An MSWI process controlled object modeling system and method based on LLM hybrid driving

By using an MSWI process controlled object modeling system based on LLM hybrid drive, the problems of strong coupling nonlinearity and drastic fluctuations in operating conditions of MSWI processes are solved, achieving efficient, low-carbon and precise control of MSWI processes. This improves the model's ability to capture local nonlinear dynamics and generalize across operating conditions, meeting the requirements of industrial-grade real-time control and compliance supervision.

CN122174647APending Publication Date: 2026-06-09BEIJING UNIV OF TECH
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN Β· China
Patent Type
Applications(China)
Current Assignee / Owner
BEIJING UNIV OF TECH
Filing Date
2026-03-05
Publication Date
2026-06-09

AI Technical Summary

Technical Problem

Existing technologies are ill-suited to the strongly coupled nonlinear characteristics, drastic fluctuations in operating conditions, and multi-objective conflicts of MSWI processes. General-purpose large language models lack deep embedding of vertical domain mechanisms, have insufficient generalization ability for extreme operating conditions, and their inference latency and uninterpretability make it difficult to meet the needs of industrial-grade real-time control and compliance supervision.

Method used

A process controlled object modeling system based on LLM hybrid drive is adopted. The system performs deep semantic fusion by driving the MSWI full-process model module through LLM, and combines it with the solid waste combustion and waste heat exchange model module driven by historical data to build a multi-source information fusion strategy. This enables dynamic coordination of combustion, waste heat and purification processes, breaks down information barriers, and constructs a hybrid architecture of 'LLM global drive + data local modeling'. It also innovatively designs a dual fusion mechanism for controlled variables and environmental indicators.

Benefits of technology

It achieves efficient, low-carbon, and precise control of MSWI processes, enhances the model's ability to capture local nonlinear dynamics and generalize across operating conditions, breaks through the limitations of traditional single-objective models, provides a new paradigm for intelligent modeling, and supports efficient, low-carbon, and precise control in the MSWI industry.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122174647A_ABST
    Figure CN122174647A_ABST
Patent Text Reader

Abstract

This invention provides a controlled object modeling system and method for MSWI processes based on LLM hybrid driving, relating to the field of solid waste incineration modeling technology. The system includes: an LLM-driven MSWI full-process module for constructing a working condition cognition model and outputting predictions of the first controlled variable and the first environmental indicator through dual channels; a historical data-driven solid waste combustion and waste heat exchange model module for receiving first purified material flow parameters, constructing a dynamic model, and outputting predictions of the second controlled variable; a controlled variable output fusion module for fusing the two prediction data streams to generate the final controlled variable output; a historical data-driven flue gas purification model module for using the second controlled variable prediction as input to the preceding working condition and combining it with the second purified material flow parameters to predict the second environmental indicator; and an environmental indicator output fusion module for fusing multi-source environmental indicators to construct a performance evaluation system. This invention solves the characterization challenges of strongly coupled, nonlinear, and multi-constraint MSWI processes.
Need to check novelty before this filing date? Find Prior Art