Research On Flame Temperature Spectrum Measurement Method Based On Area Array Sensor

The instability of the combustion field has always been an insurmountable challenge in the design of combustion chambers such as aviation engines and gas turbines.Fine measurement helps to accurately analyze combustion conditions,optimize fuel ratios,reduce the generation and emissions of pollutants,and has important value in engineering applications.Traditional temperature measurement methods can easily affect combustion characteristics in fine flame measurement.In non-contact temperature measurement,Tunable Diode Laser Absorption Spectroscopy（TDLAS）technology has high sensitivity,fast time response,and small disturbance,and has become an important tool for combustion diagnosis in internal combustion engines,ramjet engines,and other fields.However,the flame temperature measurement method based on TDLAS has the problem of low resolution,which makes it difficult to meet the needs of fine measurement and even more difficult to reconstruct the three-dimensional flame temperature field.Therefore,this dissertation proposes a flame temperature measurement method using array sensors coupled with TDLAS technology.Based on a 64 pixel array sensor,utilizing wavelength modulation spectroscopy technology and algebraic reconstruction algorithms,two-dimensional temperature measurement of flames is achieved;Based on a planar CMOS image sensor,direct absorption spectroscopy and filtered backprojection algorithm were used to achieve three-dimensional temperature measurement of flames;a new method for synchronously measuring flame temperature using active absorption spectroscopy and passive radiation spectroscopy has been achieved by combining Lambert Beer’s law and blackbody radiation law.On this basis,a U-Conv LSTM temperature distribution reconstruction prediction model combining a U-shaped and Convolutional Long Short Term Memory（Conv LSTM）network is constructed to effectively solve the time lag problem of photoelectric temperature measurement systems.The main tasks are as follows:1.Aiming at the temperature measurement mechanism and optical model of the characteristic gas phase molecule oxygen（O₂）absorption spectrum in the combustion field,the sensing mechanism of gas molecular temperature in combustion field was analyzed,combined with High-resolution Transmission Molecular Absorption Database（HITRAN）,select the optimal characteristic spectral line of O₂ at high temperature,obtain the spectral line intensity,spectral line shape function and low energy level energy of gas phase molecules and other related parameters,and finally determine the O₂ molecules at 13142.5cm^-1 and 13144.6 cm^-1,the two characteristic spectral lines are used as temperature measurement spectral lines,which provide a theoretical basis for subsequent measurement of flame temperature using a photoelectric temperature measurement system.At the same time,based on Lambert-Beer’s law and blackbody radiation law,the optical model of laser and flame thermal and optical radiation transmission in the combustion field was studied,laying the foundation for multi-dimensional imaging of high-temperature fields and dual-mode simultaneous measurement of flame temperature with active absorption spectroscopy and passive radiation spectroscopy.2.Aiming at the problem of high-precision imaging of flame temperature in the constructed photoelectric temperature measurement system,two flame temperature measurement methods coupled with TDLAS technology using array sensors are proposed.Firstly,the effects of temperature,concentration,and pressure on absorbance were simulated and analyzed.Then,a photoelectric temperature measurement system based on a 64 pixel array sensor and an array CMOS image sensor was constructed,and the relationship between the integrated absorbance and temperature of the two photoelectric temperature measurement systems under a standard high-temperature furnace was calibrated.Using wavelength modulation spectroscopy technology,a 64 pixel array sensor was used to obtain a cylindrical laser beam passing through the combustion field,and the two-dimensional temperature field of the flame was reconstructed.Using direct absorption spectroscopy,a conical laser beam passing through the test field was collected using a planar CMOS image sensor to reconstruct a higher spatial resolution three-dimensional temperature field of the flame.The two reconstructed temperature fields were compared using high-precision thermocouples,and the measurement error was less than 6%.3.A detection system based on an array CMOS image sensor was designed to address the synchronization issue of flame temperature measurement verification technology in detection systems.The system simultaneously measures the flame temperature distribution through active absorption spectroscopy and passive radiation spectroscopy.Firstly,an energy transfer optical model for synchronous measurement of flame temperature using active absorption spectroscopy and passive radiation spectroscopy was established.Then,using a standard blackbody light source,calibration experiments were conducted on the front array CMOS image sensor to determine the corresponding grayscale values at different temperatures.Finally,based on the measurement mechanism of characteristic gas-phase molecular absorption spectra and element doped flame temperature radiation spectra,the corresponding grayscale values were determined,Select characteristic spectral lines of doped element K⁺for bimodal spectral measurement of flame temperature distribution.The experimental results show that the temperature imaging difference between the two spectra is less than 8%,which verifies the feasibility of the flame temperature measurement verification synchronization method in the detection system,improves the reliability of the measurement results,solves the time synchronization problem of measuring the true temperature value in the dynamic combustion process,and reduces the complexity and cost of verification.It provides a new research approach for developing high-precision and low-cost flame temperature measurement verification technologies.4.A fast spatiotemporal prediction algorithm for the two-dimensional distribution of flame temperature field is proposed to address the issue of time lag in flame temperature imaging.A U-Conv LSTM model was designed to reconstruct the combustion field and achieve short-term prediction.By dividing the combustion field into spatiotemporal slices,using discretized spatiotemporal slices to reconstruct the two-dimensional temperature distribution,and then expanding it to higher dimensions.The simulation results show that the PSNR of the three-step prediction accuracy is not less than 30 d B,and the SSIM is greater than 0.75,which proves the feasibility of the U-Conv LSTM spatiotemporal sequence prediction algorithm in temperature field distribution prediction and provides technical support for achieving closed-loop control systems of combustion fields.