171: Chapter 171 Let Machines Go Deeper Instead of Humans

September 22nd, 5:48 PM.

Room 402, Building 17, Zijing Apartment.

Jiang Lin clicked open the underground utility tunnel data packet that had completed decryption and verification the night before yesterday.

After entering the working condition directory, the data packet had only completed source, hash value, and read permission confirmation, and no technical judgment had been made yet.

According to the established scenario review queue of the Low Entropy Workshop, it was only handed over to Jiang Lin for processing today.

The operator only allowed reading through designated, filed terminals, forbade forwarding, did not open the main control system, and did not provide real geographic coordinates. The standalone working terminal in front of Jiang Lin was precisely one of the terminals authorized to read the data.

The first batch of data all came from real field tests, but the scope was limited.

The data packet only contained on-site video recordings, fault work orders, and three equipment modification records.

Raw operation logs such as drive current, module temperature, and motion status were still retained on the operator's server.

The screen immediately darkened.

A few seconds later, the screen lit up.

The footage came from a recorder worn on the chest of a maintenance personnel.

The narrow and long underground passage stretched all the way to the distance that lighting fixtures could not reach. On both sides were multi-layered pipes and cable trays, with tiny water droplets condensing on the walls and dark traces left by rainwater seepage and erosion on the ground.

The recorder swayed gently with the maintenance personnel's footsteps.

Further ahead, a dark gray tracked inspection robot was parked in the center of the passage.

The warning light on top was still flashing.

Neither the left nor right tracks were damaged.

The robotic arm was folded above the machine body.

The camera was also still working.

Yet it had completely lost its mobility.

The lower right corner of the screen retained the status information when the fault occurred.

[Executed Task Distance: 1.37 km]

[Battery Remaining: 63%]

[Main Drive Module: Overtemperature Protection]

[Task Status: Aborted]

Two maintenance personnel hooked a nylon tow strap onto the rear of the robot.

On the first exertion of force, the robot only moved backward a few centimeters.

The tracks rubbed against the damp ground, making a dull scraping sound.

The two people readjusted their stances. One pulled in front, and the other pushed from behind. The machine originally responsible for replacing them to go underground to complete inspections was thus dragged back by them step by step.

The total length of the video was seventeen minutes.

There was no editing in the middle.

The recorder's breathing became heavier and heavier.

One of the maintenance personnel stopped when passing the third fire compartment, took off their gloves, and rested against the wall for two minutes.

When the progress bar reached the end, the robot still had not been dragged out of the verification section.

Jiang Lin rewound the video to the beginning and looked at the fault status once more.

The battery still had sixty-three percent.

After participating in the World Robot Conference, the Low Entropy Workshop had registered one hundred and thirty-four scenario requirements. After eliminating shell financing, show-case procurement, projects without clear working conditions, and those attempting to extract core parameters, only seven ultimately remained, with the underground utility tunnel's comprehensive priority ranking second.

What the project party wanted was also very clear.

They did not need the robot to run very fast, let alone make eye-catching movements.

The total length of the verification section was 2.8 kilometers.

The machine set out from the entrance, completed image capture, instrument reading, leak identification, and environmental recording, and then returned along the original route.

One round trip, 5.6 kilometers.

In three field tests meeting the established waterproof and dustproof requirements, the existing platform only went as far as 1.5 kilometers at most.

It could go in.

But it couldn't come back.

Jiang Lin continued to scroll down.

In addition to the on-site video, the project party also submitted three modification records.

The first time, suspecting that the battery output capacity was insufficient, they replaced the original battery pack with a model having a higher discharge rate.

The fault remained.

The second time, they increased the motor and driver margins.

The robot ran faster and was more powerful.

But the overtemperature protection was triggered eleven minutes earlier.

The third time, they installed a larger aluminum heatsink inside the control cabin and added two circulation fans.

The shutdown time was postponed by eight minutes.

However, the total weight of the machine increased by 6.4 kilograms.

The project briefing concluded with the current assessment.

[The heat dissipation capacity of the sealed cabin body is insufficient; it is recommended to redesign the shell and further reduce system power consumption.]

At six o'clock sharp, the scheduled technical window opened.

Liang Zhixia dialed into the meeting from the Jiangcheng office.

The project party's engineering lead and thermal design engineer also appeared on the other side of the screen.

After brief pleasantries, the engineering lead directly opened the floor plan of the verification section.

"Four types of inspection platforms have already been tried in this section of the utility tunnel."

"Four-wheeled ones cannot pass through the drainage ditch, and tracked ones have too much attitude fluctuation when passing through expansion joints. There is another model that can complete the whole journey, but the control cabin cooling vents must be opened, which fails to meet our waterproof and dustproof requirements."

