PBS高端访谈:3D打印如何推动制造业设计革命性升级

（单词翻译）

HARI SREENIVASAN: Whether it's with plastic, metal, or even living tissue, 3-D printing has been around since the 1980s. It's been used mostly for prototyping. And, so far, it's still cheaper to make most large-volume consumer goods like bottle caps using traditional methods. But, as Miles O'Brien reports, recent advances could launch 3-D printing into a new era. It's the subject of tonight's Leading Edge story, which airs every Wednesday.

MILES O'BRIEN: Just another day in an office park near LAX. No clue to the travelers above that a whole new approach to manufacturing is under way beneath their feet. It's happening at a young startup called Relativity, a team of for-real rocket scientists pushing space technology by pushing 3-D printing technology to its limits. Here, they are printing rockets, nose cone¹ to nozzle.

TIM ELLIS, Co-Founder, Relativity: Rockets are the lightest-weight, most expensive, largest, difficult-to-make thing, that really 3-D printing is the optimal² solution for.

MILES O'BRIEN: Founders³ Tim Ellis and Jordan Noone both realized this while working at large established aerospace⁴ manufacturers, where 3-D printing has been used for decades to make prototypes or a few parts here and there. They figure technology now makes it possible to think bigger. But to do this, they first had to build something bigger, the largest metal 3-D printer in the world.

TIM ELLIS: We made our own printing head, where we have aluminum⁵ wire fed in by this nozzle here. And then we're using 11 kilowatt⁶ fiber⁷ lasers to actually melt the aluminum. As you feed in material on the right, then the laser melts it. So, it's a very, very powerful laser. It can actually blind you from over 50 kilometers away.

MILES O'BRIEN: Good thing they aren't evil geniuses. Their mega-printer is called Stargate, a three-armed, 15-foot-tall robot. It hasn't made a whole rocket yet, but it has printed out a fuel tank and an engine. Relativity's full-throttled thrust into 3-D printing is just one milestone⁸ on the long road from prototypes and small parts to mass manufacturing. Mechanical engineer John Hart is director of the Laboratory for Manufacturing and Productivity at MIT.

JOHN HART, Director, Laboratory for Manufacturing and Productivity, MIT: I'm certain we're in early stages. I think that the things that we do with additive⁹ manufacturing, in the end, say, 10, 20, 30, 50 years from now, are in some part beyond our imagination.

MILES O'BRIEN: Hart is not talking about consumer-grade 3-D printers, a passing fad¹⁰ that peaked in 2014. He and his team at MIT are developing new materials and machines to help make 3-D printing more practical for manufacturers. They're grappling with familiar obstacles.

JOHN HART: 3-D printing is slow. It's expensive, right? There's few things that you can 3-D print and then use right away. You often have to do post-processing and finishing and painting, et cetera. But we're getting there.

MILES O'BRIEN: Hart and colleagues founded a company called Desktop¹¹ Metal to develop a solution. Traditionally, 3-D printing works by fusing metal powder together layer by layer with a laser. It's a single-point process limited by the speed of the laser. At Desktop Metal, they alternate layers of metal powder with a glue-like binder¹². The layers are sprayed with multiple print heads, inkjet-style. After the part takes shape, it is placed in a furnace, where the blast of heat fuses the metal while cooking away the binder. The company claims the process is about 100 times faster than the single-point laser technique. Based outside of Boston, Desktop Metal is growing fast. CEO Ric Fulop gave me a tour of his factory for factories. This is the main event right here, right?

RIC FULOP, CEO, Desktop Metal: This is our production system. This is the world's fastest metal printer. This machine can make a 150 metric tons of metal per year, 150 metric tons. There's nothing else like it.

MILES O'BRIEN: The production-scale metal 3-D printer is slated¹³ for its first delivery to customers early next year. The machine is well-suited to make higher-end, lower-demand parts like this.

RIC FULOP: This is a part made in our production system. This is for BMW.

MILES O'BRIEN: And that's -- it looks like some kind of cooling fan or something like that or...

RIC FULOP: That's a water impeller that goes inside a water pump.

MILES O'BRIEN: But 3-D printing is also spurring another revolution, in industrial design. The technique enables the creation of objects unimaginable using traditional tool and die techniques. The company is designing with software made smart by the artificial intelligence technique called machine learning. And here's the ironic¹⁴ twist. The machine is designing parts that appear to come from nature's playbook. Check out these two parts, on the left, a sleek¹⁵ human design, on the right, the rootlike handiwork of a smart computer. Andy Roberts is a software engineer. You have tested this, and what happens.

