Logo Logo
Help
Contact
Switch language to German
Property prediction of energetic materials with the “Research output software for energetic materials based on observational modelling” RoseBoom©
Property prediction of energetic materials with the “Research output software for energetic materials based on observational modelling” RoseBoom©
Since 2020, which is when the author of this thesis collected her first experiences with energetic materials 16.500 articles were published and registered by google scholar on “synthesis of new energetic materials”. This level of productivity and access to vast amounts of information was previously unheard of. In the early 1980s, home computers were just becoming possible and very few people had mobile phones. Researchers now face the challenge of sifting through a vast amount of information, with the crucial questions being about its reliability and practical use. This thesis presents several innovations that enable researchers to access information from the literature in a selective and automated way. These innovations also allow for comparison using established models, leading to informed decisions made by experts. The first chapter of this research serves as a comprehensive introduction to the concept of RoseBoom© and provides two examples of the application of RoseBoom© reported in literature. Chapter two delves into the development of RoseBoom©, a program that encapsulates the key discoveries of this research. The subsequent chapters consist of papers that provide detailed explanations of essential aspects of the study. Finally, the last chapter explores the wider scope of opportunities for advancing this field. In this thesis machine learning models, empirical models, and thermo-equilibrium codes are thoroughly tested and evaluated for the prediction of energetic materials. The limits and advantages of each method are carefully evaluated and should be considered. In Chapter 1.6, the experimental situation is assessed. An overview of various measurement techniques for detonation parameters is provided, along with recent research on using modeling tests with simpler experimental setups as an alternative method. Furthermore, the deviations in experimental measurements of detonation pressures have been analyzed. Chapter 1.7 provides a comparison of various unclassified software solutions for energetic materials, including RoseBoom©. These solutions compete in eight categories and are ranked based on the points they receive. The fundamental concept of RoseBoom© is presented in Chapter 2.3. A thorough evaluation of empirical models for energetic materials presented in the literature is given in Chapter 2.1., which was revalidated for novel energetic materials. An update for performance prediction for mixtures is given in Chapter 2.2 along with the automated input of large molecule datasets from .csv files. In Chapter 2.7 and 4.2 currently available open-source chemical structure recognition tools are investigated for implementation in RoseBoom© which further improve the user’s experience. An update to the software for rocket propellants is given in Chapter 2.4, where the specific impulses of the ISPBKW code to two empirical models are compared. Including the application programming interface to the ISPBKW code, which allows it to be easily accessed using the RoseBoom©. In Chapter 2.5, the impact on the calculated detonation parameters was investigated by comparing the use of density and heat of formation predicted by RoseBoom2.2© to those published with corresponding molecules. A range of traditional models was tested for sensitivity to input value accuracy. This highlights the need for agreement on one software for predicting energetic material performance, starting with input values. It also increases trust in RoseBoom© predictions while raising awareness of uncertainties in published performance values. This motivated the author to conduct a study in chapter 2.6 investigating the prediction of enthalpies of sublimation and vaporization, as they are required to convert an enthalpy of formation value obtained from a gas phase calculation into a room temperature state. The study presented in Chapter 1.5 aimed to determine whether complex machine learning models are necessary to predict material properties or if simple linear regression models can provide accurate predictions of thermochemical properties and density. The study analysed Joback's method in combination with statistical models, as well as the density predictions of Holden, Keshavarz, and Bondarchuk. Updated group increment tables for Joback's method were also included in the analysis. In the same chapter, a correlation between the plate dent test and Chapman-Jougett detonation pressure is presented, which would be an excellent candidate for lab-scale performance testing of novel energetic materials. Further studies like this are presented in Chapter 3 including the Ballistic Mortar test, the Trauzl Test, and the SSRT-Test. Chapter 4.1 is an example where some of the computational methods used in this thesis were applied to real-life problems.
Not available
Wahler, Sabrina
2023
English
Universitätsbibliothek der Ludwig-Maximilians-Universität München
Wahler, Sabrina (2023): Property prediction of energetic materials with the “Research output software for energetic materials based on observational modelling” RoseBoom©. Dissertation, LMU München: Faculty of Chemistry and Pharmacy
[thumbnail of Wahler_Sabrina.pdf]
Preview
PDF
Wahler_Sabrina.pdf

