Experimental investigations on flame characteristics and combustion mode recognition of lean blowout with strut/cavity flame stabilizers

The precise and real-time monitoring of the combustion state is essential for attaining a stable combustion organization in scramjet engines. Experimental studies are performed to investigate the flame characteristics and combustion mode during the transition process to lean blowout in a kerosene-fueled strut/cavity combustor. Heated air enters the engine with a Mach number of 3.07, total pressure of 2.2 MPa, and total temperature of 1670 K. A novel method for the combustion mode recognition is proposed, based on a miniaturized and lightweight photoelectric sensor array. The response characteristics of photoelectric signals, flame luminosity and pressure are analyzed throughout the processes of flame propagation, stabilization and lean blowout. The characteristics of flame oscillation and transition for the critical lean blowout can be accurately captured through photoelectric measurements. Based on the results observed, the evolution processes of lean blowout can be divided into four distinct stages: intensive combustion, weak combustion, flame flashback, flame lift-off and complete blowout. The weak combustion mode stage contains severe pressure oscillation and can subsequently give rise to the flame flashback, which represents an abnormal operating condition. During this process, the photoelectric signal in the core flame stabilization region at the trailing edge of the cavity decreases from 40.47 % to 8.31 %. In contrast, the pressure signal does not exhibit a significant decline. Thus, the photoelectric signal can recognize the transition of the weak combustion mode at an earlier point in time than the pressure. The main reasons are that the reduction of the pressure potential field during flame extinction is controlled by the cooling and convection of the expanding hot gas. This process is slower than the cessation of the combustion reaction. Then, the experiments of combustion mode recognition and closed-loop fuel injection strategy control are performed to validate the combustion state monitoring method based on photoelectric signals. The results show that it can allow the combustor to revert to a stable and efficient intensive combustion mode and achieve the desired combustor pressure value. The entire closed-loop control process is completed within a total duration of 31 milliseconds. The ultimate objective of achieving efficient and stable engine operation is successfully accomplished, thereby reducing the risk of significant thrust loss and abnormal shutdown. The reliability of the combustion state recognition method presented in this study is further validated through the implementation of multiple repetitive experiments.