Pratique du Shadowing: Envicool Full Chain Liquid Cooling 2GW+ Deployment Case Study presented by Envicool - Apprendre l'anglais à l'oral avec la vidéo
Choisissez un mode d'entraînement
Chaque mode ouvre une salle d'entraînement en plein écran.
206 phrases
Si les phrases sont trop courtes ou trop longues, cliquez sur Edit pour les ajuster.
1
My name is Lin Niu.
2
I'm the Product Manager for Liquid Cooling Solutions from NVQO.
3
Yeah, thank you.
4
Today I want to talk about our experiences in liquid cooling solutions for the past 20 years.
5
And so nice to have you all here.
6
And we are all here because we all know that the way we cool down the infrastructure is changing.
7
The question is, how do we get there together?
8
So for the following 20 minutes, I will introduce who we are, what we provide, and the ideas behind our work.
9
Let's first have a very brief introduction of our company,
10
in case some of you may not that familiar with us we are established in 2005 so making 20 years
11
in this industry for liquid cooling
12
and after 20 years developments we have been a team of over 2 000
13
or 5 000 employees globally and from the very beginning of our very first liquid cooling products delivered in 2015,
14
we have already delivered over two gigawatts liquid cooling solutions.
15
So today I want to discuss with more about liquid cooling solutions with you, with our customers, with our partners.
16
Our headquarters and main R &D centers is based in China and to maintain our competitive edge,
17
we also have our U .S.
18
R &D center in USA.
19
It's about 20 minutes drive from here.
20
And more critically, our manufacturing -based strategy is to distribute it globally.
21
We have our facilities in Thailand, we have facilities in USA,
22
and we also have our manufacturing bases in multiple locations in China.
23
And this ensures the supply chain resiliency and it can help us to meet critical and aggressive market timeline.
24
And about the service network, we have built a complete service network to ensure that wherever our customers are,
25
they can easily get in touch with us to get all the supports from technical to service.
26
Since 2021, we have seen a surge in customers asking about liquid cooling solutions.
27
At that time, most of the rack density are from 20 to 40 kilowatts.
28
And the concern of our customers are how much energy they can save if they use this liquid cooling solution.
29
And what is the comparison between the air cooling and the liquid cooling?
30
About the K -PACs, about the O -PACs.
31
But things changed when last year NVIDIA released their NBL 72 rack.
32
120 kilowatts per rack make the air cooling no longer available.
33
The liquid cooling solution is not just an option solution is the only solution to that high density rack.
34
And based on the explosive growth of the AI, it can be tell, it can clearly be tell that in the future,
35
one megawatt per rack will be the future.
36
And it's in near future, maybe in 2024, 27.
37
So what will be the solution for this high -density rack.
38
And from the heat transfer theory, we know that if we want to meet high cooling capacity,
39
high cooling requirements, we have three main theories.
40
The first is about the new coolant type.
41
And the second is to increase the heat transfer area.
42
And the third is about high flow rates.
43
But there are some questions.
44
First, for the new coolant type, we can talk about nanofruit, we can talk about refrigerants used in some two -phase cooling systems.
45
We need to do a long -term compatibility test to use it before we implement it in a real data center.
46
Because those liquid cooling servers, that service they they provide, they support is very critical.
47
The second is about the surface, is about the area of the heat transfer surface.
48
When we're talking about the new structural design of the heat sink, about the cold place, we need to think about the coast.
49
That's one part.
50
And the second is about the production feasibility.
51
Some of the structure is very hard to make it in real world.
52
So that's the second.
53
And for the high flow rate, a high flow rate will of course cause a very high pressure job.
54
It will cause more space used in these data centers.
55
So how to make this happen?
56
The last is about the AI training, that it's very different to traditional general servers' works.
57
Like in different stages of AI training, the GPU load is different.
58
So how do our cooling systems can respond to this constantly or frequently vary those.
59
That's our question.
60
So the deeper we dive into this system, the more questions we find.
61
All these considerations are the foundation of the work and the projects that we are working on.
