In magnetic-confinement nuclear fusion devices, plasma-facing components (PFCs) have to face intense heat loads (≥10 MW/m2 in the divertor region). Decades of plasma-wall interaction research point to a PFC design with plasma-facing tungsten (W) tiles joined to a copper (Cu)-based heat sink. In so-called "flat-tile" configurations, thermal gradients and the coefficient of thermal expansion (CTE) mismatch between W and Cu induce critical stress singularities at the free-edge interface, often leading to component failure through delamination. In this context, tailored interfaces represent a possible solution to improve PFCs durability, now obtainable thanks to the recent developments in additive manufacturing (AM) technologies that allow to achieve more complex and optimized joint geometries. To investigate the thermomechanical behavior of fusion-relevant W-Cu joints, and to provide an experimental and manufacturing pathway for the future standardized testing and comparison of different interface architectures, a custom experimental mock-up is designed. Its geometry is obtained and optimized through a simulation-driven approach, where finite element analysis (FEA) is used to mimic the loading conditions generated by the testbench. Assessment of the FEA results combined with design iterations, allowed to obtain a final geometry to be manufactured and tested. The experimental mock-up is composed of six specimens, brazed to a CuCrZr heat-sink. Each specimen is composed of an oxygen-free electronic copper (OFE-Cu) cast interlayer, joined to a DEMO-size W flat-tile. To obtain "lower bound" reference results on the performance of the selected geometry, a defined W-Cu joint is achieved by grinding the W surface smooth (Sa < 0.4 um) before casting the interlayer, thus reducing the mechanical interlocking effect of the roughness. Optimal manufacturing parameters for the brazing process are obtained through a preliminary assessment, where the process is repeated varying the alloy, amount of material, surface roughness, and temperature profile of the furnace. Best results are obtained with GEMCO alloy, smooth surfaces, and a temperature profile with 30 minutes of outgassing at 500 °C, and 20 minutes of dwell time at 985 °C. Simulations show that CTE mismatch and thermal gradients contribute both significantly to the stress field in W-Cu joints; the need to experimentally capture both phenomena at the same time motivates the selection of GLADIS as the testing facility. A high heat flux (HHF) thermal cycling test is performed on the mock-up replicating fusion-relevant conditions. The test is carried out without reaching failure of the mock-up, which survives 100 pulses with peak power density of 5 MW/m2 and 300 pulses at 10 MW/m2. Thermocouple measurements show almost perfect agreement with FEA results and steady values across pulses, indicating that the heat-sink is not degrading during the test. Tile #3 shows a 60% temperature increase on its surface (800 °C) during the cyclic test, linear with the pulses. Thermal gradients across the specimen are extracted from thermal side-images through a Python code, highlighting a growing increase of the temperature at the W-Cu interface and their evolution with cycling. This, combined with a post-experiment optical microscopy analysis showing severe delamination and plastic deformation at the interlayer, provides evidence that joint degradation is occurring during the test, and how this plays a crucial role in the thermal performance of the mock-up. The Python routine is also used to perform differential image thermography through point-wise subtraction of thermal images. This allows to detect and track the evolution of hot spots across the geometry. Two structured interface geometries are designed based on qualitative stress considerations, and later additively manufactured on DEMO-size W tiles. These specimens are set to be brazed and tested on the custom heat-sink using the same procedure obtained in this thesis; they outline the future research directions emerging from this project, possibly allowing the future development of advanced PFCs joints for nuclear fusion applications.
