The military-industrial scientific research system of the academic master.

"what?"

It was not Xu Yang who made this exclamation.

In fact, she just looked up at Chang Haonan in surprise, and her expression didn't even change much.

On the contrary, several other technicians from Institute 624 standing nearby were looking at Chang Haonan and Xu Yang, who were sitting on both sides of a table, with somewhat blank eyes.

To be honest, this can't be entirely their fault.

The two people just communicated without even using a pen, which was indeed a bit unfriendly to other people around them.

Even though quite a few of them were directly involved in the design of this simulation verification program, it was still difficult to follow the ideas introduced by Xu Yang just now.

As a result, before I fully understood the previous content, Chang Haonan had already given his opinions.

Naturally, there will be a feeling of "looking down and picking up a pen in math class".

Facing Xu Yang's questioning eyes, Chang Haonan finally picked up a pen from the side:

"There is no problem with the algorithm and control logic of the verification program. What I mean is that there are still some test contents that have not been taken into account."

When the people who were scattered around the room just now saw this scene, they almost took a few steps forward in unison and gathered behind the two of them.

Chang Haonan instantly felt that the light hitting the paper had dimmed...

"Ahem... Engines equipped with full-authority digital control systems generally will not be equipped with a complete mechanical control system. Therefore, various working conditions that may be faced must be considered, such as when network induction occurs due to various reasons during flight. Delay and packet loss."

"Needless to say, time delay. If there is a significant gap between the control command issued from the operating end and the response from the action end, I am afraid that no matter how good the pilot is, he will not be able to control the aircraft normally."

"As for packet loss... For the TTCAN bus used by FADEC, if the data packet fails to be sent, the protocol will only allow the reference message to be retransmitted immediately, but not other data. This will significantly affect the accuracy and speed of system control, and even cause the system to Unstable.”

Chang Haonan explained while writing:

"For example, we can first analyze the simplest form of delay and packet loss. Assuming that the delay is short and constant in each sampling period, and the number of consecutive packet losses also has an upper limit n, then within a sampling period , there may be two controller outputs reaching the actuator end..."

"..."

"The dynamic characteristics of the engine operating near the steady-state point can be simply described by a continuous state space equation. If this equation is discretized, a 2n+1 seed system will appear. The actual engine operating characteristics will be in these 2n+1 kinds. Random and continuous jumps between them significantly prolong the convergence process of the steady-state system..."

Having said this, Chang Haonan stopped the pen in his hand, pushed the two pieces of paper filled with formulas to the front, and then raised his head.

This calculation made the people around him stunned.

But fortunately, this time at least I was talking and writing, and there was a lot of content on paper, so compared to the complete confusion just now, at least a few people understood the key points.

"Ours is a wired connection, so there shouldn't be particularly exaggerated delays or packet loss..."

One person asked with some hesitation.

Chang Haonan just took a drink of water at this time, quickly swallowed it, and then shook his head:

"In fact, just like our normal computer Internet access, the signal transmission bandwidth of the engine's ECU unit and cables is actually not as good as that of home computers and network cables. Therefore, even during a normal flight, the problem of unsmooth information transmission is very serious. normal."

"Then...doesn't this mean that FADEC has shortcomings in stability?"

Another voice was obviously trembling:

"After all, the delay of the mechanical control system is certain, as long as you adapt to it, it will be fine, and there is no data loss problem. However, these electronic controls occur randomly. Do you need to keep a complete mechanical system as a backup? "

A huge advantage of FADEC is weight reduction. It is naturally unrealistic to retain mechanical backup.

However, since Chang Haonan has raised this problem, there must be a solution:

“Although electronically controlled delays and packet losses occur randomly, the impact on performance can still be reduced or even eliminated through highly robust control algorithms.”

"So when I designed the engine control program, I used a distributed control gain matrix to improve the stability margin for delay and packet loss. But how much is this margin specifically and can it satisfy the fighter under various operating conditions? The following requirements still need to be verified during the testing process.”

"If the stability margin is not enough, then you need to add another engine control unit. If the margin is excessive, you may have to reduce one. After all, an ECU and related cables together weigh dozens of kilograms... …”

"Distributed control gain matrix...this...is still possible?"

Jiang Fuhe, who had been sitting next to Chang Haonan just now, suddenly had a flash of enlightenment in his eyes.

After turning the research and development focus of the 624 Institute to vertical take-off and landing engines last year, the biggest problem before him was the control response and stability of the engine's vertical power output.

You must know that vertical take-off and landing are not like horizontal flight. The impact of larger or smaller thrust will not be great. If the thrust output is unbalanced during this process, the aircraft will lose stability or crash directly.

However, the high-pressure gas during vertical takeoff and landing has to undergo a deflection of 90° or even greater, and the possible external disturbances are more complicated...

Therefore, the requirements for the engine control system are much higher than those for ordinary engines.

In the current testing process using the external vertical nozzle of the old model turbojet 7 engine, even if an electronic control unit is used, the reliability of the vertical output process cannot be guaranteed.

But Chang Haonan's moment just now gave him a complete idea.

Jiang Fuhe only feels that it is a pity that he is not an expert in the field of control algorithms, so now he has an idea but it is difficult to implement it...

Xu Yang, who was sitting opposite, had just taken the two pieces of paper into her hands. While others were still asking questions, she had already read the contents quickly:

"In this case, I only need to add a packet loss and delay simulation to the test control system...and essentially, packet loss can also be equivalent to the input delay of the engine..."

"Therefore, it can be considered that there is a delay at the input end of the engine that increases synchronously with time, so that the control input signal received by the engine is only related to the measured value..."

"Of course, the result measured by this method is the maximum number of consecutive packet losses allowed by the system, not the packet loss rate. It should be fine, right?"

"can."

Chang Haonan nodded decisively:

"If you use the maximum number of consecutive packet losses, you just need to slightly change the limit parameter setting logic. This is a completely independent unit, and it only takes a few hours."

"Then wait for my news and I'll get it done for you within two days."

Xu Yang threw the two pieces of paper on the table, stood up, called the names of several 624 engineers, and hurriedly left the control room with them.

"Mr. Chang."

After the room became deserted again, Jiang Fuhe approached Chang Haonan again:

"Can you give me a complete idea about the distributed control gain matrix you just mentioned?"

The latter was slightly stunned.

Mainly because I didn't expect anyone to ask this kind of question.

Because this is part of the design process.

But after thinking about it, it makes sense. After all, Institute 624 does not only do aerospace engine testing, they are also a turbine power research institution.

Even in the previous life, many core machines were built by 624.

But now, because of him, he has become almost transparent...

"This is actually inspired by the flight control program algorithm on the host..."

Chang Haonan said and took the pen and paper again.

He wasn't lying about this.

It was indeed inspired by the flight control program.

It's just that it's not our own flight control program, but an American one.

A huge problem with the flight control of later F35 fighter jets is that there will be huge rudder control delays under certain flight conditions.

Until Chang Haonan was reborn, this problem had not been solved...

So when he developed the turbofan 10 control program, he almost thought of this issue first.

Of course, it was impossible to tell Jiang Fuhe these contents, so Chang Haonan told another version of the story.

Anyway, as long as the knowledge you teach is correct.

Looking at the draft paper that was gradually filling up in front of him, Jiang Fuhe, who was still a little angry before, became calmer.

"Mr. Chang, our institute currently has a vertical take-off and landing engine project. Look...after the Turbofan 10 project is completed, are you interested in taking over?"

People must learn to reconcile with reality.

If you can't beat me, just join, there's no shame in it...