62
In China, we have built a joint lab with some of our customers to do some fundamental research.
63
And for the component level,
64
we have to do our new code plate design performance and the optimization of the structural study.
65
And we also do the leak test of the QD when they are operated under some offsets.
66
And we also do the research of the manifold uniform distribution and the coolant, the long -term lifetime property.
67
And for the system level, we also built our long -term operations test setup to do the test like
68
if the flow rate is over 1 .5 meter per second, what will happen?
69
What influence it will have on the whole system and the performance of the code plates and the strength of the filter?
70
What will be the influence? and for the coolant parts,
71
what is the lifetime model of the coolant?
72
Because we need to replace the coolant frequently now with like pitch 25 or other kinds of the coolant.
73
So how can we make it last longer?
74
That's also one of our topic.
75
And also we will do some CDU control logic test.
76
As we just discussed, the loads of AI training will vary frequently.
77
So how can our cooling system respond to it and make a stable flow rate and a stable temperature supply?
78
Those are all the topics and projects we are working on.
79
And the goal of our company is very simple.
80
We want to support reliable, we want to support sustainable solutions to our customers.
81
We want to discuss more with our customers, with our partners, to discuss about their requirements,
82
to know more about the problem they are facing, and to provide more reliable products and solutions to them.
83
And in U .S., we are also building a test platform in our office.
84
This platform is composed of four parts.
85
The first is the liquid -to -liquid -CDU loop.
86
The second is the sidecar solution, and the in -rack solution, and also the fan wall air cooling solutions.
87
With this Mooc data center,
88
it can give a very clear understanding to our customers who visit it to know what is the liquid cooling solution,
89
what is the direct -to -chip liquid cooling solution, and what is the valve configuration,
90
and maybe all the operation procedures that will happen in a real data center.
91
And that is one of the main features of this platform.
92
And the second is about the testing and the research.
93
We will use this platform to do some fundamental research.
94
Like for the CDU, we will do the thermal test, we will do the control logic test,
95
the function test, and also maybe the whole cooling system validation test.
96
And we will also use this platform to do some long -term operation.
97
So let's look what we have for this industry.
98
We want to provide reliable solutions for our customers.
99
So we raised the concept of full -chain nuclear cooling.
100
We defined the full chain with three strategic pillars.
101
full ecosystem integration, full application scenario, and also the full lifecycle management.
102
So first is about full ecosystem integration.
103
We will collaborate with customers across the entire industry from the
104
chief provider to several manufacturers all the way to the end users.
105
Like we will discuss about their requirements.
106
We will develop these products with them from the rack level, like code place, like QD and manifolds,
107
and all the way to the facility level, like the CDU, different types of the CDU, the code source, the coolant,
108
the secondary pipes, and even the control system.
109
And with this approach, it can also help us to have a better understanding of the entire system system.
110
And that is very critical for the reliability,
111
because only in that way we can make all the components work together seamlessly.
112
And the second level is full application scenarios.
113
We have expanded our products line away from the liquid cooling loop itself.
114
We developed the products based on all related application scenarios.
115
Like for the production line, we have our whole series of the products for all the critical tasks
116
that may happen in a liquid cooling server production line.
117
Like in the production line of liquid cooling server, they need to do the failing, draining, leak tests, and the flushing,
118
and maybe in the last part, the nitrogen injection and the drying.
119
We have a whole series of the products for that.
120
And the second is the lab test application scenario.
121
Some of our server manufacturer customers need to do thermal tests, need to do the flow tests to their liquid cooling server demo.
122
So we will, based on their requirements, to develop those specific equipments for the test.
123
And the last is for operation in a real data center.
124
For the data center operation and maintenance, for the data center commissioning, we have our load banks and the emulators for the cooling system validation.
125
And the last level is the full lifecycle management.
126
Reliability is not only about the components, it's about the whole lifecycle management.
127
So we will design all the products that we share in last pages all the designs are done by ourselves,
128
and all the production and manufacturing are in our own factories.
