An engine combustion field automatic control model construction method

CN117272796BActive Publication Date: 2026-07-07BEIJING JINGHANG COMPUTING & COMM RES INST
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
BEIJING JINGHANG COMPUTING & COMM RES INST
Filing Date
2023-09-18
Publication Date
2026-07-07

AI Technical Summary

Technical Problem

Existing PID control algorithms cannot optimize multiple key parameters simultaneously, which limits the improvement of engine performance. They rely on manual parameter adjustment, which is costly and makes it difficult to achieve intelligent control.

Method used

A deep reinforcement learning agent based on the Actor-Critic framework is constructed. By receiving empirical data, an automatic control model of the engine combustion field is trained to optimize parameters such as fuel equivalence ratio, fuel injection ratio, and slider displacement. Deep reinforcement learning algorithms are used for multi-parameter control.

Benefits of technology

It achieves precise control of the engine combustion field, can simultaneously optimize multiple key parameters, reduces operating costs, improves the level of intelligence in control, reduces the influence of subjective factors, and makes the combustion process more accurate.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117272796B_ABST
    Figure CN117272796B_ABST
Patent Text Reader

Abstract

The present application relates to a kind of engine combustion field automatic control model construction method, comprising: receiving the experience data of agent;Including: the state space s of agent current period, the engine simulation environment action space a generated by agent policy network based on the state space of current period, the state space s' of agent next period determined based on engine simulation environment action space, the reward information r determined based on the state space of next period and end signal d;Experience data <s,a,s',r,d> is stored in experience pool;When experience data in experience pool reaches preset data volume, experience data in experience pool is obtained as training data, and engine combustion field automatic control model is obtained by training agent. Realize the accurate control and adjustment of multiple key parameters of engine combustion field by constructing engine combustion field automatic control model, so that engine combustion process can more accurately reach the expected target pressure value.
Need to check novelty before this filing date? Find Prior Art