PDLP Chapter 162 Equivalent Thermoelastic Approximation Method
Chapter 162 Equivalent Thermoelastic Approximation Method
2020年11月16日，晚上8点47分。
The live technical Q&A session has been going on for nearly fifty minutes.
At this point, the number of online viewers remained stable at around 131 million, with the barrage of comments so dense that the screen was almost unreadable. In the past twenty minutes, Lin Wei had thoroughly humiliated Fang Jianhua with Su Chen's handwritten technology notebook—page 47 of the third notebook contained a draft of the airflow channel dated March 2, 2020, a full three and a half months before the team was even formed.
Then, Zhou Zhiyuan's remote connection was an even bigger blow.
"It wasn't that he used me; it was that I chose him."
At this moment, Fang Jianhua sat in front of the live broadcast camera, his face ashen.
Zhou Zhiyuan's words were repeated no less than 100,000 times in the live chat, directly shattering the core narrative of "exploiting Zhou Zhiyuan's patriotism" in the open letter.
On the tablet in front of him, seven items in the "Guide to Questioning Techniques" sent by Ishikawa Akira had been crossed out, leaving only the last two. The first seven items—questioning the team's contributions, questioning the acquisition of Light Armor, and questioning whether the seed plan was an empty promise—were all shattered one by one by Lin Wei with solid evidence.
But Fang Jianhua did not give up.
Because he cannot give up. Ishikawa Akira's $800 million grant and the opportunity to study at MIT are one aspect, but more importantly—if he admits defeat now, what awaits him is not only the destruction of his academic reputation, but also the incriminating evidence that Bosch has against him.
He took a deep breath and looked at the eighth item in the conversation guide.
This is a line that Ishikawa Akira personally marked with three exclamation marks, with the words "Core Weapon - Questioning Technical Parameters" written in red next to it.
"Ms. Lin Wei," Fang Jianhua's voice regained its calmness, even carrying a hint of scholarly composure, "I admit that Su Chen did indeed have the design concept for the flow channel quite early on. I will no longer question the date on the notebook."
The live chat was instantly flooded with comments like "Did he back down?" and "Professor Fang gave in?"
But Fang Jianhua changed the subject: "However, having an idea and being able to implement it are two different things."
He stood up, walked to the whiteboard in the live broadcast room, and wrote down a set of numbers with a marker:
DRIE 250mm cavity thermal expansion pre-compensation angle: 12.05°
"This value," Fang Jianhua turned around, his gaze sharp as he looked at Lin Wei, "is a core design parameter published by Hongyuan Feiniao on page 137 of their technical white paper. A pre-compensation angle of 12.05 degrees—this is an extremely precise value."
He paused, glancing around at the camera: "You viewers may not be aware that in the field of MEMS manufacturing, the industry-standard method for calculating thermal expansion pre-compensation for large-size cavities is to use finite element simulation software—such as ANSYS, COMSOL, and ABAQUS. These are top-tier international simulation platforms, and the licensing fee for each software ranges from hundreds of thousands to millions of RMB."
Fang Jianhua's voice grew increasingly confident, because the next card he played was irrefutable evidence that Ishikawa Akira had spent two days investigating.
"but--"
His voice suddenly rose: "I reviewed all of Hongyuan Feiniao Technology Co., Ltd.'s software procurement records, technical service contracts, and procurement lists of affiliated companies since its establishment. The result is—"
Fang Jianhua stated the conclusion slowly and deliberately:
"Hongyuan Feiniao has never purchased any commercial finite element simulation software."
"ANSYS is not available."
"COMSOL is not available."
"There is no ABAQUS."
"There isn't even an educational version."
He turned to Lin Wei, his eyes flashing with a determined glint:
"Ms. Lin Wei, without simulation tools, how was the pre-compensation angle of 12.05 degrees, accurate to two decimal places, calculated?"
There was a three-second silence in the live stream.
The comments section also experienced a brief period of silence, which was unusual.
Then--
The live chat exploded with comments.
"Holy crap, you really can't do manual calculations from this angle, right?"
What is finite element simulation?
"It uses supercomputers to simulate physical processes; without this software, it's impossible to calculate precise angles."
