A method and system for automatic burning of hot blast stove based on deep reinforcement learning
A technology of reinforcement learning and hot blast stove, applied in neural learning methods, furnaces, blast furnaces, etc., can solve the bottleneck of knowledge acquisition, difficulty in summarizing and extracting control rules, etc., to achieve high control accuracy, high optimization efficiency, and high flexibility. Effect
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0035] Such as figure 1 As shown, this embodiment provides a method for automatic firing of hot blast stoves based on deep reinforcement learning, including:
[0036] S1: Obtain the historical furnace firing data of the hot blast stove;
[0037] S2: According to the historical furnace data, train the depth enhancement model of the hot blast stove in different combustion stages;
[0038] The training includes: according to the state of the hot blast stove and the firing action at the previous moment, predict the action range of the firing action in the current hot blast stove state and the feedback state of the hot blast stove at the next moment after the firing action is performed; The reward and punishment value of the firing action, and select the optimal firing action in the current state of the hot blast stove; iteratively calculate the optimal firing action at each moment in the current combustion stage;
[0039] S3: Carry out automatic burning control on the hot blast ...
Embodiment 2
[0069] This embodiment provides a hot stove automatic firing system based on deep reinforcement learning, including:
[0070] The data acquisition module is used to obtain the historical furnace burning data of the hot blast stove;
[0071] The model training module is used to train the depth enhancement model of the hot blast stove in different combustion stages according to the historical furnace data;
[0072] The training includes: according to the state of the hot blast stove and the firing action at the previous moment, predict the action range of the firing action in the current hot blast stove state and the feedback state of the hot blast stove at the next moment after the firing action is performed; The reward and punishment value of the firing action, and select the optimal firing action in the current state of the hot blast stove; iteratively calculate the optimal firing action at each moment in the current combustion stage;
[0073] The automatic control module is...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More 


