AI
AI

Neural Network Enhances Quantification Accuracy in Tunable Diode Laser Absorption Spectroscopy

Photo credit: phys.org

Advancements in Absorbance Recovery for Enhanced TDLAS Measurements

A team of researchers from the Hefei Institutes of Physical Science (HFIPS) at the Chinese Academy of Sciences has pioneered a new method utilizing neural networks to enhance the accuracy of single-path tunable diode laser absorption spectroscopy (TDLAS). This innovative approach is aimed at addressing the common challenges associated with measurement errors in existing TDLAS methodologies.

The findings of their research were published in the journal Fuel, emphasizing the significance of accurate combustion flow field temperature and concentration measurement. Such data are critical for the design, monitoring, and diagnosis of advanced combustion systems, offering insights that go beyond traditional measurement techniques.

TDLAS is praised for its speed, high sensitivity, and resistance to interferences. However, the conventional single-path approach has been hindered by substantial measurement errors, largely due to baseline distortion affecting absorbance readings. Prof. Liu Kun, a key researcher on the project, noted that “the main challenge involves correcting baseline errors that impact absorbance measurements.”

The research team identified that the derivative of absorbance is more responsive to changes in the line shape curvature, particularly near the absorption peak where baseline error-induced variations are minimal. Building on this understanding, the HFIPS researchers, under the guidance of Prof. Gao Xiaoming and Prof. Liu Kun, devised a model capable of extracting the absolute absorbance profile from the derivative data.

This new absorbance recovery method underwent rigorous testing through simulations and practical temperature assessments. It was applied effectively in measuring exhaust temperature and water concentration in a small diesel turbojet engine, yielding impressive results with an error margin of only 0.9% when compared to traditional thermocouple readings.

“Our results present a promising technique for enhancing the accuracy of TDLAS measurements, and it can be seamlessly integrated into existing tomographic absorption spectroscopy systems,” stated Prof. Gao Xiaoming.

More information: Ruifeng Wang et al, Measurement of engine exhaust plume temperature and concentration distributions with tomographic absorption spectroscopy and learning-based absorbance recovery, Fuel (2024). DOI: 10.1016/j.fuel.2024.132775

Citation: Neural network improves tunable diode laser absorption spectroscopy quantification accuracy (2024, September 12) retrieved 12 September 2024 from https://phys.org/news/2024-09-neural-network-tunable-diode-laser.html

This content is provided for informational purposes only, subject to copyright restrictions. No part may be reproduced without prior written permission.

Source
phys.org

Related by category

Innovative 3D Display Allows Interaction with Virtual Objects

Photo credit: www.sciencenews.org Floating displays that allow users to reach...

Exploring Ways to Harness Electricity from Rainfall

Photo credit: www.sciencenews.org A Revolutionary Way to Utilize...

New Atomic Fountain Clock Joins Elite Ranks of Global Timekeeping

Photo credit: phys.org Thanks to NIST-F4, a newly developed atomic...

Latest news

Kolkata Hotel Fire Claims at Least 14 Lives, According to Police

Photo credit: www.cbsnews.com New Delhi — A devastating fire engulfed...

Raphinha Transforms from Unsung Hero to Ballon d’Or Contender for Barcelona

Photo credit: www.theguardian.com Raphinha: A Journey Through Missed Opportunities and...

An Existential Moment: Greens Challenge Reform for Disenchanted Voters

Photo credit: www.theguardian.com With its picturesque thatched cottages and rural...

Breaking news