Nei dispositivi per la fusione nucleare a confinamento magnetico, le componenti a contatto con il plasma (Plasma-Facing Components, PFCs) sono soggette a carichi termici sostenuti (≥10 MW/m2 nella regione del divertore). Decenni di studi sulle interazioni plasma-parete suggeriscono un PFC design basato su piastrelle in tungsteno (W) accoppiate a un dissipatore di calore a base di rame (Cu). Nelle cosiddette configurazioni "flat-tile", i gradienti termici e la differenza nei coefficienti di espansione termica (CTE) tra W e Cu generano singolarità di stress critiche al bordo libero dell'interfaccia, spesso portando a delaminazione e cedimento del componente. In questo contesto, interfacce appositamente progettate possono migliorare la durabilità delle PFCs; ciò è reso possibile dai recenti sviluppi nell'additive manufacturing (AM), che consentono la realizzazione di geometrie di giunzione più complesse e ottimizzate. Per studiare il comportamento termomeccanico dei giunti W-Cu rilevanti per la fusione, fornendo un percorso sperimentale e di produzione per effettuare test e confronti standardizzati tra diverse architetture di interfaccia, un dissipatore è stato appositamente progettato. La sua geometria è stata ottenuta e ottimizzata mediante un approccio guidato da simulazioni: analisi agli elementi finiti (FEA) sono state impiegate per simulare il carico dell'apparato sperimentale. Esaminando i risultati FEA e iterando il design, si è ottenuta una geometria finale da produrre e testare. Il mock-up sperimentale è costituito da sei campioni, brasati al dissipatore in CuCrZr. Ciascun campione è formato da uno strato intermedio in rame oxygen-free electronic (OFE-Cu) ottenuto per colata, accoppiato a una piastrella in W di dimensioni rilevanti per DEMO. Per ottenere risultati di riferimento sulle "prestazioni minime" della geometria selezionata, la superficie della piastrella in W è stata levigata (Sa < 0.4 um) prima della colata, riducendo l'effetto di interlocking meccanico dovuto alla rugosità. I parametri di fabbricazione ottimali per il processo di brasatura sono stati determinati mediante una valutazione preliminare, dove il processo è stato replicato variando lega, quantità di materiale, rugosità superficiale e profilo termico della fornace. I migliori risultati si sono ottenuti con la lega GEMCO, superfici lisce e un profilo con 30 minuti di degassamento a 500 °C e 20 minuti di mantenimento a 985 °C. Le simulazioni indicano che sia la differenza di CTE sia i gradienti termici contribuiscono in modo significativo al campo di stress; la necessità di catturare perimentalmente entrambi i fenomeni in modo simultaneo ha motivato la scelta della facility GLADIS per i test. Un test di thermal-cycling ad alto flusso termico (high heat flux, HHF) è stato eseguito sul mock-up. La componente non ha raggiunto il collasso, resistendo a 100 impulsi con densità di potenza massima 5 MW/m2 e a 300 impulsi a 10 MW/m2. Le misure con termocoppie mostrano un accordo quasi perfetto con i risultati FEA e valori stabili fra gli impulsi, suggerendo l'assenza di degrado del dissipatore durante la prova. Durante l'esperimento, la piastrella #3 mostra un incremento della temperatura superficiale del 60% (circa 800 °C), lineare con il numero di impulsi. I gradienti termici dei campioni sono stati estratti da immagini termiche tramite codice Python, evidenziando un netto aumento all'interfaccia W-Cu. Tale osservazione, combinata con l'analisi postuma al microscopio ottico, che mostra accentuata delaminazione e deformazione plastica nello strato intermedio, fornisce evidenza del degrado del giunto durante il test e del suo cruciale impatto sulle prestazioni termiche del mock-up. Il codice Python è stato anche utilizzato per eseguire termografia differenziale tramite sottrazione punto-punto di immagini termiche, permettendo di rilevare e tracciare l'evoluzione di hot spots sulla geometria. Due geometrie di interfacce strutturate sono state progettate basandosi su considerazioni qualitative sullo stress, e successivamente stampate mediante additive manufacturing su piastrelle DEMO. Questi campioni saranno in futuro brasati e testati sull'apposito dissipatore, seguendo la procedura sviluppata in questa tesi; essi indicano le direzioni future di ricerca derivanti da questo progetto, che permetterà il test, confronto ed ulteriore sviluppo di PFCs avanzate per applicazioni sulla fusione nucleare.
