Search

CN-121981097-A - Intelligent diagnosis report generation method for coal motor unit

CN121981097ACN 121981097 ACN121981097 ACN 121981097ACN-121981097-A

Abstract

The invention relates to a coal motor unit-oriented intelligent diagnosis report generation method which comprises a device early warning module, a knowledge base building module, an MCP service tool module, a large model building module and a report output module, wherein the device early warning module comprises a fault diagnosis model obtained based on training of equipment asset trees and historical measurement point data of a coal motor unit, the knowledge base building module comprises a maintenance log knowledge base, a fault procedure knowledge base and a diagnosis report generation template, the MCP service tool module comprises at least one MCP tool used for acquiring measurement point original data, asset tree basic data and energy efficiency optimization data, the large model building module comprises a large language model which is deployed locally and supports calling of the MCP tool, and the report output module is used for responding to an early warning event triggered by the device early warning module to generate a diagnosis report. The intelligent diagnosis system has the beneficial effects that the intelligent diagnosis system capable of automatically calling the time sequence and energy efficiency interface is constructed by fusing the overhaul log, the fault procedure and the asset data through the large model, so that the knowledge standardized precipitation and the efficient report generation are realized.

Inventors

  • CAO YAN
  • WU LONGFEI
  • Tao Menli
  • GUO QING
  • YE XINNAN
  • WANG CHENGZHANG
  • PAN JUNWEI
  • Shao Chengan
  • GUO LIANG
  • Wu Chaohao
  • WANG JI
  • He tianjiao
  • JIA XIAOYAN
  • GU BAO

Assignees

  • 浙江浙能嘉华发电有限公司
  • 浙江浙能数字科技有限公司

Dates

Publication Date
20260505
Application Date
20260120

Claims (9)

  1. 1. The intelligent diagnosis report generation method for the coal motor unit is characterized by comprising the following steps of: step 1, constructing an equipment early warning module, wherein the early warning module comprises a fault diagnosis model which is obtained based on training of a coal electric machine group equipment asset tree and historical measuring point data and is used for monitoring and triggering early warning events in real time; step 2, constructing a knowledge base module, wherein the knowledge base module comprises a maintenance log knowledge base, a fault procedure knowledge base and a diagnosis report generation template; Step 3, constructing an MCP service tool module, wherein the MCP service tool module comprises at least one MCP tool for acquiring measuring point original data, asset tree basic data and energy efficiency optimization data; Step 4, constructing a large model module, wherein the large model module comprises a large language model which is deployed locally and supports calling of an MCP tool; And 5, constructing a report output module, wherein the report output module is used for responding to an early warning event triggered by the equipment early warning module, automatically calling related knowledge and templates in the knowledge base module by the large language model, and calling a corresponding MCP tool based on an analysis result to acquire related data so as to generate a diagnosis report.
  2. 2. The intelligent diagnosis report generating method for a coal motor unit according to claim 1, wherein step 1 comprises: step 1.1, carding equipment measuring points of a coal motor unit, and constructing a tree-shaped equipment asset tree according to membership of equipment and the measuring points; and 1.2, collecting historical data of measuring points of each device based on the device asset tree, training and deploying a device fault diagnosis model, and forming a device early warning module.
  3. 3. The intelligent diagnosis report generating method for a coal motor unit according to claim 1, wherein step 2 comprises: Step 2.1, multi-source heterogeneous data of a unit maintenance log are collected, and a maintenance log knowledge base comprising three elements of maintenance objects, maintenance activities and maintenance results is constructed after data cleaning, entity identification and relation extraction; Step 2.2, collecting a unit fault specification document system, and constructing a fault specification knowledge base containing three elements of fault description, operation specification and parameter threshold after structural modeling; And 2.3, constructing a diagnosis report generating template comprising an integral frame primary template, a chapter structure secondary template and a sentence pattern tertiary template, and marking dynamic fields in the template.
  4. 4. The intelligent diagnosis report generating method for a coal motor unit according to claim 1, wherein step 3 comprises: Step 3.1, packaging a relevant debugging interface of a power generation original time sequence database and a budget function developed later into a large model callable MCP tool to form a time sequence data MCP tool; Step 3.2, packaging a KKS coding system, spare part associated information and asset tree node relation data constructed based on a knowledge graph in the existing system of the power plant into a large model callable MCP tool to form asset tree basic data MCP tools; And 3.3, packaging the time period energy efficiency index, the consumption difference model, the environment-friendly emission intensity, the optimal working condition reference library and the predicted energy efficiency loss evaluation algorithm in the performance calculation engine into a MCP tool which can be called by the large model to form energy efficiency optimization data MCP tool.
  5. 5. The intelligent diagnosis report generating method for a coal motor unit according to claim 4, wherein step 4 comprises: Step 4.1, selecting a proper open source large model and carrying out privately-arranged locally; and 4.2, registering the MCP tool.
  6. 6. The intelligent diagnosis report generating method for a coal motor unit according to claim 4, wherein step 5 comprises: Step 5.1, triggering an early warning event in the running of the coal motor group in real time by an equipment early warning module; Step 5.2, analyzing the early warning event by utilizing the large model capacity module, and calling related knowledge in the knowledge base module and a diagnosis report generation template; step 5.3, the large language model analyzes the required data according to the early warning event and the called knowledge base knowledge, and calls a corresponding MCP tool to acquire related data; step 5.4, generating a diagnosis report the first edition according to the called knowledge base knowledge, the diagnosis report generation template and the acquired related data; And 5.5, performing fault confirmation and field condition supplementation on the diagnosis report the first edition by a field expert to generate a final fault diagnosis report.
  7. 7. A coal motor unit-oriented intelligent diagnostic report generating system for performing the method of any of claims 1 to 6, comprising: the early warning module is used for monitoring and triggering early warning events in real time based on a fault diagnosis model obtained by training the asset tree of the coal electric motor unit equipment and the historical measurement point data; The knowledge base module is used for storing unit maintenance log knowledge, fault procedure knowledge and a hierarchical diagnosis report generation template; the MCP service tool module comprises at least one MCP tool for acquiring measurement point original data, asset tree basic data and energy efficiency optimization data; A large model module comprising a large language model deployed locally and supporting invocation of MCP tools; And the report output module is used for responding to the early warning event triggered by the equipment early warning module, automatically calling related knowledge and templates in the knowledge base module by the large language model, calling a corresponding MCP tool based on the analysis result to acquire related data, and further generating a diagnosis report.
  8. 8. A computer storage medium, characterized in that the computer storage medium has stored therein a computer program which, when run on a computer, causes the computer to perform the method according to any of claims 1 to 6.
  9. 9. An electronic device, comprising: A memory for storing a computer program; a processor for executing the computer program to implement the method of any one of claims 1 to 6.

