A method for calculating weights of evaluation indexes for site selection of large-span underground powerhouse

By constructing a geological risk knowledge graph and a GEO-KG-BERT model, the subjectivity problem in weight calculation in the site selection of large-span underground powerhouses is solved, realizing efficient and objective indicator weight calculation and risk assessment, which is applicable to scenarios such as tunnel engineering and mining.

CN122114256APending Publication Date: 2026-05-29THREE GORGES JINSHAJIANG CHUANYUN HYDROPOWER DEV CO LTD

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
THREE GORGES JINSHAJIANG CHUANYUN HYDROPOWER DEV CO LTD
Filing Date
2026-01-06
Publication Date
2026-05-29

Smart Images

  • Figure CN122114256A_ABST
    Figure CN122114256A_ABST
Patent Text Reader

Abstract

The application provides a kind of large-span underground powerhouse site selection evaluation index weight calculation method, belong to the intersection technical field of geology engineering and artificial intelligence, including the following steps: define the ontology model of geological risk knowledge graph;Based on BERT model, model fine-tuning is carried out to generate GEO-KG-BERT model, so that GEO-KG-BERT model can accurately identify the core entity and key relationship in geological text;The core entity and key relationship are extracted using the GEO-KG-BERT model, and the key relationship of the ontology model of "node-edge-attribute" is constructed;Based on the geological risk knowledge graph, the number of relationships, path intensity and importance are calculated;The importance of all geological indicators is normalized to generate an initial weight vector;The initial weight vector is corrected to obtain the final weight vector.The application calculates the weight of the large-span underground powerhouse site selection evaluation index based on the knowledge graph and GEO-KG-BERT model, and dynamically corrects the semantic support parameters based on the GEO-KG-BERT model, which improves the objectivity and accuracy of the risk evaluation result.
Need to check novelty before this filing date? Find Prior Art