This project starts from the need, in the industrial field, to avoid or at least limit, the stall phenomena and its harmful effects on the lifespan of rotating turbomachines. The significance of the stall phenomena depends by the issues rising as a result of the machine continuous working in the unsteady areas of the characteristic curve plane, beyond the stall limit, for prolonged periods or in a persistently oscillating way. During stall the stall cell formed on the blades, moves from a blade to the next in the opposite direction in respect to the blades rotating direction. The moving stall cell, besides leading to a reduction of the performance due to the annulus obstruction (and flow rate reduction) induces vibrations that are transmitted to the whole system. Such vibrations induced by the rotting stall represent a limit of fundamental importance on the performance of the rotating turbomachines, especially on compressors and fans, for whom the rotating stall is a persistent issue in design and development. The comprehension of the key physical phenomena that manifest during stall and before (stall inception) have a critical role in the attempt to contain and avoid stall, already in the design phase or through active and passive control methods. Traditionally the methods used to contrast stall are based on the passive control techniques. This is true especially in the industrial field. The passive control techniques operate a contrast to rotating stall by influencing the flow, since they intervene on the various factors that may lead to a stall recovery. In more recent years, thanks to the technological progress, the passive control techniques are giving way to the active control techniques. These last with the aid of high-response, high-sensitivity sensors, combined with the available increased computational power, have allowed to develop systems of control, early detection and warning, that have the ambition of preventing the damaging phenomena bound to the stall occurrence, if not even prevent the stall itself, recognizing the signals that announce its impending. The high economical and computational cost related to the high performance of instrumentation and analysis methods makes these technologies less attractive, industrially speaking. From these considerations is born the idea to develop a measurement chain for stall detection based on DIY (Do It Yourself) sensors. The unconventional sensors have been used to measure the pressure instabilities as pseudo-sound signals in the near field, on board of an industrial axial fan on which stall was forced. The aim was to develop and set-up a measurement system able to acquire the low frequency signals typical of stall with the aid of sensors that have typically a narrow range of acquiring frequency, if compared with more advanced and evolved technical solutions. The final aim was the realization of robust and reliable stall warning system. The data acquired during stall have been analysed with the traditional methods of the spectral auto-correlation, that although the sensors limits about the acquisition frequency, provided a first validation of the goodness of the used measurement system. In addition it has been used an analysis system based on the comparison of the patterns provided by a phase space reconstruction inspired technique. The reconstructed phase space is a widely used technique in the study of chaotic dynamic systems, and the rotating stall totally fits in this class. Finally it has been investigated the possibility to detect a secondary phenomena, such as fouling, and its influence on the rotating stall, again using the same DIY measurement chain and the above mentioned analysis methods. The fouling on the rotating blades is caused by the presence, in the flow passing through the machine, of particles of dust, pollen or dirt that adhere to the blades forming a deposit on them. This deposit is an obstacle to the passing flow and alters the rotor geometry leading to unbalance problems, instabilities and moving the stall limit. By means of the realized stall detection system it has been possible to identify the rotating stall and recognize the presence of fouling, since there was a modification in the stall dynamic and stall precursors were identified. The experimental test demonstrated that the realized system represent a candidate solution for the detection of patterns typical of rotating stall and fouling in turbomachines.
Questo studio prende il via dalla necessità in ambito industriale di evitare o perlomeno limitare il fenomeno dello stallo e i suoi effetti nocivi sulla vita delle turbomacchine rotanti. L'importanza di questo fenomeno dipende dalle problematiche che insorgono a seguito del continuo operare della macchina in zone instabili della curva caratteristica, oltre il limite di stallo per periodi prolungati o in maniera oscillatoria persistente (passando frequentemente dalla zona instabile alla zona stabile del campo operativo). Durante lo stallo la cella di stallo che si forma a ridosso della palettatura, spostandosi da una pala alla successiva, in direzione opposta al senso di rotazione, oltre a comportare una diminuzione delle prestazioni a causa dell'ingombro nell'annulus e quindi della riduzione di portata elaborata, induce delle vibrazioni all'interno del sistema. Tali vibrazioni indotte dallo stallo costituiscono una limitazione di importanza fondamentale sulle performance delle turbomacchine rotanti, in particolare nei compressori e nei ventilatori, per i quali lo stallo rappresenta una problematica persistente nella loro progettazione e sviluppo. La comprensione dei fenomeni fisici chiave che si manifestano durante lo stallo e prima, durante lo stall inception, hanno un ruolo critico nel tentativo di contenere o evitare lo stallo già in fase di progettazione o attraverso metodi di controllo attivo e passivo. Tradizionalmente i metodi di contrasto dello stallo impiegati si basano sulle tecniche di controllo passivo, e questo è vero in particolare in ambiente industriale. Queste tecniche di controllo applicate vanno a contrastare lo stallo, influenzando il flusso intervenendo su diversi fattori che possono condurre ad un recupero. In tempi più recenti per merito dello sviluppo tecnologico, le tecniche di controllo passivo stanno cedendo il passo alle tecniche di controllo attivo. Queste ultime con l'ausilio di sensori ad alta risposta e alta sensitività in combinazione con le aumentate capacità computazionali a disposizione, hanno permesso di sviluppare sistemi di controllo, di early detection e di warning che hanno l'ambizione di prevenire fenomeni dannosi legati al insorgere dello stallo, se non addirittura prevenire lo stallo stesso, andando a riconoscere i segnali che ne annunciano l'incombere. I costi economici e computazionali elevati, legati alle alte prestazioni di strumentazione e analisi, rendono queste più moderne tecnologie meno appetibili a livello industriale. Da qui è nata l'idea di implementare una catena di misura per lo stallo che fosse basata su sensori