"We have seen the test of G-01C at Hengtai."

"We need to confirm whether it can make an underground utility tunnel variant."

Liang Zhixia asked, "What are the task requirements?"

"Continuous operation for no less than four hours. One round trip of the verification section, with no manual intervention allowed midway."

"Environmental boundaries?"

"Normal temperature between 28 and 35 degrees Celsius. Local humidity approaches saturation during the rainy season. There may be shallow accumulated water on the ground, but complete submersion is not considered."

"Project cycle?"

"We hope to complete the first verification within fifteen days."

Liang Zhixia did not accept this time requirement.

"The Low Entropy Workshop can currently only conduct feasibility reviews. Whether we can enter joint verification depends first on reviewing the underlying fault logs that preserve the common temporal relationship."

"We have already provided the fault report."

"We need the fault data."

The project party's engineering lead glanced at the thermal design engineer beside him.

"The control strategy and some hardware parameters involve existing suppliers."

"They can be desensitized," Liang Zhixia said, "but we cannot be provided with only conclusions."

Jiang Lin brought up the fault footage in the video.

"For the second experiment, from task start to two minutes after the overtemperature protection was triggered, are the drive current, driver temperature, cabin temperature, motion status, and position logs all retained?"

The thermal design engineer replied, "Yes."

"Where are the temperature sampling points?"

"One is inside the main drive module, output by the drive chip. The other is in the middle of the control cabin, about twenty centimeters away from the main drive module."

"Shell temperature?"

"It was only measured with an infrared thermal imager in a few experiments, without full-process recording."

"What about the timestamps when the robot passed through expansion joints and drainage ditches?"

"They are in the motion logs."

Jiang Lin placed the three experimental briefings side by side on the screen.

The three shutdown locations were different.

The first time was 1.37 kilometers.

The second time was 1.19 kilometers.

The third time was close to 1.5 kilometers.

There was also significant differences in the running durations of the three times.

The three briefings showed that the three experiments used the same verification route and identical task processes, and the ambient temperature did not exhibit changes sufficient to explain the shutdown differences.

In the second experiment, a motor and driver with higher power were replaced, and the robot's average speed increased, yet the overtemperature protection was triggered eleven minutes earlier than the first time.

If the fault only stemmed from the slow overall heat accumulation of the sealed control cabin in the same environment, the second experiment with a shorter running time should not have triggered protection the earliest.

Yet the facts were precisely the opposite.

"Send over the raw logs of the second experiment," Jiang Lin said.

"All of them?"

"Just this complete segment of logs. Absolute time, location, and equipment numbers can be replaced, but the drive current, the two temperatures, motion status, and position logs must retain the same relative time zero point."

The engineering lead did not directly retrieve the data.

He turned sideways and confirmed a few words with the data security lead waiting outside the meeting room.

A minute later, that lead joined the meeting.

"Will a common relative timeline expose the actual operation shifts?"

"Set the moment when overtemperature protection is triggered as T0. The logs start from task initiation and are retained for 120 seconds after T0. All channels use the same T0. We don't need the date, nor do we need to know whether it happened on the day shift or night shift."

"What about position data?"

"Only retain the relative distance starting from the starting point, as well as anonymous event tags for expansion joints, drainage ditches, and water accumulation areas."

"Supplier parameters?"

"Currents can be normalized by rated values. The overtemperature protection line is retained, while the model and specific rated values can be hidden."

The data security lead wrote these contents into a temporary application form.

[Application Scope: Second Shutdown Experiment Fault Window]

[Time Range: Task Initiation to T+120s]

[Time Processing: Common Relative Time Zero Point]

[Zero Point Definition: Main Drive Module Triggers Overtemperature Protection]

[Position Processing: Relative Distance and Anonymous Event Tags]

[Parameter Processing: Rated Value Normalization]

[Reception Permission: Low Entropy Workshop Read-Only Data Area]

The application required joint confirmation from the engineering lead, data security lead, and supplier liaison.

The meeting was temporarily suspended.

Jiang Lin did not urge them.

Lacking official authorization, even if the data was already placed on the other party's server, one should not cross the permission boundary to enter the Low Entropy Workshop.

At 8:14 PM, the three-party confirmation was completed.

The data packet that had undergone truncation, translation, anonymization, and hash verification was sent into the Low Entropy Workshop's read-only data area. The original files remained on the operator's server, and the Low Entropy Workshop could only read this approved section of the fault window.

At 8:20 PM, the technical meeting restarted.

Jiang Lin did not look at the temperature first.

He overlaid the motor current curve and the position record together.

