Method and system for cleaning fuel oil of diesel vehicle
By combining sensor arrays and reinforcement learning Q-value tables, the cleaning threshold is dynamically calculated and the optimal cleaning action is selected, which solves the problems of delayed cleaning timing and synergistic corrosion of multiple pollutants in diesel vehicle fuel cleaning technology. This achieves adaptive fuel system cleaning control, improving cleaning effect and energy consumption optimization.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-18
- Publication Date
- 2026-03-27
AI Technical Summary
Existing diesel vehicle fuel cleaning technologies cannot adapt to dynamic changes in engine operating conditions, resulting in delayed or excessive cleaning. Furthermore, they lack quantitative assessment of the synergistic corrosion effects of multiple contaminants, impacting component lifespan and maintenance costs.
By acquiring fuel quality data and vehicle operating parameters through a sensor array, calculating dynamic cleaning thresholds, determining pollution levels using a multi-pollutant synergistic effect index, selecting the optimal cleaning action based on a reinforcement learning Q-value table, and adjusting cleaning intensity using PID feedback, adaptive cleaning control is achieved.
It improves the accuracy and robustness of cleaning threshold determination, optimizes cleaning effect and energy consumption, extends component life, and reduces maintenance costs.
Smart Images

Figure CN121738792A_ABST