酷兔英语

特汽车公司(Ford Motor Co.)工程师凯文·塔利奥(Kevin Tallio)举着一串用硬化过的沙子制成的扭曲圆环,宣称这个东西──发动机新部件气缸盖的铸模──不久前还是不可能做出来的。

 

作为福特的高级发动机工程师,塔利奥正在参与一场已经席卷整个汽车界的机动车设计革命。计算机辅助工程技术的进步和在计算能力方面的巨大投资为制造商提供了可进行创新设计的新工具,让他们能以更少的时间和成本测试自己的设计构思。

 

结果是:更多的设计概念被构思出来并得到了测试,最好的概念迅速得到采纳,帮助制造商提高汽车、卡车、巴士和摩托车的燃油效率和性能。

 

福特计算机辅助工程执行技术总监南德·科克哈尔(Nand Kochhar)说:"这种新流程使我们得以进行许多创新。"

 

汽车厂商现在正使用计算机运行数十种可能的设计方案,而这所花的时间以前只够生产出一个设计原型。

 

Altair Engineering Inc.

Altair的设计软件为车辆部件设计理想的外形,例如这款摩托车框架

塔利奥说,就在几年前,新的气缸盖从拿出设想到制作原型可能要耗时长达八个月之久,而且花费高达数百万美元。

 

开发新部件所需的时间和成本使得机动车制造商不愿考虑许多能小幅提高燃油经济性的大胆或创新的设计,因为失败的成本很高。

 

如今,新部件通过计算机模拟被构建,从而得出最高效的设计方案。接着工程师根据制造技术局限对设计进行修改,并在模型中对修改后的设计进行测试。而这种模型是使用积累了几十年的材料属性和发动机性能数据作为指导的。接着福特做出铸模以制造出实体部件,再将实体部件组装在发动机上进行进一步测试 。

 

Altair Engineering Inc.

出于制造方面的考量,设计师会对软件设计的框架进行修改

整个过程只需几天时间而不是几个月,花费也只有几千美元而非几百万美元──省下的时间和金钱值得冒一些设计上的风险。塔利奥说:"这就像是为我们的设计构思提供缓冲的降落伞。"

 

它是2010年福特探险者(Ford Explorer)的降落伞。在福特准备推出这款运动型多用途车的最新款时,刹车片出了一个问题。公司发言人杰克·迪里克(Jake Dylik)说,由于福特能够用虚拟技术设计和测试部件,并迅速生产出原型,新的刹车片在一周之内就制造了出来,公司最重要的一款汽车的发布也得到了拯救。

 

最新近利用到新设计流程的一个产品是福特将于今年引进中国的一款1.5升的四缸发动机。设计师得以迅速开发和测试将排气歧管(从发动机的气缸收集废气)置入发动机的原型,而不是将其装在发动机上。这使得这款发动机和它即将取代的大发动机一样省油和有力。小发动机可以使车主避免支付高额税务罚款,同时又不会牺牲性能。

 

美国联邦关于提高燃油效率的规定正促使机动车厂商们努力提高设计流程的效率。美国近期推出的更严厉的法规要求,新上市汽车和卡车的燃效到2025年需达到现有机动车的约两倍。这促使厂商们耗资数十亿美元以各种各样的方式进行投资,以便减轻车身重量,提高车辆空气动力性能,从而改善发动机效率。

 

从2000年开始,随着计算机模拟现实世界技术的提高,汽车厂商纷纷表示考虑放弃实物原型。但后来证明虚拟数码和实际之间的鸿沟要比人们起初认为的大得多。需要收集和分析更多的数据才能用数字技术对部件进行可靠的设计和测试。

 

这个过程在过去四五年里得到了加速。积累了数十年的关于各类材料成本、表现和安全性能的数据,以及设计数据已被进行数字化处理,从而使设计软件得以对这些数据进行权衡。

 

如今,一辆车被制造出来,它可以冲过雪堤,可以在极冷或极热环境中发动,可以重复碰撞──所有这一切都在庞大的计算机网络中进行。福特数字创新发展及部署全球总监鲍勃·特雷卡佩利(Bob Trecapelli)说,"我们可以在计算机模型中安装发动机管理软件并启动发动机",从而查看发动机在采用新设计部件后在各种不同状况下的性能。

 

为了充分利用这些技术进步,汽车厂商们在计算能力方面进行了大规模投资。

 

福特就是一个例子。科克哈尔说,在至少八年的时间里,福特每年的计算能力都提升了约50%。

 

在密歇根迪尔本(Dearborn)的福特总部,好几栋多层建筑里都堆满了运行公司全球工程程序的计算机。而且还不止如此。

 

福特与橡树岭国家实验室(Oak Ridge National Laboratory)──世界最强的计算机阵列──进行合作,实验室会在福特需要大量计算时协助其增强计算能力。

 

与此同时,高级软件为汽车厂商提供了大量新的设计选择。软件会根据大量数据提出最高效的设计,通常能够设计出工程师们无法想象的形状。

 

Altair Engineering Inc.的设计软件供世界各地的大型汽车制造商使用。这种软件通常会设计出复杂和非对称的纺锤形造型,然后还必须进行修改,从而变得更易于制造并与车的其他设计兼容。

 

这时福特大规模的计算能力就派上了用场。福特全球采购负责人唐浩泰(Hau Thai-Tang)说,公司将把Cray Inc.的超级计算机加入自己的阵列,从而迅速修改Altair软件拿出的设计。他说:"我们可以输入一系列限制条件,放进Cray计算机里,经过一晚上的时间,它就会输出最高效的解决方案。"

Ford Motor Co. engineer Kevin Tallio holds up a twisty series of loops made of hardened sand and declares that the object -- a mold for a new engine part, a cylinder head -- was an impossibility not long ago.