DIY (Do It Yourself). I sensori non convenzionali sono stati usati per misurare le instabilità di pressione nel campo vicino, a bordo di un ventilatore assiale a bassa velocità, su cui veniva forzato lo stallo. Si è tentato quindi di sviluppare e testare un sistema di misura che fosse in grado di acquisire i segnali a bassa frequenza tipici dello stallo, con dei sensori che hanno tipicamente un range di frequenza di acquisizione limitato rispetto a soluzioni tecnologiche più avanzate ed evolute. L'obiettivo finale era la realizzazione di un sistema di stall warning affidabile e robusto. I dati acquisiti durante lo stallo sono quindi stati analizzati con i metodi più tradizionali della auto-correlazione spettrale che, nonostante i limiti dei sensori rispetto alle frequenze di acquisizione, hanno fornito una prima validazione della bontà del sistema di misura impiegato. In aggiunta si è voluto impiegare un sistema di analisi basato sul confronto dei pattern forniti da una tecnica ispirata alla ricostruzione dello spazio delle fasi, una metodologia molto impiegata nello studio dei sistemi dinamici caotici, categoria in cui lo stallo rotante rientra pienamente. Si è infine voluta indagare l'influenza di un fenomeno secondario, quale il fouling della palettatura, rispetto allo stallo, sempre ricorrendo alla catena di misura realizzata con i sensori DIY e con i sistemi di analisi sopracitati. Il fouling è dovuto alla presenza nel flusso elaborato di particelle di polvere, polline o sporcizia. Queste particelle aderiscono alla palettatura fino a formare uno strato di deposito che è di ostacolo a regolare passaggio del flusso e altera la geometria del rotore, creando problemi di sbilanciamenti, instabilità e spostando il limite di stallo. Con il sistema di stall detection realizzato, è stato possibile identificare lo stallo e riconoscere il caso in cui fosse presente lo sporcamento sulla palettatura, in quanto si è dimostrata una variazione della dinamica dello stallo e sono stati identificati dei precursori dello stallo. Le prove sperimentali hanno confermato che il sistema realizzato rappresenta una possibile soluzione per l'individuazione dei pattern tipici dello stallo rotante e del fouling nelle turbomacchine.
Development of non-conventional experimental techniques for industrial fans Application to stall and fouling phenomena / Tortora, Cecilia. - ELETTRONICO. - (2016).
Development of non-conventional experimental techniques for industrial fans Application to stall and fouling phenomena
TORTORA, CECILIA
01/01/2016
Abstract
This project starts from the need, in the industrial field, to avoid or at least limit, the stall phenomena and its harmful effects on the lifespan of rotating turbomachines. The significance of the stall phenomena depends by the issues rising as a result of the machine continuous working in the unsteady areas of the characteristic curve plane, beyond the stall limit, for prolonged periods or in a persistently oscillating way. During stall the stall cell formed on the blades, moves from a blade to the next in the opposite direction in respect to the blades rotating direction. The moving stall cell, besides leading to a reduction of the performance due to the annulus obstruction (and flow rate reduction) induces vibrations that are transmitted to the whole system. Such vibrations induced by the rotting stall represent a limit of fundamental importance on the performance of the rotating turbomachines, especially on compressors and fans, for whom the rotating stall is a persistent issue in design and development. The comprehension of the key physical phenomena that manifest during stall and before (stall inception) have a critical role in the attempt to contain and avoid stall, already in the design phase or through active and passive control methods. Traditionally the methods used to contrast stall are based on the passive control techniques. This is true especially in the industrial field. The passive control techniques operate a contrast to rotating stall by influencing the flow, since they intervene on the various factors that may lead to a stall recovery. In more recent years, thanks to the technological progress, the passive control techniques are giving way to the active control techniques. These last with the aid of high-response, high-sensitivity sensors, combined with the available increased computational power, have allowed to develop systems of control, early detection and warning, that have the ambition of preventing the damaging phenomena bound to the stall occurrence, if not even prevent the stall itself, recognizing the signals that announce its impending. The high economical and computational cost related to the high performance of instrumentation and analysis methods makes these technologies less attractive, industrially speaking. From these considerations is born the idea to develop a measurement chain for stall detection based on DIY (Do It Yourself) sensors. The unconventional sensors have been used to measure the pressure instabilities as pseudo-sound signals in the near field, on board of an industrial axial fan on which stall was forced. The aim was to develop and set-up a measurement system able to acquire the low frequency signals typical of stall with the aid of sensors that have typically a narrow range of acquiring frequency, if compared with more advanced and evolved technical solutions. The final aim was the realization of robust and reliable stall warning system. The data acquired during stall have been analysed with the traditional methods of the spectral auto-correlation, that although the sensors limits about the acquisition frequency, provided a first validation of the goodness of the used measurement system. In addition it has been used an analysis system based on the comparison of the patterns provided by a phase space reconstruction inspired technique. The reconstructed phase space is a widely used technique in the study of chaotic dynamic systems, and the rotating stall totally fits in this class. Finally it has been investigated the possibility to detect a secondary phenomena, such as fouling, and its influence on the rotating stall, again using the same DIY measurement chain and the above mentioned analysis methods. The fouling on the rotating blades is caused by the presence, in the flow passing through the machine, of particles of dust, pollen or dirt that adhere to the blades forming a deposit on them. This deposit is an obstacle to the passing flow and alters the rotor geometry leading to unbalance problems, instabilities and moving the stall limit. By means of the realized stall detection system it has been possible to identify the rotating stall and recognize the presence of fouling, since there was a modification in the stall dynamic and stall precursors were identified. The experimental test demonstrated that the realized system represent a candidate solution for the detection of patterns typical of rotating stall and fouling in turbomachines.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.