Description

Intelligent diagnosis report generation method for coal motor unit Technical Field The invention belongs to the technical field of power generation information, and particularly relates to an intelligent diagnosis report generation method for a coal motor unit. Background The modern power plant equipment system is complex, the coupling is strong, and the fault analysis is a key link for guaranteeing safe and economic operation. The traditional manual diagnosis report has three major bottlenecks, namely low compiling efficiency, long time consumption and quality fluctuation along with personnel experience from data extraction and trend analysis to conclusion writing, discretization of knowledge, difficult multiplexing and inheritance of diagnosis experience in an unstructured form or model training, formation of knowledge blind areas, lack of standardization, non-uniform report format and content, influence on timeliness and subsequent comparison analysis. Disclosure of Invention The invention aims to overcome the defects in the prior art and provides an intelligent diagnosis report generation method for a coal motor unit. In a first aspect, a method for generating an intelligent diagnosis report for a coal motor unit is provided, including: step 1, constructing an equipment early warning module, wherein the early warning module comprises a fault diagnosis model which is obtained based on training of a coal electric machine group equipment asset tree and historical measuring point data and is used for monitoring and triggering early warning events in real time; step 2, constructing a knowledge base module, wherein the knowledge base module comprises a maintenance log knowledge base, a fault procedure knowledge base and a diagnosis report generation template; Step 3, constructing an MCP service tool module, wherein the MCP service tool module comprises at least one MCP tool for acquiring measuring point original data, asset tree basic data and energy efficiency optimization data; Step 4, constructing a large model module, wherein the large model module comprises a large language model which is deployed locally and supports calling of an MCP tool; And 5, constructing a report output module, wherein the report output module is used for responding to an early warning event triggered by the equipment early warning module, automatically calling related knowledge and templates in the knowledge base module by the large language model, and calling a corresponding MCP tool based on an analysis result to acquire related data so as to generate a diagnosis report. Preferably, step 1 includes: step 1.1, carding equipment measuring points of a coal motor unit, and constructing a tree-shaped equipment asset tree according to membership of equipment and the measuring points; and 1.2, collecting historical data of measuring points of each device based on the device asset tree, training and deploying a device fault diagnosis model, and forming a device early warning module. Preferably, step 2 includes: Step 2.1, multi-source heterogeneous data of a unit maintenance log are collected, and a maintenance log knowledge base comprising three elements of maintenance objects, maintenance activities and maintenance results is constructed after data cleaning, entity identification and relation extraction; Step 2.2, collecting a unit fault specification document system, and constructing a fault specification knowledge base containing three elements of fault description, operation specification and parameter threshold after structural modeling; And 2.3, constructing a diagnosis report generating template comprising an integral frame primary template, a chapter structure secondary template and a sentence pattern tertiary template, and marking dynamic fields in the template. Preferably, step 3 includes: Step 3.1, packaging a relevant debugging interface of a power generation original time sequence database and a budget function developed later into a large model callable MCP tool to form a time sequence data MCP tool; Step 3.2, packaging a KKS coding system, spare part associated information and asset tree node relation data constructed based on a knowledge graph in the existing system of the power plant into a large model callable MCP tool to form asset tree basic data MCP tools; And 3.3, packaging the time period energy efficiency index, the consumption difference model, the environment-friendly emission intensity, the optimal working condition reference library and the predicted energy efficiency loss evaluation algorithm in the performance calculation engine into a MCP tool which can be called by the large model to form energy efficiency optimization data MCP tool. Preferably, step 4 includes: Step 4.1, selecting a proper open source large model and carrying out privately-arranged locally; and 4.2, registering the MCP tool. Preferably, step 5 includes: Step 5.1, triggering an early warning event in the running of the coal motor group i