Camellia oleifera dynamic drying parameter intelligent optimization method

By constructing a double-layer stacked LSTM neural network model and a reinforcement learning agent guided by a multi-objective reward function, the problems of insufficient data-driven capability and insufficient dynamic adaptability of control strategy in the drying process of camellia seeds are solved. Adaptive tracking and continuous and accurate optimization of parameters in the drying process of camellia seeds are realized, thereby improving prediction accuracy and energy efficiency.

CN122241370APending Publication Date: 2026-06-19RES INST OF SUBTROPICAL FORESTRY CHINESE ACAD OF FORESTRY
0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
RES INST OF SUBTROPICAL FORESTRY CHINESE ACAD OF FORESTRY
Filing Date
2026-03-26
Publication Date
2026-06-19

AI Technical Summary

⚠Technical Problem

Existing technologies for optimizing camellia seed drying parameters suffer from insufficient data-driven capabilities, lagging quality control, single prediction targets, and insufficient dynamic adaptability of control strategies. This results in inaccurate moisture control during the camellia seed drying process, difficulty in preserving oil quality, and high energy consumption.

⚗Method used

By collecting experimental drying data and simulated drying data, a basic training dataset is constructed. A two-layer stacked LSTM neural network model is used for prediction. Combined with a multi-objective reward function, the reinforcement learning agent is guided to determine the optimal control action, thereby achieving real-time optimization of drying temperature and hot air velocity.

🎯Benefits of technology

It achieves adaptive tracking and continuous, precise optimization of parameters in the drying process of camellia seeds, improving prediction accuracy and energy efficiency, ensuring oil quality and reducing energy consumption.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
Patent Text Reader

Abstract

This application discloses an intelligent optimization method for dynamic drying parameters of camellia seeds, comprising: collecting experimental drying data and actual real-time drying data during the drying process of camellia seeds; generating simulated drying data through orthogonal experimental design; supplementing with extreme working condition data; and fusing the simulated drying data with the experimental drying data to construct a basic training dataset, including moisture content, acid value, and energy consumption under different temperatures and wind speeds; training a prediction model based on the basic training dataset, using control parameters from historical time periods as input, and outputting a predicted sequence of state parameters for future time periods until convergence to obtain the optimal prediction model; determining the optimal control action based on real-time drying data through a multi-objective reward function; converting the optimal control action into control commands, obtaining updated state data after the equipment executes the commands, and feeding back the updated optimal control action. This application can improve the practicality, comprehensiveness, and accuracy of drying parameter optimization.
Need to check novelty before this filing date? Find Prior Art