{"title":"Sliding mode control for longitudinal oscillating combustion","authors":"Long Zhang, Xingyu Su, Hua Zhou, Z. Ren","doi":"10.1080/13647830.2023.2197409","DOIUrl":null,"url":null,"abstract":"Considerable research has been reported on developing effective active control means to suppress oscillating combustion. The typical pressure oscillation can be divided into linear growth, transition and saturation stages. In this study, a sliding mode control strategy, consisting of a state estimate model, disturbance observers and a sliding mode controller, is proposed to suppress the longitudinal oscillating combustion. The control strategy is first tested with a nonlinear 0D state space model as the controlled plant. Results show that the state estimate model combined with the singular spectrum analysis (SSA) method can accurately estimate the system state quantities by grouping the SSA modes according to the frequency difference and calculating mode envelopes. To ensure the estimate accuracy, the number of truncated SSA modes varies according to the oscillation stage. The disturbance observers are designed to improve the robustness of the controller by introducing broadband spectrum disturbance to account for the external noise in the observed values. The sliding mode controller can limit the disturbance amplitude, and effectively suppress the pressure oscillation. A 1D Rijke tube acoustic network is also tested to further validate the controller adaptability. With this controller, the Rijke tube pressure oscillation can be effectively eliminated when control starts at the linear growth, transition, or saturation stages.","PeriodicalId":50665,"journal":{"name":"Combustion Theory and Modelling","volume":"27 1","pages":"653 - 684"},"PeriodicalIF":1.9000,"publicationDate":"2023-04-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Combustion Theory and Modelling","FirstCategoryId":"5","ListUrlMain":"https://doi.org/10.1080/13647830.2023.2197409","RegionNum":4,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q4","JCRName":"ENERGY & FUELS","Score":null,"Total":0}
引用次数: 0
Abstract
Considerable research has been reported on developing effective active control means to suppress oscillating combustion. The typical pressure oscillation can be divided into linear growth, transition and saturation stages. In this study, a sliding mode control strategy, consisting of a state estimate model, disturbance observers and a sliding mode controller, is proposed to suppress the longitudinal oscillating combustion. The control strategy is first tested with a nonlinear 0D state space model as the controlled plant. Results show that the state estimate model combined with the singular spectrum analysis (SSA) method can accurately estimate the system state quantities by grouping the SSA modes according to the frequency difference and calculating mode envelopes. To ensure the estimate accuracy, the number of truncated SSA modes varies according to the oscillation stage. The disturbance observers are designed to improve the robustness of the controller by introducing broadband spectrum disturbance to account for the external noise in the observed values. The sliding mode controller can limit the disturbance amplitude, and effectively suppress the pressure oscillation. A 1D Rijke tube acoustic network is also tested to further validate the controller adaptability. With this controller, the Rijke tube pressure oscillation can be effectively eliminated when control starts at the linear growth, transition, or saturation stages.
期刊介绍:
Combustion Theory and Modelling is a leading international journal devoted to the application of mathematical modelling, numerical simulation and experimental techniques to the study of combustion. Articles can cover a wide range of topics, such as: premixed laminar flames, laminar diffusion flames, turbulent combustion, fires, chemical kinetics, pollutant formation, microgravity, materials synthesis, chemical vapour deposition, catalysis, droplet and spray combustion, detonation dynamics, thermal explosions, ignition, energetic materials and propellants, burners and engine combustion. A diverse spectrum of mathematical methods may also be used, including large scale numerical simulation, hybrid computational schemes, front tracking, adaptive mesh refinement, optimized parallel computation, asymptotic methods and singular perturbation techniques, bifurcation theory, optimization methods, dynamical systems theory, cellular automata and discrete methods and probabilistic and statistical methods. Experimental studies that employ intrusive or nonintrusive diagnostics and are published in the Journal should be closely related to theoretical issues, by highlighting fundamental theoretical questions or by providing a sound basis for comparison with theory.