"Professor Fang has hit the nail on the head this time."
"Wait, if there really isn't any simulation software... then is the 12.05 degree setting fabricated?"
In the WeChat group, messages from Lao Chen, the drone enthusiast, also stopped. He works with drones and is no stranger to finite element simulation—to calculate thermal expansion pre-compensation, it is indeed necessary to build a three-dimensional finite element model, set material properties, boundary conditions, and temperature field distribution, and then solve iteratively.
This is not a problem that can be solved by manual calculation.
"Old Chen, is Fang Jianhua right?" Xiao Wang from Shenzhen sent a message.
Old Chen, the aviator, typed a few words, then deleted them, and finally posted: "...That makes sense."
The calm analyst paused for a few seconds before saying, "If Hongyuan Feiniao truly doesn't have simulation software, then the value of 12.05 degrees does indeed need explanation."
At this moment, Zhao Mingyuan glanced at Lin Wei and noticed that Lin Wei's expression showed no panic.
In fact, from the start of the live broadcast until now, Lin Wei's expression has remained calm—not the kind of calm that is forced, but the kind of calm that says, "I knew you would play this card."
"Professor Fang." Lin Wei stood up and walked towards the whiteboard.
She was holding an iPad with a high-resolution scan of Su Chen's tech laptop on it.
"You're right, Hongyuan Feiniao has indeed not purchased any commercial finite element simulation software."
Fang Jianhua's lips curled up slightly—the other party had admitted it!
The comments section was flooded with phrases like "It's over" and "There's no way to explain this now."
"However," Lin Wei said calmly, "who says that calculating thermal expansion pre-compensation necessarily requires finite element simulation?"
Fang Jianhua's smile froze.
Lin Wei projected the iPad screen onto the large screen in the live broadcast room and turned to page 112 of the fifth notebook.
It was a page filled with densely written handwritten derivations.
The pages were slightly yellowed, with minor creases at the edges—traces of repeated reading. The date was written in the upper right corner: July 14, 2020.
The title of the derivation was a few words written by Su Chen in black pen:
Equivalent Thermoelastic Approximation Method—Analytical Solution for Thermal Expansion of Large-Size DRIE Cavities
Lin Wei turned the pages one by one.
Page 112 is the starting point of the derivation—starting from the coefficient of thermal expansion of silicon-based materials, a simplified axisymmetric thermoelastic model is established.
Page 113 contains the crucial mathematical transformation—Su Chen cleverly reduces the three-dimensional thermal stress problem to a quasi-two-dimensional problem, and then uses the Bessel function to expand and solve it.
Page 114 contains the error estimation—Su Chen marked a line in red on the right side of the page: "When the cavity depth-to-width ratio is <0.3, the deviation between the analytical solution and the finite element method is <0.5%, and it can be directly substituted."
Page 115 contains the verification results—the calculation results for a 200mm cavity.
Su Chen wrote down a set of numbers on this page:
"Estimated 200mm: 8.73°"
Actual measurement (October 27): 8.71°C
Error: 0.02°
A circle was drawn in red next to it, with the words: "Method feasible. Extended to 250mm."
Page 116 – Derivation results of the 250mm cavity.
"Estimated 250mm: 12.05°"
Below it was written a line of small print: "To be verified."
Lin Wei finished flipping through the five pages and then turned to Fang Jianhua.
"Professor Fang," her voice wasn't loud, but everyone in the live stream heard her clearly, "Su Chen didn't use finite element simulation because he didn't need it."
"He used an analytical method he derived himself—the equivalent thermoelastic approximation method. This method skips the three-dimensional iterative solution of finite element simulation and directly calculates the pre-compensation angle through the analytical solution."
"The estimated value for a 200mm cavity is 8.73 degrees, the actual measured value is 8.71 degrees, and the error is 0.02 degrees."
"This is where the 12.05 degree comes from."
The live stream fell silent again.
But this time the silence wasn't due to doubt—it was due to shock.
Fang Jianhua stared at Su Chen's handwritten derivation on the big screen, his pupils slightly contracting.