ANDY ROBERTS, Desktop Metal: What we find is that the parts have been self-organized so that they distribute the strain evenly across the parts. So, there are no sort of hot spots where you get a crack forming, for example.

MILES O'BRIEN: So, this is better than a human could do?

ANDY ROBERTS: Oh, yes. It is better than a human could do.

MILES O'BRIEN: It may be some time before organic-looking parts take root. But, in the short-term, some big players, like BMW and Caterpillar¹⁶, are anxious to try new ways of manufacturing their current designs.

JOHN HART: A lot of customers for industrial printing do get it. They have been working with the technologies for many years, studying them, prototyping with them. And there's this urge and thirst for mass production. I wouldn't have said this three to five years ago, but I'm convinced of it now, because you see more demonstrated applications.

MILES O'BRIEN: If 3-D printing delivers on these promises, it will do much more than upend the process of manufacturing. The ripple¹⁷ effects are far-reaching.

JOHN HART: From how the designer or the engineer goes about their work, to what the factory looks like, to how business agreements are structured, to where factories are placed, to what production workers do on a daily basis, it's all going to happen. And I like to think the goal is to get ahead about the understanding and the vision and help make it happen.

MILES O'BRIEN: At Relativity, they are still developing designs and printing processes, but they have reason to believe they have launched a good idea. They printed this giant, 14-foot-tall fuel tank in a matter of days. A traditional manufacturer would have taken a year. But, for Relativity, the real proof is in the testing, and they have successfully fired their printed rocket engine 85 times at NASA's fabled¹⁹ rocket testing center in Mississippi.

TIM ELLIS: So, that's like a fully¹⁸ printed design would normally be almost 3,000 parts, but we have gotten it down to three, and really shown that that's robust²⁰ and that it works.

MILES O'BRIEN: By the end of 2020, the team hopes to be delivering satellites and other payloads to low-Earth orbit with fully 3-D-printed rockets. They predict they can cut the cost of even the cheapest flights today by more than 80 percent. A game-changing number like that would destine manufacturing for a tectonic retooling²¹ layer by layer. For the "PBS NewsHour," I'm Miles O'Brien.

哈里·斯瑞尼瓦桑：自20世纪80年代起，无论是塑料还是金属，甚至是活组织，3D打印技术都可将其囊括。一直以来，它主要用于原型设计。而且，到目前为止，像如瓶盖这样，大批量消费品的生产，使用传统方法仍然更加便宜。但是，据迈尔斯·奥布赖恩报道，最近的技术进步，可将3D打印推向一个新时代。这是今晚《Leading Edge》的主题，《Leading Edge》每星期三播出。

迈尔斯·奥布赖恩：这是LAX附近，一个办公园区里平常的一天。对于来到这里的人来说，一种全新的制造方法正在他们脚下悄然进行。这一切就发生在一个叫做相对论的年轻创业公司里，一个真正的火箭科学家团队正将3D打印技术推向极限，挑战空间技术。在这里，他们正在打印火箭，从鼻锥体到喷嘴。

提姆·埃利斯，相对论联合创始人：火箭是最轻质，最昂贵，最庞大，也是最难做的东西，而3D打印技术真的是其最佳的解决方案。

迈尔斯·奥布赖恩：在为众多大型航空航天制造商工作时，创始人提姆·埃利斯和乔丹·努恩就已经意识到了这一点，在那些大型航空航天制造商那里，3D打印技术已经使用了几十年，人们利用3D打印技术制造模型样本，有时也会生产一些零件。他们认为，如今技术让人们得以脑洞大开。但要做到这一点，他们首先必须要建造一个更大的家伙，世界上最大的金属3D打印机。

提姆·埃利斯：我们制造了我们自己的打印头，我们在这里用这种管嘴喂进铝线。然后我们用11千瓦的光纤激光器来熔化铝。你在右边进料，然后激光把它熔化。所以，它是一个非常非常强大的激光器。它实际上可以从50公里以外的地方将你灼瞎。

迈尔斯·奥布赖恩：很好，他们不是邪恶的天才。他们的巨型打印机叫做星际之门，是个三条胳膊，15英尺高的机器人。完整的火箭尚未完成，但它已经打印出了油箱和发动机。相对论对3D打印技术的全面使用是漫漫长路上的一个里程碑，脱离了单纯对模型样本的制作和小批量零件的生产。机械工程师约翰·哈特是麻省理工学院制造和生产力实验室主任。

约翰·哈特，麻省理工学院制造和生产力实验室主任：我确信我们处于初期阶段。我认为，我们在附件制造业的所作所为，最终，比如说，10, 20, 30年，50年后，在某种程度上会超出我们的想象。

