Research on Algorithm Fusion for the Control of a Roots-Type Waste Heat Power Generation System

Abstract

At present, the recovery of low-quality waste heat is a major problem in energy utilization. To solve this problem and improve energy efficiency, this research group designed a low-quality waste heat power generation device with a roots-type power machine as the core component. However, the power generation device produces a large hysteresis in power generation regulation. While the hardware can be improved, the design of the measurement and control system is also critical. In view of the problems existing in low-quality waste heat power generation devices, this research group introduced an internal model controller into the control system and designed an internal model controller and filter through the analysis of each module. In addition, to improve the performance of the controller, this research group applied the deep learning method to optimize the control system and used the prediction function of the deep learning method to further improve the stability of the device. The simulation and experimental results show that the control strategy can make this device for the recovery of low-quality waste heat respond quickly to fluctuations in the gas source and improve the hysteresis problem.

Keywords

• Deep learning
• internal model control
• low-quality waste heat
• root-type power machine
• waste heat power generation