33MB

Abstract

Since 2020, which is when the author of this thesis collected her first experiences with energetic materials 16.500 articles were published and registered by google scholar on “synthesis of new energetic materials”. This level of productivity and access to vast amounts of information was previously unheard of. In the early 1980s, home computers were just becoming possible and very few people had mobile phones. Researchers now face the challenge of sifting through a vast amount of information, with the crucial questions being about its reliability and practical use. This thesis presents several innovations that enable researchers to access information from the literature in a selective and automated way. These innovations also allow for comparison using established models, leading to informed decisions made by experts. The first chapter of this research serves as a comprehensive introduction to the concept of RoseBoom© and provides two examples of the application of RoseBoom© reported in literature. Chapter two delves into the development of RoseBoom©, a program that encapsulates the key discoveries of this research. The subsequent chapters consist of papers that provide detailed explanations of essential aspects of the study. Finally, the last chapter explores the wider scope of opportunities for advancing this field. In this thesis machine learning models, empirical models, and thermo-equilibrium codes are thoroughly tested and evaluated for the prediction of energetic materials. The limits and advantages of each method are carefully evaluated and should be considered. In Chapter 1.6, the experimental situation is assessed. An overview of various measurement techniques for detonation parameters is provided, along with recent research on using modeling tests with simpler experimental setups as an alternative method. Furthermore, the deviations in experimental measurements of detonation pressures have been analyzed. Chapter 1.7 provides a comparison of various unclassified software solutions for energetic materials, including RoseBoom©. These solutions compete in eight categories and are ranked based on the points they receive. The fundamental concept of RoseBoom© is presented in Chapter 2.3. A thorough evaluation of empirical models for energetic materials presented in the literature is given in Chapter 2.1., which was revalidated for novel energetic materials. An update for performance prediction for mixtures is given in Chapter 2.2 along with the automated input of large molecule datasets from .csv files. In Chapter 2.7 and 4.2 currently available open-source chemical structure recognition tools are investigated for implementation in RoseBoom© which further improve the user’s experience. An update to the software for rocket propellants is given in Chapter 2.4, where the specific impulses of the ISPBKW code to two empirical models are compared. Including the application programming interface to the ISPBKW code, which allows it to be easily accessed using the RoseBoom©. In Chapter 2.5, the impact on the calculated detonation parameters was investigated by comparing the use of density and heat of formation predicted by RoseBoom2.2© to those published with corresponding molecules. A range of traditional models was tested for sensitivity to input value accuracy. This highlights the need for agreement on one software for predicting energetic material performance, starting with input values. It also increases trust in RoseBoom© predictions while raising awareness of uncertainties in published performance values. This motivated the author to conduct a study in chapter 2.6 investigating the prediction of enthalpies of sublimation and vaporization, as they are required to convert an enthalpy of formation value obtained from a gas phase calculation into a room temperature state. The study presented in Chapter 1.5 aimed to determine whether complex machine learning models are necessary to predict material properties or if simple linear regression models can provide accurate predictions of thermochemical properties and density. The study analysed Joback's method in combination with statistical models, as well as the density predictions of Holden, Keshavarz, and Bondarchuk. Updated group increment tables for Joback's method were also included in the analysis. In the same chapter, a correlation between the plate dent test and Chapman-Jougett detonation pressure is presented, which would be an excellent candidate for lab-scale performance testing of novel energetic materials. Further studies like this are presented in Chapter 3 including the Ballistic Mortar test, the Trauzl Test, and the SSRT-Test. Chapter 4.1 is an example where some of the computational methods used in this thesis were applied to real-life problems.