Testing of W-Cu joints for plasma-facing components by means of high heat flux thermal cycling
Cocchi, Simone
2025/2026
Abstract
In magnetic-confinement nuclear fusion devices, plasma-facing components (PFCs) have to face intense heat loads (≥10 MW/m2 in the divertor region). Decades of plasma-wall interaction research point to a PFC design with plasma-facing tungsten (W) tiles joined to a copper (Cu)-based heat sink. In so-called "flat-tile" configurations, thermal gradients and the coefficient of thermal expansion (CTE) mismatch between W and Cu induce critical stress singularities at the free-edge interface, often leading to component failure through delamination. In this context, tailored interfaces represent a possible solution to improve PFCs durability, now obtainable thanks to the recent developments in additive manufacturing (AM) technologies that allow to achieve more complex and optimized joint geometries. To investigate the thermomechanical behavior of fusion-relevant W-Cu joints, and to provide an experimental and manufacturing pathway for the future standardized testing and comparison of different interface architectures, a custom experimental mock-up is designed. Its geometry is obtained and optimized through a simulation-driven approach, where finite element analysis (FEA) is used to mimic the loading conditions generated by the testbench. Assessment of the FEA results combined with design iterations, allowed to obtain a final geometry to be manufactured and tested. The experimental mock-up is composed of six specimens, brazed to a CuCrZr heat-sink. Each specimen is composed of an oxygen-free electronic copper (OFE-Cu) cast interlayer, joined to a DEMO-size W flat-tile. To obtain "lower bound" reference results on the performance of the selected geometry, a defined W-Cu joint is achieved by grinding the W surface smooth (Sa < 0.4 um) before casting the interlayer, thus reducing the mechanical interlocking effect of the roughness. Optimal manufacturing parameters for the brazing process are obtained through a preliminary assessment, where the process is repeated varying the alloy, amount of material, surface roughness, and temperature profile of the furnace. Best results are obtained with GEMCO alloy, smooth surfaces, and a temperature profile with 30 minutes of outgassing at 500 °C, and 20 minutes of dwell time at 985 °C. Simulations show that CTE mismatch and thermal gradients contribute both significantly to the stress field in W-Cu joints; the need to experimentally capture both phenomena at the same time motivates the selection of GLADIS as the testing facility. A high heat flux (HHF) thermal cycling test is performed on the mock-up replicating fusion-relevant conditions. The test is carried out without reaching failure of the mock-up, which survives 100 pulses with peak power density of 5 MW/m2 and 300 pulses at 10 MW/m2. Thermocouple measurements show almost perfect agreement with FEA results and steady values across pulses, indicating that the heat-sink is not degrading during the test. Tile #3 shows a 60% temperature increase on its surface (800 °C) during the cyclic test, linear with the pulses. Thermal gradients across the specimen are extracted from thermal side-images through a Python code, highlighting a growing increase of the temperature at the W-Cu interface and their evolution with cycling. This, combined with a post-experiment optical microscopy analysis showing severe delamination and plastic deformation at the interlayer, provides evidence that joint degradation is occurring during the test, and how this plays a crucial role in the thermal performance of the mock-up. The Python routine is also used to perform differential image thermography through point-wise subtraction of thermal images. This allows to detect and track the evolution of hot spots across the geometry. Two structured interface geometries are designed based on qualitative stress considerations, and later additively manufactured on DEMO-size W tiles. These specimens are set to be brazed and tested on the custom heat-sink using the same procedure obtained in this thesis; they outline the future research directions emerging from this project, possibly allowing the future development of advanced PFCs joints for nuclear fusion applications.| File | Dimensione | Formato | |
|---|---|---|---|
|
Thesis.pdf
non accessibile
Descrizione: Thesis classical format
Dimensione
186.64 MB
Formato
Adobe PDF
|
186.64 MB | Adobe PDF | Visualizza/Apri |
|
ExecutiveSummary.pdf
non accessibile
Descrizione: Executive summary
Dimensione
47.63 MB
Formato
Adobe PDF
|
47.63 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/10589/261076