Research on Surge Control of Centrifugal Compressor Based on Reinforcement Learning

Jiang, Kun; Xiang, Yang; Chen, Tianyou; Jiang, Chaojun

doi:10.1007/978-3-030-57745-2_25

Cited by 4 publications

(2 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Certain industrial operations, such as refining, power generation and chemical processes, require that fluids flow at very high pressures. Typically, compressors increase the pressure in flowing fluids by mechanically decreasing their volumes [1], [2]. In a compressor system, fluid passes through the suction eye at a lower velocity and exits the impeller circumference faster [3].…”

Section: Introductionmentioning

confidence: 99%

Optimal design of full order state observer for active surge control in centrifugal compressors using genetic algorithm

et al. 2022

View full text Add to dashboard Cite

This paper presents the design of a genetic algorithm (GA) based full order state observer for a compressor system. The observer measures the compressor system parameters and computes the difference between the measured output and the setpoint, which is used as feedback to the active surge controller. The proposed control structure with a GA tuned compressor characteristic is required to minimize system oscillations and energy losses induced by the piston actuation gas recycling technique. The design method for GA based full order observer begins with a state observer modelling, observer error estimation based on the magnitude of fluid friction and mechanical effects on measured parameters. The aim of the controller is to regulate the compressor driver speed such that the measured output achieves its required set point. In order to achieve a maximum compressor operating capacity, the compressor characteristics were optimized using GA, where the mass flow was maximized to improve the compressor efficiency. Simulations were carried out, and results showed the viability of GA based full order state observer compared with the piston actuation recycle method to control active surge in centrifugal compressor systems.

show abstract

Section: Introductionmentioning

confidence: 99%

Optimal design of full order state observer for active surge control in centrifugal compressors using genetic algorithm

et al. 2022

View full text Add to dashboard Cite

show abstract

“…Due to the destructive effects of the surge phenomenon of the compressor system, many attempts have been made to control this instability in the centrifugal compressor. [7][8][9][10][11][12] The two main categories for surge instability control are surge prevention and active surge control. 13,14 In the first method, by selecting the surge control line and designing the controller, an attempt is made to prevent the system from entering the unstable zone.…”

Section: Introductionmentioning

confidence: 99%

Robust active finite-time control of gas compressor system surge in the presence of unmatched disturbance and uncertainty

Sun¹,

Gu²,

Zhang³

et al. 2021

Proceedings of the Institution of Mechanical Engineers, Part I:

View full text Add to dashboard Cite

Active and robust control of surge instability is a special necessity for optimal and safe operation of centrifugal compressors, and for the purpose, this article presents a new hybrid scheme based on fuzzy and terminal sliding mode methods. The Greitzer model is used to design a novel controller when the disturbance instability in the flow and pressure alike the uncertainity in the compressor characteristic curve and throttle valve are embedded in it. The fuzzy approximator is used to estimate the effects of parametric uncertainty and the nonlinear terms, and the robustness of the proposed method is guaranteed using the terminal sliding mode control method. The Lyapunov criterion is utilized to verify the finite-time stability of the closed-loop system. The performance of the presented method is compared with other methods in the literature through simulations in MATLAB software. The results suggest that our designed controller outperforms the existing ones in terms of surge prevention and robustness against unmatched uncertainties and disturbances.

show abstract

Application of artificial intelligence in turbomachinery aerodynamics: progresses and challenges

Zou,

Xu,

Chen

et al. 2024

Artif Intell Rev

View full text Add to dashboard Cite

Turbomachinery plays a vital role in energy conversion systems, with aerodynamic issues being integral to its entire lifecycle, spanning the period of design, validation, and maintenance. Conventionally, the reliance on skilled aerodynamic engineers has been pivotal in the successful development of turbomachines. However, in the current era of burgeoning artificial intelligence (AI) technology, researchers are increasingly turning to AI to replace human expertise and decision-making in these aerodynamic issues and to solve previously intractable aerodynamic problems. This paper presents a systematic literature review of the latest advancements in applying AI to turbomachinery aerodynamics, encompassing the design, validation, and maintenance of compressors and turbines. It underscores how AI is revolutionizing the research paradigm of turbomachinery aerodynamics. AI’s powerful learning capability facilitates more precise and convenient aerodynamic analyses and inspires innovative aerodynamic design ideas that go beyond the capabilities of classical design techniques. Additionally, AI’s autonomous decision-making capability can be employed for aerodynamic optimization and active flow control of turbomachines, generating optimal aerodynamic solutions and complex control strategies that surpass human brains. As a main contribution, we provide a detailed exposition of the future intelligent turbomachinery research and development (R &D) system, along with highlighting potential challenges such as physics embedding, interactive 3D design optimization, and real-time prognoses. It is anticipated that harnessing AI’s full potential will lead to a comprehensive AI-based turbomachinery R &D system in the future.

show abstract

Research on Surge Control of Centrifugal Compressor Based on Reinforcement Learning

Cited by 4 publications

References 8 publications

Optimal design of full order state observer for active surge control in centrifugal compressors using genetic algorithm

Optimal design of full order state observer for active surge control in centrifugal compressors using genetic algorithm

Robust active finite-time control of gas compressor system surge in the presence of unmatched disturbance and uncertainty

Application of artificial intelligence in turbomachinery aerodynamics: progresses and challenges

Contact Info

Product

Resources

About