Chapter 941 A battle of wits and courage with the air

Of course, the statement of backstabbing was just Chang Haonan’s inner complaint.

If I really say it, it would be a bit of an exaggeration.

Although Dassault has cooperated with Torch Group and is also one of the important users of TORCH Multiphysics software, it has even purchased a supercomputer specially launched by Sugon Group.

But when it comes to numerical computing, it is ultimately just a relationship between manufacturer and customer.

The two parties have not formed any form of commercial alliance, and the Falcon Z joint development agreement does not specify the form in which engineering documents must be submitted.

However, this kind of "partner abandons oneself and chooses similar products from competitors" is really too NTR.

Therefore, Chang Haonan still downloaded the project file.

In fact, TORCH Multiphysics is even compatible with opening and editing cmh files after a certain minor version.

At most, some COMSOL Multiphysics-specific features may not be available.

Soon, a wing with grid division appeared on Chang Haonan's computer screen.

Moreover, it already has details.

If you zoom in, you can even see the flaperon modeling and winglets.

Strictly speaking, it is a wingtip sail.

It should be inspired by some designs of Airbus A320.

Of course this is normal.

Aircraft design, especially civil aircraft design.

Useful elements will quickly spread from one person to ten, and from ten to a hundred.

"Let me see what you are calculating..."

Chang Haonan began to further check some of the data contained in the engineering files——

In 1997, when the first trial version of TORCH Multiphysics was released, the only commonly used and convenient storage media was 3.5-inch floppy disks.

In order to avoid the situation where the project file is so large that it needs to be stored on several disks, Chang Haonan separated the editing and running log files from the project files.

Later, this became the practice in almost all numerical computing software.

As a result, the boomerang now hit his head.

In the absence of information, it took Chang Haonan about half an hour to figure out what problems the Dassault engineers were studying.

Simply put, they should be preparing to design a wing that can have extremely low drag in the transonic range.

Since the small-thrust turbofan modified from the M88 core engine is powerful, as long as the resistance of the fuselage, especially the wing part, meets the standard, the aircraft can achieve economical and quiet navigation at the "quasi-sonic speed" of 0.9 Mach or even 0.95 Mach.

This means that compared with its current competitors on the market, whose speeds generally range from Mach 0.7 to 0.8, the Falcon Z can save about 15% of flight time.

It is obviously attractive to the customer groups targeted by business jets.

And it was also a very good publicity stunt.

However, it is impossible for the Falcon Z to have a delta wing or a swept wing.

To achieve excellent near-sonic performance on a generally straight wing, optimizing the airfoil alone is not a solution.

Attention must also be paid to the details of the wing surface.

This engineering document calculates the air flow near the wing surface.

It is similar to the work that Chang Haonan did when he optimized the wing shape of the J-8C.

But the level of sophistication is much higher.

After all, within the limits allowed by the structure, fighter jets do not need to consider the noise and vibration caused by airflow.

In fact, even if we only consider the structure, a straight wing with a large aspect ratio is much more complicated than a delta wing.

Therefore, judging from the content presented in this engineering document, Dassault engineers seem to be still in relatively early stages of research.

The reason for switching to COMSOL Multiphysics is probably because of its advertised high computational efficiency.

Otherwise, this wing shape will most likely not catch up with the Falcon Z schedule.

Of course, the structure of a passenger plane is relatively simple.

Many models (such as the Xian MA60 of that year) even offered wings of different designs as optional parts to customers.

But it would definitely be better if it could be launched as standard.

but……

Looking at the wing model on the screen, Chang Haonan gently placed one hand on his chin.

"This calculation result seems... to have some problems?"

Before Samsung Marine Engineering did this, Chang Haonan had been studying out of interest how to solve the problem of high computational weight when using the finite volume method to solve the viscoelastic constitutive equation.

In fact, he had found a constitutive equation that did not involve diffusion terms, but he had not yet had time to find a specific solution.

During that process, Chang Haonan used a large number of existing examples at hand to provide "learning materials" for the high-order data dimensionality reduction process.

Therefore, although he himself cannot give numerical calculation results with accuracy to a certain number of digits like a computer, he still has some intuition about the characteristics, or trends, of the results.

But the several sets of results calculated by Dassault did not match his intuition.

"Professor Liu."

Chang Haonan raised his head and looked at Liu Hongbo, who was sitting on the sofa and had already drunk who knows how many cups of tea.

"Ok?"

The latter was just feeling bored and wondering whether to take the initiative to say goodbye, when he heard his boss's voice:

"Mr. Chang, please tell me."

"You have COMSOL Multiphysics version 2.5, right?"

Chang Haonan asked.

"some."

Liu Hongbo answered somewhat bewilderedly.

"That's good."

Chang Haonan said as he turned his gaze back to the screen, and then forwarded the email from Dassault to Liu Hongbo's mailbox:

"I sent you an email with a COMSOL Multiphysics project file. When you go back, open it with the latest version of their software. Don't change anything and run the calculation again from the beginning."

Although Liu Hongbo still didn't know what Chang Haonan wanted to do, he nodded:

"Okay, I'll send you the results as soon as possible."

However, Chang Haonan shook his head directly:

"No...it turns out I don't need it."

"When the time comes, extract the log of this calculation process and send it to my email."

……

Liu Hongbo moved very quickly.

Less than two days later, Chang Haonan received the log file he sent.

txt format.

Since COMSOL Multiphysics was released independently several years earlier than in history, many software-level details are not good.

For example, typically, logs are stored directly in txt format.

Of course, for a hardcore ruthless person like Chang Haonan who checks logs with his naked eyes, it is indeed a lot more convenient.

The results of the inspection...

It also confirmed his guess two days ago.

"Fuck..."

Chang Haonan leaned back in his chair, muttering to himself and cursing:

Nearly two months ago, he began to think about how COMSOL would solve the problem of high computational weight of the viscoelastic constitutive equation.

And the result is...

Not solved at all...

It can be seen from the log that the calculation process is to directly follow the general method and use the grid-centered FVM discretization method to numerically solve the NS equations.

However, situations with these characteristics are indeed extreme and are not actually very common.

In particular, COMSOL's past customers were generally schools and small and medium-sized enterprises with limited strength and financial resources.

Even less common.

Most of the new customers who started to try COMSOL Multiphysics after version 2.5 may not have fully adapted to it yet.

Like Dassault, they have already started using it for projects, but are still in the exploratory stage.

I don’t even know if there is a problem with the result.

So, the Swedes probably haven't gotten any feedback so far.

In other words, they probably didn’t know or thought about the question from the beginning to the end…

And Chang Haonan...

It has been fighting wits and courage with the air.

Thinking of this, he smiled bitterly and shook his head, then opened the Word document and prepared to write a memo to inform Dassault of the situation.

After all, the Falcon Z was a joint project.

If something goes wrong on a big job like a wing, it will inevitably affect the progress of the entire project.

By the way, it allows them to use the old computing tools with justification.

Thanks to the fact that it is not bound to the finite volume method, TORCH Multiphysics does not have to face the problem of high computational weight when calculating viscoelastic constitutive equations.

The cost is that it consumes more computing resources.

But anyway, it's better than figuring out the problem in the end.

As for the competitor COMSOL...

Chang Haonan already had a rather "evil" idea.