| Περίληψη: | The current dissertation, titled “Fire Response of Composite aerostructures” deals with a crucial subject of the aeronautics industry that is the fire response of composite aerostructures, more specifically the issue of interest in this work is the fuselage fire burnthrough from an external liquid jet-fuel pool fire. Other fire issues that “bother” the aeronautics industry are the fire spread inside the cabin, smoke generation and toxicity of the fumes, but these are not handled in the current dissertation.
Aircraft structures are designed to withstand various loading scenarios during their operational life. These loading scenarios are associated to a great extent with normal aircraft operation (flight manoeuvres, take-off and landing). However there are situations where the aircraft structures are required to assure the safety of the passengers and crew. In the case of an emergency crash landing, the threat of an external jet-fuel fire always exists. Considering that the aircraft structure survives the impact, the survivability of the passengers and crew onboard the aircraft depends solely on the fire resistance of the aircraft structure. A measure of the fire resistance of an aircraft structure is the time needed for the flames to penetrate the fuselage and spread inside the cabin, the so-called, burn-through time.
So far, the aircraft fire resistance has been extensively studied by conducting lab, medium and full scale tests. The early lab scale tests were performed by the Federal Aviation Administration (FAA) and involved the Bunsen-burner flammability test of coupons for developing fire safe interior materials. As the application of polymer materials on aircrafts kept increasing, the problem of fire burn-through due to external fire emerged. Marker was one of the first to perform full-scale fuselage burn-through tests to access the insulating performance of materials. Also a statistical analysis was performed by Cherry and Warren that accessed and analyzed data from past accidents and their work resulted in proving the importance of fuselage fire hardening and the passengers’ lives that could be saved using low-cost solutions. These works led the FAA to proposed new fire testing procedures for aircraft materials.
The scope of this dissertation was to assess the performance of various structural materials in a pool-fire scenario. A simplified approach is made, approximating the pool-fire conditions with a flat panel burn-through test in accordance to the ISO2685:1998(E) Standard.
The originality of the present work comes from the fact that it incorporates a multistage approach in order to investigate the behaviour and response of composite aircraft structures in the possibility of a fire event. The current approach goes down on material level in order to investigate and model the deterioration (decomposition) of the polymer composite. Thus, it investigates and proposes a methodology of how the thermophysical properties of the composite are deteriorated due to the fire event. It proceeds into developing a progressive-damage material model (material properties varying with the deterioration degree) and finally implementing this custom material model into a commercial FE package and solving the loading scenarios.
Being more specific the current work begins with a quick review of the literature where incidents and work done on the burnthrough event for the past 20-30 years are summarized. It progresses then to presenting the various types of polymers used in the aircraft industry and their basic decomposition mechanisms, from the unsaturated polyesters to the epoxies and phenolics and in the end reference to the thermoplastics is made. Every organic material, hence, polymers used in aerospace applications, present a set of response characteristics when subjected to fire, specifically the heat release rate, thermal stability index, limiting oxygen index, flammability index, time-to-ignition, surface flame spread, mass loss, smoke density and smoke toxicity.
Following is the backbone of this dissertation, the kinetics modelling. Two approaches are made, one simplified using single stage kinetics where the decomposition degree a is calculated based on the Arrhenius reaction theory and using the kinetic triplets (kinetic parameters) extracted from thermogravimetry, TGA, data using the Friedman multi-curve method. The second approach is more complicated and considers multi-stage decomposition of the polymer composite. Specifically a 3-stage reaction network is considered for every material, the LY-Ref, and the two modified batches, one with ammonium polyphosphate AP423 and the other both with AP423 and multi-wall carbon nanotubes MWCNT. Again the kinetic parameters, activation energy EA, frequency factor A, and reaction order n, are extracted for every step using the van Krevelen methodology. In the end using the reaction rates equations the reconstruction of the TGA curves is achieved with an error of less than 5% from the test data. Correlations that consider the material deterioration and affect the thermophysical properties of the materials are proposed. Those expressions are being developed for both of the two kinetic approaches, the single and multi stage.