As the vice dean of the School of Microelectronics at Beihang University, Fang Jianhua could certainly understand these mathematical derivations. He could even spot the dimensionality reduction technique Su Chen used on page 113 at a glance—an extremely ingenious mathematical transformation that mapped the complex three-dimensional thermal stress field onto a quasi-two-dimensional plane.
This method... he had never seen before.
But he couldn't admit defeat.
"Ha." Fang Jianhua sneered, forcibly composing himself. "The industry uses finite element simulation because analytical methods cannot handle complex boundary conditions. This so-called 'equivalent thermoelastic approximation method' you presented—"
He emphasized the word "so-called".
"—No one in the industry has ever used this method before, and no peer-reviewed paper has verified its reliability. A method that is not recognized by its peers cannot be called scientific."
Another wavering voice appeared in the comments:
Indeed, methods that no one has used before are not necessarily reliable.
"But an error of 200mm is only 0.02 degrees!"
One experiment cannot prove anything.
Fang Jianhua finally breathed a sigh of relief when he saw the comments shifting in his favor again. As long as he stuck to the point of "no peer review," even if Lin Wei wore out her notebook, it wouldn't make a difference—
"Professor Fang."
Lin Wei's voice interrupted Fang Jianhua's thoughts.
She looked at Fang Jianhua with a peculiar expression in her eyes—not anger, not mockery, but a calm that was almost pity.
"You said this method 'has never been used in the industry before,' so it 'cannot be called scientific.'"
Lin Wei said, enunciating each word clearly:
"Then, Professor Fang, what's a method that everyone has used called?"
She answered the question herself:
"That's common sense."
A low gasp of surprise rippled through the live stream.
"It's precisely because no one has ever used it before that it's called innovation."
The comments section went crazy.
"Wow, that's brilliant!!"
"It's etched into our DNA"
"Innovation means doing things that no one has ever done before! Professor Fang's logic is inherently contradictory!"
"This woman is amazing..."
"Lin Wei! Lin Wei! Lin Wei!"
Fang Jianhua's face turned extremely ugly.
He opened his mouth, wanting to refute, but Lin Wei had already continued speaking:
"Furthermore, Professor Fang, you said there was no experimental verification. But I just showed you—the estimated value for a 200mm cavity is 8.73 degrees, and the measured value is 8.71 degrees, with an error of 0.02 degrees."
She looked into Fang Jianhua's eyes: "Is a 0.02-degree error enough to illustrate the point?"
Fang Jianhua fell silent.
0.02 degrees.
In the DRIE field, the computational accuracy of commercial finite element simulation software is typically between 0.1 and 0.3 degrees. In other words, if the measured error of this "equivalent thermoelastic approximation method" is indeed only 0.02 degrees, then its accuracy is actually higher than that of finite element simulation.
That's impossible.
Fang Jianhua's first reaction was—this is impossible.
How could a handwritten parsing method be more accurate than commercial simulation software that has taken decades to develop and billions of dollars to invest?
"Data can be forged," Fang Jianhua said, uttering the last rebuttal he could think of.
Lin Wei was not angry.
She simply stated, "The DRIE measurement data for the 200mm cavity is based on all complete third-party testing records from AVIC Institute 611. Professor Fang can request these from Institute 611 at any time if he has any questions."
Fang Jianhua fell silent.
He certainly knew what Institute 611 was—it was one of the core research institutes of the Aviation Industry Corporation of China, and their testing data had national-level credibility.
If the data from Institute 611 is also fake, then Fang Jianhua's questioning is not directed at Su Chen, but at AVIC (Aviation Industry Corporation of China).
He doesn't have the guts.
When Zhao Mingyuan saw Fang Jianhua fall silent, he knew the course of the live broadcast was irreversible. But as an outstanding media professional, he wanted to do one last thing.
"Ms. Lin Wei," Zhao Mingyuan began, "I have a question I'd like to ask you."
Lin Wei looked at him.
"How significant is Su Chen's 'equivalent thermoelastic approximation method' in the MEMS industry?"
Lin Wei thought for a moment and said, "I'm not qualified to answer that question."
She glanced at the camera: "But I know someone who can."