129
And for the installation and maintenance,
130
we also have our own teams to do all these works to minimize the complexity for our customers.
131
So that's the three -fold.
132
We call it a three -fold.
133
We call it a full -chain nuclear cooling solution.
134
In the following parts I will give some detailed introduction
135
Of our products
136
First is the CDU CDU is the core components
137
For this direct -to -chip liquid cooling
138
So we provide in -road CDU We provide in -react CDU Also both we have
139
Liquid -to -liquid type
140
And liquid -to -air type
141
For our liquid in -road CDU Based on our This year's project experiences
142
we have already delivered multiple models from 300 kilowatts all the way to 2 .5 megawatts.
143
And especially for our 2 .5 megawatts CDU, its cooling capacity is based on 4 -degree Celsius ATD,
144
and it has a wide flow rate range from 1 % to 100 % with very stable temperature and flow rate support.
145
And the second is about our sidecar, it's also the in -row liquid -to -air CDU.
146
Our current model can support 100 kilowatts under 15 degrees Celsius ATD, and we also have N plus 1 pumps,
147
N plus 1 the power module, and N plus 1 the vents to support concurrent maintenance.
148
And in this Q4, we will have our 240 kilowatt sidecar.
149
It's about two rack wide.
150
Please, you can check our social media.
151
We will release in Q4.
152
And then it's about our INRAC CDU.
153
We have our INRAC Liquid2Air and Liquid2LiquidCDU.
154
For our Liquid2AirCDU, this is mainly used in some lab test scenarios.
155
Now our standard model is 20 kilowatts, but of course we can customize based on customer's requirements.
156
And for our URAC liquid to liquid CDU,
157
the most powerful unit that we have now is 160 kV under 4 degree Celsius ATD.
158
And it also has N plus 1 POM redundancy, and it can be compatible for all our WRAC.
159
So for the manifold and QD, now the most commonly used QD in this industry now is UQD series series,
160
and also now we have MQD series.
161
Of course, we have our own design for the QDs, like low flow resistance QD and other types of QD.
162
If you have interest, you can come to our booth to have a visit.
163
And for our manifold, manifold is a high customized product.
164
Now, the largest manifold that we have provided is about 800 LPM per rack.
165
The secondary pipe is also a highlight product that we provide.
166
For the secondary pipes, because this liquid cooling system requires very high cleanliness, so all these pipes are manufactured in modular
167
and they are prefabricated in our own factories and shipped to the project site to do the installation.
168
And we have modular pipes.
169
And in order to decrease the requirements for the manpower and the lead time,
170
we also provide this framed piping.
171
It integrates the pipes into the supports, and all the integration will be happening in our own factories.
172
You only need to do some tiny connections on site.
173
It will decrease about 50 to 70 % of the time
174
that it requires to do the installation of the whole secondary loop.
175
This is our load bank.
176
And our load bank is also a modular requirement.
177
For our load bank, we provide this for you 20 kilowatts liquid cooling load bank.
178
It can be integrated in one rack,
179
multiple this modular liquid cooling load bank can be integrated in one rack to reach about 240 kilowatts per rack.
180
And we are also looking for a more higher density of
181
this liquid cooling load bank for the test of our future rack, like one megawatts or even higher.
182
So the last part is about our products for the production line.
183
We have discussed with our server manufacturing customers about the typical process during their production of the server.
184
Typically, including the leak test, flushing, draining, fading, burning test,
185
and also last part is the draining and nitrogen injection.
186
We want to introduce our concept of automation to our customers.
187
So there is a video introduce this concept.
188
We'll be right back.
189
Thank you.
190
Thank you.
191
Thank you.
192
Thank you.
193
Thank you.
194
Oh, So thank you.
195
So that's our production testing line for the liquid cooling servers.
196
And I will also introduce some of the details of the parameters of the products
197
that you have seen in this testing line.
198
So, this one, we call it automatic test tools.
199
So, in this unit, we have integrated all the tasks that it requires,
200
like all the filling and the drying nitrogen injection to this equipment.