Mr. Tallio, a senior engine developer at Ford, is taking part in a revolution in vehicle design that has swept the auto industry. Advances in computer-aided engineering and big investments in computing power have given manufacturers new tools to create designs and the ability to test their ideas in a fraction of the time and at far less cost than they could before.

The result: Many more design ideas are being conceived and tested than ever before, and the best are being adopted quickly, helping manufacturers improve the fuel efficiency and the performance of cars, trucks, buses and motorcycles.

'This new process is allowing us to do a lot of innovation,' says Nand Kochhar, executivetechnicaldirector of computer-aided engineering at Ford.

Car makers are using computers to run through dozens of design possibilities in the time it once took to produce a single prototype.

Only a few years ago, it might have taken as long as eight months to get from the idea for a new cylinder head to the building of a prototype, and it would have cost millions of dollars, Mr. Tallio says.

The time and expense involved in developing new parts discouraged vehicle makers from looking at many daring or innovative designs that could yield small gains in fuel economy, because a failure could be so costly.

Today, the part was created in a computer simulation that came up with the most efficient design possible. Engineers then altered that design to account for manufacturing constraints and tested the revised design virtually in models that used decades of data on material properties and engine performance as a guide. Ford then created the mold to make a real part that could be bolted onto an engine for further testing.

The entire process took days instead of months and cost thousands of dollars instead of millions -- savings that make design risks worth taking. 'It's like a parachute for our ideas,' says Mr. Tallio.

It was a parachute for the Ford Explorer in 2010. When Ford was preparing to launch the latest version of the sport-utility vehicle, there was a problem with the brake rotor. Because of its ability to virtually design and test a part and quickly produce a prototype, a new brake rotor was created within a week, and the launch of one of the company's most important vehicles was saved, company spokesman Jake Dylik says.

A more recent product of the new design process is a 1.5-liter, four-cylinder engine that Ford will introduce in China this year. Designers were able to quickly develop and test a prototype that has the exhaustmanifold -- which collects exhaust from the engine's cylinders -- built into the engine, instead of being bolted on. That helps make the engine as efficient and powerful as the larger engine it will replace. The smaller engine will allow drivers to avoid an expensive tax penalty on larger engines, but with no sacrifice in performance.

The U.S. federalmandate for greater fuel efficiency is driving the efforts by vehicle makers to make their design processes more efficient. Recently toughened regulations in the U.S. now require new cars and trucks to have roughly double the fuel efficiency of today's vehicles by 2025. That is causing manufacturers to invest billions of dollars in various means to cut the weight of their vehicles, make them more aerodynamic and improve the efficiency of their engines.

Since 2000, car makers have been talking about ditching physical prototypes as computer simulations of real-world conditions improved. But the gap between the digital and the actual proved wider than initially thought. Much more data needed to be collected and crunched before parts could be reliably designed and tested digitally.

That process has accelerated in the past four or five years. Decades of data on the costs, performance and safety of various materials and designs have been digitized so they can be weighed by design programs.

Today a vehicle can be built, run through snow banks, started in frozen or hellishly hot conditions and crashed repeatedly -- all inside a vast network of computers. 'We can install the engine-management software into the computer model and run the engine' to see how it performs with newly designed parts under a wide variety of conditions, says Bob Trecapelli, Ford's global director of digital innovation development and deployment.

To take advantage of those advances, car makers have made massive investments in computing power.

Ford, for example, has increased its computing power by about 50% every year for at least eight years, says Mr. Kochhar.

In Dearborn, Mich., where Ford is based, there are several multistory buildings stacked floor to ceiling with computers to run the company's engineering programs around the world. And it doesn't stop there.

Ford has arrangements with the Oak Ridge National Laboratory -- the world's most powerful computing array -- to boost the auto maker's computing power when huge calculations are required.

Meanwhile, advanced software is giving auto makers a whole new universe of designs to work with. The software offers up the most efficient design possible, based on massive amounts of data, often creating shapes human engineers would be unlikely to imagine.

Altair Engineering Inc. sells design software to every major auto maker in the world. The software often comes up with elaborate, spindly, asymmetrical designs that then must be modified into something that is more easily manufactured and is compatible with the rest of a vehicle's design.

That's where Ford's expanding computing power comes in. Hau Thai-Tang, the global purchasing chief for Ford, says the company is adding Cray Inc. supercomputers to its array for the purpose of modifying the designs the Altair software suggests -- in a hurry. 'We can put in a set of constraints, throw it in the Cray overnight, and it will spit out the most efficient solution,' he says.



文章标签: