***Investigation of the Effects of Steam Injection on NOx Emissions in Hydrogen-Methane Blend Flames***

Author: Doruk Çelebi

TU Delft, Faculty of Aerospace Engineering, Flight Performance & Power and Propulsion

MSc Thesis Report

Contact Information: doruk828@hotmail.com


*** General Information***

This dataset contains the data collected during the combustion experiments conducted at HSL facility in TU Delft, as a part of Doruk Çelebi's MSc garduation thesis. (October 2025)
Link to thesis paper: https://resolver.tudelft.nl/uuid:1a09c904-57aa-46f0-abec-899111f67625


It is being made public both to act as supplementary data and in order for other researchers to use this data in their own work


The data in this data set were collected at the HSL lab in TU Delft, between 13.06.2025-26.06.2025

***Purpose of the test campaign***
The purpose of the experiments was to investigate the effect of steam injection in H2-CH4 flames on the NOx emissions, using gas analyzer emission data. A variety of load cases were tested with varying power, h2-blend fraction, steam injection amount and steam injection location


***Test Equipment***
All experimental data was gathered using ABB2000 gas analyzer. Calibraion using reference calibration gases were conducted at the beggining of each experimental day to mitigate drift of the readings. 
Gas analyzer readings include species of: NO, NO2, CO, CO2, O2 and CH4. It was found during the experiments that steam injection intereferes greatly with the NO2 measurement module, and caused unsteady and unrealiable readings, namely sudden spikes and drops during measurements, with more than expected mean NO2 readings. Hence NO2 data is unsuable for steam injection cases. The NO2 readings seemed mostly reasonable for no-steam cases, but regardless comparison against steam injection cases is not possible. (refer to paper for potential reasons for this, "Data Processing" chapter)
Sampling by the gas analyzer was done at 10 Hz

Units of the measurements are:
NO in ppm
NO2 in ppm
CO2 in %vol
O2 in %vol
CH4 in ppm
CO in ppm



***Description of the experiments***

The experiments were conducted over different days, each day focusing on a different set of variable to be changed:
The tabs in the excel file are named after the day of the experiments:

13-06-25: 2 set of experiments were conducted. First one is 11 kW constant power burner with varying hydrogen blend fraction. Second one is pure hydrogen combustion with varying burner power. Both experiments used no-steam vs steam-9 comparisons. Side injection configuration was used in both experiments.
18-06-25: 2 set of experiments were conducted. First one is full methane combustion with varying burner power. Second one is 9 kW constant power burner with varying hydrogen blend fraction. Both experiments used no-steam vs steam-9 comparisons. Side injection configuration was used in both experiments.
20-06-25: 1 set of experiment was conducted. 10 kW constant burner power with varying hydrogen blend fraction. This experiment used no-steam vs steam-9 vs steam-7 comparisons. Side injection configuration was used.
23-06-25: 1 set of experiment was conducted. This experiment was a repeat of 20-06-25 for data sanity check.
25-06-25: 1 set of experiment was conducted. Injection location was changed to axial steam injection (as described in the paper). 10 kW constant burner power with varying hydrogen blend fraction was tested. This experiment used steam-6,7,8 and 9.
26-06-25: 1 set of experiment was conducted. Side Steam Injection was once again used. 10 kW constant burner power with varying hydrogen blend fraction. This experiment used steam-6,7,8 and 9.

Steam injection amounts correspond to:
Steam-9: 24.4 g/min
Steam-8: 22.8 g/min
Steam-7: 17.7 g/min
Steam-6: 13.7 g/min


***Description of the data in the data set***

Gas analyzer data is compiled in an excel table for each data point. Gas analyzer data is extracted as tdms files which contains the data recorded over the 60-100 seconds. The data listed here are processed data from those files.
The filename column contains the details regarding the configuration used in the test point, with indication total air injection amount in nlpm, burner power in kW, axial air injection (AAI) fraction, H2 blend percentage, and whether steam is injected or not. For example:

2025_06_20_11_18_58250mainair-10kW-0aai-30h2.tdms file tested the point with 250 nlpm total air injection, 10 kW burner power, 0% AII, and 30% H2 blend with no steam injection
2025_06_18_15_58_09160mainair-7.5kw-0aai-0h2-steamp9.tdms file tested the point with 160 nlpm total air injection, 7.5 kW burner power, 0% AAI and 0% H2 with steam injection level of steam-9

Several statistical tools are used to in the analysis. Mean values of all the species are extracted to be used in the analysis. 
To determine the reliability of the data, other statistical tools are also utilize and reported over the sampled time sampling frequency of 10 Hz, such as the standard deviation, range,, coefficient of variation (CV) and absolute CV.

These statistical tools acted as signals to indicate whether something was wrong with the recorded data. "Bad" data with too high CV, standard deviation and range were discarded. Sometimes, data was fixed manually to remove any spikes observed in the readings, hence any outlier samplings from reader artifacts were mitigated.

The "bad" data is highlighted with dark red.

In each sheet below the file data, the compliation of the results are presented, with manual fixes included.


The rest of the tables and sheets are calculations to verify and sanity check values.