迈尔斯·奥布赖恩：哈特谈的并不是那种消费级的3D打印机，一种在2014年达到顶峰的时尚潮流。他和他在麻省理工学院的团队正在研发新的材料和机器，以帮助制造商进行更真实的3D打印。他们在应付熟悉的障碍。

约翰·哈特：3D打印很慢。很贵，对吧？很少有东西可以让你3D打印，然后立即使用。你经常要做后期处理，整理，喷绘，等等。但我们在努力实现这些。

迈尔斯·奥布赖恩：哈特和他的同事们创立了一个叫做Desktop Metal的公司，寻找解决方案。传统上，3D打印的工作原理是将金属粉末融化，并在激光的作用下，一层一层粘合起来。受到激光速度的限制，这是一个单点进程。在Desktop Metal公司，它们用胶状粘合剂替代了金属粉末层。这些层由多个打印头喷涂，喷墨式的。零件成形后，放在炉子里，热风把金属熔化，除去粘合剂。该公司声称，这一过程要比单点激光技术快100倍。坐落于波士顿城外，Desktop Metal正在快速成长。首席执行官瑞克·福勒普带我参观了他的工厂。这里就是核心，对吗？

瑞克·福勒普，Desktop Metal首席执行官：这是我们的生产系统。这是世界上最快的金属打印机。这台机器每年能制造150公吨的金属，150公吨。首屈一指。

迈尔斯·奥布赖恩：明年初，制造生产级的3D金属打印机将首次交付给客户。这台机器非常适合制造高端、低需求部件，比如像这个。

瑞克·福勒普：这是我们生产系统的一部分。这是给宝马做的。

迈尔斯·奥布赖恩：它看起来像某种冷却风扇或类似的东西...

瑞克·福勒普：它是个排水叶轮，放在水泵里的。

迈尔斯·奥布赖恩：但是3D打印在工业设计中也正在引发另一场革命。3D打印技术使得曾经利用传统工具及淘汰技术，不可想象的创造成为了可能。该公司正在使用的是由人工智能技术开发的智能软件，这种技术叫做机器学习。这是讽刺性的扭曲。机器设计的零件似乎浑然天成。看看这两个部件，左边这个是人工设计的光滑成品，而右边这个是由智能计算机完成的粗糙成品。安迪·罗伯茨是一名软件工程师。你已经测试过了，发生了什么？

安迪·罗伯茨，Desktop Metal：我们发现，这些部件是自组织的，所以他们把压力均匀地分布在各个零件上。所以，比如说，任何地方都不会出现导致裂纹的热点。

迈尔斯·奥布赖恩：那么，这比人类做得更好吗？

安迪·罗伯茨：哦，是的。比人类做得更好。

迈尔斯·奥布赖恩：可能需要一段时间才能看到机器人制造出看起来很天然的零件。但短期内，一些大公司，如宝马和卡特彼勒，急于尝试新的方法来制造生产目前他们的设计。

约翰·哈特：很多工业打印产品的顾客的确会这样做。他们多年来一直致力于这些技术的研究和使用。对大规模生产，有着强烈的欲望和渴望。三五年前，我不会这么说，但我现在相信了，因为你看到了更多被证明了的应用。

迈尔斯·奥布赖恩：如果3D打印实现了这些承诺，比起颠覆制造过程，它能改变更多的事情。连锁效应非常深远。

约翰·哈特：从设计师或工程师如何开展工作，到工厂的外观，商业协议的设定，工厂的选址，再到生产工人每天日常工作内容，这一切都会发生改变。我想，我们的目标是增强人们对它的理解和愿景，并帮助实现这一目标。

迈尔斯·奥布赖恩：在相对论，他们还在开发设计和印刷过程，但他们有理由相信，他们已经想到了一个好主意。他们在几天内，就完成了这个巨型14英尺高的燃料箱打印工作。而传统制造商要花一年时间才能完成。但是，对于相对论来说，实践出真知，他们85次，成功地在NASA密西西比州火箭试验中心，发射了由他们打印的火箭。

提姆·埃利斯：所以，这就像一个完全打印的设计，通常是近3000个零部件，但我们把它降到了3个，并且真的证明它很结实，能用。

迈尔斯·奥布赖恩：到2020年底，该小组希望，使用完全3D打印的火箭，向低地球轨道，发射卫星及其他有效载荷。他们预测，应用这项技术，即使今天最便宜的航班，也能削减80%以上的费用。像这样一个具有颠覆性意义的数字，注定会逐层引发生产制造业的构造调整。PBS NewsHour，我是迈尔斯·奥布莱恩。