For the first seven hundred meters after the robot entered the verification section, the main drive current changed steadily.

At 730 meters, passing the first deformation joint, the current showed three consecutive peaks.

At 860 meters, the ground water accumulation deepened, a slight speed difference occurred between the left and right tracks, and the control system continuously corrected the direction.

At 1,080 meters, the right track rolled over the edge of a drainage ditch.

Every time it passed a ground change, the current curve would spike upward with a row of sharp peaks.

Viewed individually, none of them exceeded the driver's rated range.

Yet another temperature curve did not recover after the spikes ended.

It climbed upward tier by tier.

After the first correction, it increased by three degrees.

The second time, it increased by five degrees.

The third time, it only dropped by one degree before the next round of current peaks had already arrived.

The air temperature in the middle of the control cabin never exceeded 47 degrees.

Inside the drive module, however, it crossed the protection line at 1.19 kilometers.

Jiang Lin magnified the two temperature curves separately.

One was gentle.

One was steep.

In the same sealed cabin, the two sampling points were twenty centimeters apart, and the maximum temperature difference between them approached forty degrees.

The project party's thermal design engineer stared at the curves.

"The internal temperature of the driver is always higher than the cabin temperature. We know this."

"Where is your heatsink installed?"

"Above the drive board."

"How is it connected to the shell?"

"There is no direct connection. The heatsink releases heat into the cabin air, which is then dissipated out by the shell."

"Is this a structure left over after the sealed version modification?"

"The original platform had air inlet and outlet vents. After the project required upgrading the sealing level, we retained the internal air ducts and changed the external circulation to internal circulation."

He opened the control cabin structural diagram.

Two fans were installed near the heatsink.

The airflow passed through the aluminum fins, flowed backward along the top of the control board, and then returned from the side of the battery compartment.

From the airflow diagram, this was a complete cycle.

And the problem lay right here.

Heat left the driver and entered the control cabin air.

The fans then sent the already heated air back near the driver.

A portion of the heat could slowly dissipate to the outside through the cabin wall, but more heat lingered inside.

On stable, low-load flat roads, this structure could barely maintain balance.

Once the robot continuously crossed expansion joints, water accumulation, and drainage ditches, the local thermal pulses generated by the driver would exceed the transport capacity of this heat dissipation path.

"After you enlarged the heatsink, how much did the module temperature drop?" Jiang Lin asked.

"It dropped by six degrees under stable load. Before the on-site shutdown, it dropped by less than two degrees."

"What about the shell?"

"Little change."

"The heatsink became heavier, yet the shell did not noticeably heat up."

The project party's engineering lead caught the meaning behind this sentence.

"The heat didn't reach the shell?"

"At least not in time."

Jiang Lin marked the main drive module on the structural diagram.

Then, along the heatsink, cabin air, and shell, he drew the current heat transfer path.

That path took a very large detour.

After the driver generated heat, it first had to pass through the thermal pad to reach the aluminum heatsink. Then the heatsink heated the air inside the sealed cabin, and finally relied on the air contacting the cabin wall to transfer the heat to the shell.

It possessed a large heat dissipation area.

What it lacked, however, was an exit that was short enough and capable of withstanding instantaneous heat flux.

"Can you provide a set of the same model drive module, sealed control cabin, and motor load components in Beijing?" Jiang Lin asked.

The engineering lead said, "We have the same model motor, and the load components are also at the Beijing test station."

As he spoke, he glanced at the control cabin structural diagram and asked again, "The whole machine won't be used?"

"Not for now," Jiang Lin said. "I want to replay the load spectrum experienced by the main drive module during continuous obstacle crossing, without needing to reproduce the entire utility tunnel."

"Are the motor and drive module configured according to the original platform?"

"Keep them consistent. The load peaks, durations, and intervals should be restored according to the newly approved fault window."

The engineering lead looked at the time.

"It can be delivered tomorrow morning."

Liang Zhixia took over the conversation.

"Send it to the Beijing R&D Center, and confirm the sample list, ownership description, and disassembly permissions tonight."

"What about the fifteen-day verification cycle?"

"Let's talk after the minimum reproduction is completed."

When the meeting ended, the project party did not get an answer as to whether G-01C could run continuously in the utility tunnel for four hours.

The two sides only confirmed the paid fault reproduction of the first stage.

[Task Scope: Local Heat Accumulation Replication in the Sealed Control Cabin]

[Test Object: Project Party's Existing Main Drive Module]

[Deliverables: Thermal Path Bottleneck Determination and Original Test Records]

[Exclusions: G-01C Complete Machine Solution and Four-Hour Continuous Operation Commitment]

Liability for sample damage, disassembly permissions, and test data ownership have entered the electronic confirmation process.

