The aim of the whole work done for this thesis consist in the correction of rigid CFD simulations with the data coming from the pressure transducers positioned on the right wing. This correction hides inside the pressure variations arising from all the possible contributions like, for example, structure deformations and different Cp trends along the surfaces with respect to that expected by the numerical solver. Taking the data recorded during the flights and handling them to obtain continuous corrections on the affected surfaces, the final results will be some new files, structured like the original ones, containing the corrected pressure coefficient distributions. To obtain a continuous correction on the whole surfaces (not only at the positions of the sensors) some interpolation methods will be used: they are first of all presented and tested in dedicated chapters. The interpolators that will be used on the wing are neural networks and kriging because both showed the best results during the test procedure. All the cases that will be corrected are flight conditions requested by the company for a more in-depth analysis of the loads. The last part of the thesis concerns the interpolation of the sensors pressure coefficients recorded during the flights (only for symmetrical maneuvers with a null flaps deflection), using again neural networks and kriging: after this procedure they will be available in the whole domain defined by the flight envelope of the M-345 aircraft. In this way, corrections, even in conditions not really flown during the dedicated flight campaign, will be obtained applying the same method described before.
L'obiettivo di tutto il lavoro svolto per questa tesi consiste nella correzione di simulazioni CFD rigide con i dati provenienti da alcuni trasduttori di pressione posizionati sull'ala destra. Questa correzione nasconde all'interno delle variazioni di pressione derivanti da tutti i possibili contributi come, ad esempio, deformazioni della struttura e andamenti di Cp, sulle superfici coinvolte, diversi rispetto a quanto previsto dal risolutore numerico. Prelevando i dati registrati durante i voli e manipolandoli per ottenere una correzione continua sulle superfici interessate, i risultati finali saranno alcuni nuovi file, strutturati come quelli originali, contenenti le distribuzioni corrette dei coefficienti di pressione. Per ottenere una correzione continua su tutte le superfici (non solo nelle posizioni dei sensori) saranno utilizzati alcuni metodi di interpolazione: questi sono prima di tutto presentati e testati in capitoli dedicati. Gli interpolatori che verranno utilizzati sull'ala sono reti neurali e kriging perché entrambi hanno mostrato i migliori risultati durante la procedura di test. Tutti i casi che verranno corretti sono delle condizioni di volo richieste dalla azienda per un'analisi più approfondita dei carichi. L'ultima parte della tesi riguarda invece l'interpolazione dei coefficienti di pressione registrati durante i voli dai sensori (solo per manovre simmetriche con deflessione dei flap nulla) utilizzando ancora una volta reti neurali e kriging: dopo questa procedura saranno disponibili in tutto il dominio definito dall'inviluppo di volo del M-345. In questo modo, le correzioni, anche in condizioni non realmente volate durante la campagna di voli dedicata, si potranno ottenere applicando lo stesso metodo precedentemente descritto.
A new approach for matching and validation of CFD results by pressure flight acquisition and its interpolation
PIZZETTI, EDOARDO
2018/2019
Abstract
The aim of the whole work done for this thesis consist in the correction of rigid CFD simulations with the data coming from the pressure transducers positioned on the right wing. This correction hides inside the pressure variations arising from all the possible contributions like, for example, structure deformations and different Cp trends along the surfaces with respect to that expected by the numerical solver. Taking the data recorded during the flights and handling them to obtain continuous corrections on the affected surfaces, the final results will be some new files, structured like the original ones, containing the corrected pressure coefficient distributions. To obtain a continuous correction on the whole surfaces (not only at the positions of the sensors) some interpolation methods will be used: they are first of all presented and tested in dedicated chapters. The interpolators that will be used on the wing are neural networks and kriging because both showed the best results during the test procedure. All the cases that will be corrected are flight conditions requested by the company for a more in-depth analysis of the loads. The last part of the thesis concerns the interpolation of the sensors pressure coefficients recorded during the flights (only for symmetrical maneuvers with a null flaps deflection), using again neural networks and kriging: after this procedure they will be available in the whole domain defined by the flight envelope of the M-345 aircraft. In this way, corrections, even in conditions not really flown during the dedicated flight campaign, will be obtained applying the same method described before.| File | Dimensione | Formato | |
|---|---|---|---|
|
A new approach for matching and validation of CFD results by pressure flight acquisition and its interpolation.pdf
non accessibile
Descrizione: Testo della tesi
Dimensione
44.81 MB
Formato
Adobe PDF
|
44.81 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/148371