Another crucial part of this work is the measurement and calibration of the applied fire load. Again two fire load approaches are used, one according to the ISO2685 Standard where a propane burner was manufactured and calibrated according to the Standard for medium scale samples testing and a lab scale butane burner for small samples. The ISO2685 burner was also CFD simulated and the models calibrated against analytical expressions, ISO requirements and real measurements. The CFD simulations were performed so the heat flux or heat transfer coefficient to be extracted and used as input for the later thermal FE burnthrough models. The heat flux distribution of the lab-scale AML burner on the specimen surface was measured via a water cooled Schmit-Boelter SBG01 heat flux sensor manufactured by Hukseflux.
Manufacturing and material details are presented concerning the samples used for every test campaign. Metallic (AL2024-T3) samples, CFRP neat and modified, and hybrid GLARE ones where manufactured. Also the experimental work performed is described. Cone calorimetry testing data are available, results from thermogravimetry tests, differential scanning calorimetry, and finally the burnthrough tests with both the testing apparatuses, the ISO2685 one and the AML lab-scale burner.
The modelling work in this dissertation involved thermal models that were developed into a commercial FE package. It was not part of this work to develop a thermal solver so a commercial one was selected and all the developed methodology was adapted to its requirements and specifications. The boundary conditions on the models are presented both for the ‘hot’ front surface and the rear ‘cooling’ one. For the ‘hot’ one the heat flux distribution is used and for the ‘cooling’ one an equivalent convection is applied that accounts for both convective and radiative cooling. The decomposing material model is implemented into to FE solver via user defined subroutines for the single stage kinetics and the multi-stage approach. Finally the simulations were run and the results and models were compared against the available experimental results.
Since so far the burnthrough response of aerostructures was limited to coupon, samples and medium size flat panels. A more realistic approach was performed by developing a mathematical model of a real size test. The certification tests conducted by the FAA are for full size fuselage sectors under the fire load of a burning jet-fuel pan pool-fire. A burning jet-fuel pool fire is a complex phenomenon on its own, combining it with a decomposing fuselage structure make the modeling approach even more difficult to simulate if not impossible. Required data for the pool-sizes under investigation were not available, so data for large external hydrocarbon pool fires from literature were used. Also, because the main characteristic of a jet-fuel (kerosene) pool fire is that the flames are not clear, on the contrary, great amount of shoot is produced making combustion modeling and radiative heat transfer to the fuselage even more of a challenge to model, it was decided to try and tackle this full-scale approach by a simplified the modeling approach. Instead of liquid fuel combustion, an equal hot air stream with mass flow, velocity and temperature properties extracted from literature correlation data was performed.
Conclusively, in terms of completeness the benefit analysis performed by Cherry and Warren is presented in brief. The objective of their analysis was to assess the potential benefits, in terms of reduction of fatalities and injuries, resulting from improvements in fuselage burnthrough resistance to ground pool fires. Fire hardening of fuselages will provide benefits in terms of enhanced occupant survival and may be found to be cost beneficial if low-cost solutions can be found. The maximum number of lives saved per year in worldwide transport aircraft accidents, over the period covered by the data, if hardening measures were applied, was assessed to be 12.5 for the aircraft in its actual configuration (when the accidents occurred) and 10.5 for the aircraft configured to later airworthiness requirements.
These figures are completely significant and give an extra confirmation that this work on investigating the fire response of composite aerostructures is on the right track. As the work of Cherry and Warren concluded, the fire hardening measures in order to be applicable need to be cost efficient. The concept under which this whole dissertation stepped on was to investigate the fire response of composite aerostructures and the possibility of hardening the structure itself without the use of extra protective layers that add cost and weight to the overall aircraft and its maintenance. In the end it was concluded that there is the possibility of hardening the fuselage structure by design and by material. Incorporating composites into the structure it is possible to prolong the burnthrough time at least for 4-5 minutes before auto ignition occurs on the inner side of the fuselage. Auto ignition of the inner side fuselage cabin materials is mentioned since in NONE of the burnthrough tests of the CFRP composites and the GLARE samples flame penetration was observed.
|
|---|