201
But for customers who will require that all the function parts need to be separated,
202
we also have our separated products for those parts like this is the automatic failing tools.
203
And so we are all the explorers in this AI era and we are all the explorers in this liquid cooling era.
204
We want to have more collaborations with our customers and with our partners to make innovation.
205
So I warmly welcome you to our booth at B14 to have a more detailed view of our products.
206
Thank you.
Télécharger l'application
Tout pour parler couramment
Prononciation IAChaque phrase notée
Phonétique (IPA)Maîtrisez chaque son
VocabulaireEnrichissez votre lexique
Jeu de vocabulaireApprenez en jouant
FOR YOU
Vidéo recommandée
C13
130 phrases
The Truth About How Data Centers Use Water And Power To Keep Cool
Daily ConversationUniversity StudentsWorking Adults
C12
28 phrases
What is IMC (Intermetallic Compounds) part II? The different between IMC, alloy, eutectic
Daily ConversationUniversity StudentsWorking Adults
C11
20 phrases
Manufacturing Equipment at Z-AXIS
Daily ConversationUniversity StudentsWorking Adults
C12
318 phrases
實戰練英文｜黃仁勳：你們要加油，不然我會拿走所有機會，然後我會變很有錢｜德國沒有下一個機會了｜投資德國高峰會｜長文、註解、大字版、中英對照
Daily ConversationUniversity StudentsWorking Adults
C15
72 phrases
Businesses drive innovation for climate resilience | Vietnam Today
Daily ConversationUniversity StudentsWorking Adults
C14
127 phrases
How to design a Deployment Pipeline (GitOps)
Daily ConversationUniversity StudentsWorking Adults
C11
398 phrases
Sec Growth DataScience staff meeting Sep 14 2022
Daily ConversationUniversity StudentsWorking Adults
C11
63 phrases
Power House Episode 2: The Future of Global Commerce: Trusted Data & Digital Supply Chains
Daily ConversationUniversity StudentsWorking Adults
C1304 phrases
From Home Appliance to Grid Asset: The Evolution of Heat Pumps
Daily ConversationUniversity StudentsWorking Adults
C11
31 phrases
New Modula Copilot: Inside the HMI Redesign for Smarter Warehouse Automation
Daily ConversationUniversity StudentsWorking Adults
C11
282 phrases
什麼是屏蔽罩(shielding can)？工程師對屏蔽罩的愛恨情仇
Daily ConversationUniversity StudentsWorking Adults
C11
150 phrases
Parallel Pump Operation
Daily ConversationUniversity StudentsWorking Adults
Vocabulaire et conseils d’expression pour cette leçon
Cette leçon d’expression orale s’appuie sur la vidéo « Envicool Full Chain Liquid Cooling 2GW+ Deployment Case Study presented by Envicool ». Les mots qui reviennent le plus souvent : cooling, liquid, test, solution, customers. Cette vidéo contient 206 phrases et 2480 mots à répéter en shadowing. La partie parlée dure 23:22. Le locuteur parle lentement, environ 106 mots par minute, ce qui laisse le temps d’imiter chaque son. 83 % des mots font partie des 3 000 mots les plus courants en anglais ; le reste mérite d’être vu avant de commencer.
