Chapter 900 Competitors' Prediction
Similar scenes naturally occur not only at Ohio State University.
As Professor Harold guessed, many people directly followed Chang Haonan's account on arXiv a year ago.
After the latter publishes an update, followers will immediately receive a reminder email in their mailbox—
Originally, functions like this may have to wait for a few more years, and will not be launched until the Internet era has fully arrived.
But because of Chang Haonan's emergence, technical staff at Cornell University's Science and Technology Campus were “forced” to develop this function in advance…
That's why thousands of people clicked to view it within ten minutes.
You must know that arXiv is not a self-media. Although there is no threshold for registration, its content inherently excludes most non-academic people.
Even scholars, in 2001, not everyone was proficient in using the Internet.
Besides, everyone is very busy, and it is impossible to spend all day on the Internet like an idle person.
So, being able to generate a 4-digit reading volume in a short period of time is already a very amazing communication efficiency.
However, not everyone has the same mentality as Harold.
Slightly later.
San Francisco, Stanford University.
Professor Anthony Jameson was frowning slightly and looking at the PDF document displayed on the computer screen.
Of course it is also the preprinted paper released by Chang Haonan.
But unlike Harold, there were several other similarly dressed people sitting next to him.
They are from the Swedish COMSOL company and are the development business leaders of COMSOL Multiphysics numerical calculation software.
In fact, COMSOL has recognized the potential of multi-physics simulation tools very early. In the mid-to-late 1980s, one of its founders, Svante Littmarck, began to write similar modules based on the MATLAB software that had just been commercially released at that time.
Unfortunately, it was still in the early stages of the development of computer applied science. Those who could apply computers on a large scale in their work were either top universities or super large companies such as Boeing, McDonnell Douglas, and IBM.
Everyone has idol baggage and wants to independently develop their own application software.
You mean outsourced software?
Can't afford to lose that person!
Hence, the road to commercialization of COMSOL products has never been smooth.
In this regard, on the contrary, China has provided Torch Group with a huge amount of business space because its universities and research institutions have limited capabilities and know that they do not have the ability to independently develop software.
In the true sense of the word “progress is rapid because the starting point is low”, it is yes.
It was not until TORCHMUltiphysics was released and began to seize the market with its absolute performance advantage that a MATLAB commercial toolkit called FEMLAB was launched in 1998 and was renamed COMSOLMultiphysics a year later in 1999.
It is precisely in order to quickly catch up with the pioneers that COMSOL hired Anthony Jameson as chief technology officer with a high salary—
Before the name Chang Haonan appeared, Anthony Jameson could be said to be one of the greatest names in the field of CFD, especially the finite volume method.
Since the 1970s, he has developed a series of CFD-related core algorithms for Boeing, McDonnell Douglas, NASA, France Aerospace, Dornier and other companies and institutions.
In particular, he wrote a code framework for predicting ideal transonic flow on swept wings - FLO, and based on this, he developed the most successful CFD software "A488" for Boeing before the 1990s.
The latter is the core calculation tool in the development process of Boeing 757, 767 and 777.
It can be said that during those years, Jameson's role to Boeing was almost similar to that of Chang Haonan to China today...
And such a big boss really had an almost immediate effect after joining COMSOL.
In 1999, the newly renamed COMSOL software was almost overwhelmed by competing products, and could only survive at an ultrlow price among some universities that could not afford TORC Multiphysics.
In less than two years, the newly released COMSOL Multiphysics 2.0 has already achieved more than 90% of the efficiency of its opponents in the electromagnetics-structural mechanics-fluid mechanics module.
ˆAlthough according to industry consensus, TORC Multiphysics definitely has a "professional enhanced version" with stronger performance, it has not been publicly released and does not constitute competition at the commercial level.
Hence, in recent months, COMSOL has captured a lot of markets due to its high cost performance and its status as a pure-blood European company.
Especially some small and medium-sized start-up companies, as well as research institutions that are willing to try various new things.
Although as a European company, it is indeed a bit shameful that it can only rely on cost-effectiveness to win.
But it's about making money.
Not shabby.
However, before they could breathe a sigh of relief, they discovered that the developer of TORCH Multiphysics opposite seemed to have come up with something new... This feeling of oppression was simply mind-numbing.
