Air conditioner compressor torque compensation method and device based on DQN network and air conditioner
By applying DQN-based agents to perform torque compensation calculations in the air conditioning compressor, the problems of manual calculation dependence and inefficiency in the prior art are solved, and fast and efficient torque compensation is achieved, reducing the impact of vibration.
Patent Information
- Application Number
- CN202411953355.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-27
- Publication Date
- 2025-04-04
AI Technical Summary
In the prior art, compressor torque compensation mainly relies on manual calculations, and there are problems of experience dependence and low calculation efficiency, making it difficult to quickly adapt to the changes in compressor operating frequency.
Using an agent based on a deep Q learning network (DQN) to initialize the network weights of the compressor vibration numerical test environment and torque compensation model, the action values generated by the agent based on the current state are obtained, and input them into the test environment for compressor control and vibration numerical acquisition, and optimize the generation strategy until the torque compensation parameters that meet preset requirements are obtained.
The calculation of compressor torque compensation is quickly completed, the calculation efficiency is improved, and the impact of the pipe group on outdoor unit vibration is significantly reduced.