Vocabulaire clé de cette vidéo
15 mots de la vidéo à retenir, avec leur prononciation et leur sens :
| Mot | Prononciation | Sens |
|---|---|---|
| liquid nom | /ˈlɪk.wɪd/ | liquide |
| rack nom | /ɹæk/ | étagère |
| server nom | /ˈsɝvɚ/ | serveur |
| requirement nom | /ɹɪˈkwaɪɹmənt/ | exigence, besoin |
| pipe nom | /ˈpaɪp/ | cornemuse |
| scenario nom | /sɪˈnɛəɹioʊ/ | scénario |
| introduce verbe | /ˌɪntɹəˈdus/ | présenter |
| density nom | /ˈdɛn.sə.ti/ | masse volumique, densité |
| component nom | /kəmˈpoʊ.nənt/ | composant |
| integrate verbe | /ˈɪn.tɪ.ɡɹeɪt/ | intégrer |
| modular adjectif | /ˈmɑd͡ʒəlɚ/ | modulaire |
| stable nom | /ˈsteɪ.bəl/ | étable, écurie |
| loop nom | /luːp/ | boucle |
| reliable adjectif | /ɹɪˈlaɪəbl̩/ | fiable, sûr |
| integration nom | /ˌɪn.tɪˈɡɹeɪ.ʃən/ | intégration |
Les verbes à particule que vous entendrez
| Mot | Prononciation | Sens |
|---|---|---|
| cool down verbe | /kuːl ˈdaʊn/ | se refroidir, refroidir |
La grammaire de cette vidéo
Les structures que le locuteur utilise le plus, avec les mots exacts de la vidéo :
| Structure | Dans la vidéo |
|---|---|
| Present perfect have/has + participe passé — une action passée qui compte encore maintenant | have been · have already delivered · have built |
| Voix passive be + participe passé — l’accent est mis sur ce qui arrive, pas sur qui le fait | are established · is based · are operated |
Prononciation à surveiller
Le locuteur utilise 3 contractions et formes réduites, comme I'm, we'll, we're. Prononcez-les sous leur forme courte, telles que vous les entendez.
- Les sons « th »: booth /buθ/, thermal /ˈθɜːməl/
- Les sons « sh » et « zh »: integration /ˌɪn.tɪˈɡɹeɪ.ʃən/, installation /ˌɪn.stəˈleɪ.ʃən/, injection /ɪnˈd͡ʒɛk.ʃən/, structural /ˈstɹʌk(t͡)ʃəɹəl/, flush /ˈflʌʃ/
- Mots longs — placez bien l’accent: integration /ˌɪn.tɪˈɡɹeɪ.ʃən/, installation /ˌɪn.stəˈleɪ.ʃən/, automatic /ˌɔː.təˈmæt.ɪk/, fundamental /ˌfʌn.dəˈmɛn.təl/, manufacturer /ˌmænjʊˈfæktʃəɹə/
Les sons difficiles pour les francophones :
- /h/ — il se souffle, il n’est pas muet: headquarters /ˈhɛdˌk(w)ɔɹ.tɚz/, wherever /(h)wɛɹˈɛvɚ/, highlight /ˈhaɪˌlaɪt/
- /tʃ/ et /dʒ/ — à ne pas adoucir en « ch » et « j »: modular /ˈmɑd͡ʒəlɚ/, injection /ɪnˈd͡ʒɛk.ʃən/, nitrogen /ˈnaɪ.tɹə.d͡ʒən/, chip /t͡ʃɪp/, manufacturer /ˌmænjʊˈfæktʃəɹə/
- /r/ anglais — langue recourbée, sans frotter la gorge: rack /ɹæk/, requirement /ɹɪˈkwaɪɹmənt/, scenario /sɪˈnɛəɹioʊ/, introduce /ˌɪntɹəˈdus/, integrate /ˈɪn.tɪ.ɡɹeɪt/
Comment s’entraîner avec cette vidéo
- Écoutez la vidéo en entier une fois sans parler et notez les mots que vous ne connaissez pas.
- Répétez phrase par phrase à vitesse normale, en reprenant chacune jusqu’à ce que votre rythme corresponde à celui du locuteur.
- Enregistrez-vous et comparez avec l’original, en faisant attention à des mots comme liquid, rack, server.
Qu'est-ce que la technique du Shadowing ?
Le Shadowing est une technique d'apprentissage des langues fondée sur la science, développée à l'origine pour la formation des interprètes professionnels. Le principe est simple mais puissant : vous écoutez de l'anglais natif et le répétez immédiatement à voix haute — comme une ombre suivant le locuteur avec un décalage de 1 à 2 secondes. Les recherches montrent une amélioration significative de la précision de la prononciation, de l'intonation, du rythme, des liaisons, de la compréhension orale et de la fluidité.