So, the group rushed to Stanford without even having time to have lunch and asked Professor Jameson about coping strategies.
“Professor, what do you think?”
Klaus Helz, the head of COMSOL based in the United States, looked at Jameson who had not spoken for a long time, and tentatively broke the silence.
“A very nice model.”
Jameson came up with a compliment.
This almost caused Hertz to have a cardiorespiratory arrest.
However, the former changed the subject immediately:
“But only on a theoretical level.”
It can only be said that Jameson knows how to suppress desires before promoting them.
Heltz, who had gone through a lot of ups and downs in less than a second, had no time to complain about this and quickly asked:
“So, professor, what you are saying is...there are still many difficulties in implementing this algorithm?”
“That's right.”
Jameson let go of the mouse and turned to look at Hertz, who was sitting not far behind him:
“Using LSM, oh, this level set method to simulate evaporation, atomization and combustion on the deformation interface is indeed closer to reality than the empirical vapor layer model and simple heat transfer model.”
"But in terms of specific algorithm implementation...traditional meshing, I mean, even moving meshes such as the Lagrangian mesh method, are difficult to meet the requirements of mesh generation accuracy and accuracy in this paper. Speed requirements.”
“You must know that most two-phase flow interfaces themselves change rapidly with time. The original LSM method is not strictly conserved and cannot guarantee accuracy in long-term simulations of more than 1 second.”
“Professor Chang's new algorithm achieves conservation when dealing with two-phase flow problems with divergent free velocity fields, but the generation efficiency in the Cartesian coordinate system is reduced..."
At this point, the former took off his glasses and rubbed the corners of his sore eyes:
"Of course, this paper is still a preprint, and it does not involve much about the specific implementation process of the algorithm. But unless the computing speed of the computer is improved by 3-4 orders of magnitude compared to now, there will definitely be no way to solve the problem of long wheelbase time parameters. CFD problem, so I speculate that the author may have found some specific conditions as a calculation example to get such beautiful results in the article..."
“You know, even for the N-S equation, people have found hundreds of analytical solutions under specific circumstances. According to Professor Chang's mathematical level, I think this is not difficult..."
It should be said that Jameson is worthy of being the best CFD expert in the last era.
He almost saw the weakest part of Chang Haonan's thesis at a glance in just twenty minutes.
That is, the grid generation efficiency in the Cartesian coordinate system.
It's just that, after all, he is already a CFD expert in the “last era”.
As for the power of new technologies in the new era, there has been a misjudgment after all...
Of course, at this point in time, Hertz must still be happy:
“So, our next work…”
“Proceed as usual.”
Professor Jameson replied decisively:
"My Synplane model has entered the final and most critical optimization stage. Judging from the current results, the traditional finite volume method is used to fix the coordinate system describing the fluid on the fluid particle and let it move with the fluid. , which can achieve the effect of accurately tracking the interface evolution process.”
"Currently, my research group is trying to topologically reconstruct the grid to solve the distortion problem caused by this process. It just needs to introduce special node search technology to determine the adjacency relationship of nodes. This part of the work is a bit complicated, but in the end It should be ready for commercial application by the end of this year.”
Facing the dominant position of Torch Group, COMSOL Multiphysics, in addition to its cost-effectiveness strategy, has another important advantage in its extremely fast update frequency.
More than a year after it was officially launched, the second major version has been launched.
Especially for industry-leading functions, as long as they can pass the performance test, they should be included first.
This is very attractive to most research institutions.
TORCHMultiphysics is the complete opposite. After the official version 1.0 was released, it only pushed a total of three small-scale updates. However, its performance is extremely stable. As long as you follow the operation manual, there will be almost no oscillations or divergences.
However, this also causes COMSOL to maintain such high-intensity updates, so the business is under great pressure.
They plan to push another major update at the end of 2001 to increase the software version number to 2.5.
Synplane, a model that can optimize the fluid mechanics and structural mechanics of the entire aircraft at the same time, is the "trump card" of version 2.5.
Hence, after hearing this assurance, Herz's tense expression finally relaxed.
(End of this chapter)
Chapter end
Report
|
Donate
Oh o, this user has not set a donation button.
|