelligence according to claim 1 is characterized in that: The review opinion generation step can generate diverse and innovative review opinions by generating adversarial network technology, thereby improving the quality of the review results.

4. The method for automatically generating power review opinions based on artificial intelligence according to claim 1 is characterized in that: The weight coefficient W in the formula in the review opinion generation step i Dynamic adjustments are made based on historical review opinions and model performance to optimize the review opinions that are finally generated. The review opinion optimization step uses a reinforcement learning algorithm and user feedback information to continuously optimize the parameters of the review opinion generation model to improve the quality of the review opinions.

5. The method for automatically generating power review opinions based on artificial intelligence according to claim 1 is characterized in that: The user interface module allows users to review and edit review opinions for multiple projects in batches, and automatically optimizes the review opinion generation process based on user feedback. The compliance check step can compare the power engineering industry standards and regulations in real time, and automatically prompt non-compliant content in the review opinions.

6. The method for automatically generating power review opinions based on artificial intelligence according to claim 1 is characterized in that: The review opinion generation step can automatically adjust the generated review opinion content by comparing historical similar project data to make it more in line with the actual project situation. The system can automatically generate relevant project reports based on the generated review opinions and provide further analysis and planning suggestions.

7. The method for automatically generating power review opinions based on artificial intelligence according to claim 1 is characterized by: The system can adaptively adjust the format and content of the generated opinions according to the review criteria input by the user to ensure that they meet the user's needs.

8. The method for automatically generating power review opinions based on artificial intelligence according to claim 1 is characterized by: The terminology recognition model can automatically adjust the terminology recognition strategy according to the type and scale of the power engineering project to ensure the accuracy of recognition.

Images