They first figured out why that robot had stopped 1.37 kilometers away.

...

September 23, 3:17 PM.

Low Entropy Workshop Beijing R&D Center.

The access control system at the entrance had been installed, but the protective film on the floor had not been completely peeled off. Leaning against the wall were several unopened equipment boxes, with procurement numbers and acceptance statuses pasted on the outsides of the boxes.

In the center of the room, two heavy-duty workbenches were already in use.

Independent circuit.

Programmable DC power supply.

Multi-channel temperature recorder.

A small infrared thermal imager.

And a small motor dynamometer table sent by the project party along with the sample, plus a temperature-controlled enclosure capable of housing the sealed control cabin.

Dual motors of the same model were already fixed on the load rack, and the torque and speed sensors had completed calibration.

The tool wall only hung a batch of the most commonly used wrenches, screwdrivers, and measuring tools, leaving plenty of empty spaces.

The sealed control cabin sent by the project party was placed on the left workbench.

Jiang Lin opened the upper cover.

The internal structure was basically consistent with the drawings.

The battery compartment was located at the rear.

The computing module was in the middle.

The main drive board and two circulation fans were crammed into the front end.

The heat sink was large, and the fin arrangement was neat enough. Just looking at the materials used, it was hard to associate it with insufficient heat dissipation capacity.

Jiang Lin put on an antistatic wrist strap and fixed the four thermocouples in sequence.

The first one was attached close to the power device of the main drive module.

The second one was fixed at the root of the aluminum heat sink.

The third one was suspended at the air temperature sampling point inside the cabin.

The fourth one was attached to the inner wall of the shell.

The thermal design engineer from the project party stood on the other side, completing the circuit check.

"External data only shows the rated current ratio," the thermal design engineer said, "The dynamometer table has been restored according to the true rated values of this set of motors and drives. The amplitudes, durations, and intervals of the three sets of load peaks are all retained."

"What about the speed changes?"

"Replay according to the second field experiment. All channels use the common relative timeline approved last night."

"Overtemperature protection value?"

"Keep the original drive settings."

"Power voltage and fan control?"

"Consistent with the field."

Jiang Lin closed the cabin cover.

The bolts were tightened in diagonal order.

The sealed control cabin was moved into the temperature-controlled enclosure.

The dual motors and dynamometer table remained outside the box, and the power and control lines were connected to the main drive module inside the cabin through sealed wire-passing ports.

The ambient temperature of the enclosure was set to 33 degrees, consistent with the records of the second field experiment.

The test began.

The enclosure ran at 33 degrees until the shell of the control cabin, the air inside the cabin, and the main drive module returned to the set initial temperatures.

After the three measuring points were continuously stable for ten minutes, the second field load spectrum began to replay.

During the first ten minutes after the load spectrum began to replay, the two motors maintained a steady speed, and the four temperature curves slowly rose together.

Subsequently, the dynamometer applied three groups of short-term braking torques in turn, simulating the load changes when the robot crossed expansion joints, corrected direction in accumulated water, and when the right track rolled over the edge of the drainage ditch.

The main drive module was still modulated by the original controller, and the current peaks, durations, and intervals were all reproduced according to the field records.

At the thirteenth minute, the first set of load peaks appeared.

The temperature of the power device jumped up by three degrees.

The root of the heat sink rose by one degree.

The shell temperature barely changed.

At the twenty-first minute, continuous direction correction was replayed.

The temperature of the power device rose again.

The fan speed had reached the set upper limit. The air inside the sealed cabin circulated constantly, and the airflow speed behind the heat sink was not low.

Yet heat still accumulated near the drive board.

At the thirty-fourth minute, the power device exceeded 80 degrees.

The cabin air was 46 degrees.

The shell was 42 degrees.

The project party's engineering head looked at the four sets of numbers through remote video.

"The shell is still this low?"

The thermal design engineer did not explain.

His attention had entirely fallen on the temperature of the power device.

At the forty-first minute, the test bench issued a short prompt.

[Main Drive Module: Overtemperature Protection]

[Output Cut Off]

The load returned to zero.

The two circulation fans were still running at high speed.

This time there was no 1.37-kilometer underground passage, nor accumulated water and expansion joints.

There was only a sealed control cabin placed on the workbench.

The failure still occurred.

Jiang Lin stopped recording and waited for the internal temperature to drop.

Then he opened the cabin cover and removed the heat sink.

Between the bottom of the heat sink and the drive module was a thermal conductive pad. Its thickness served to balance electrical isolation, assembly tolerances, and vibration cushioning, but it also brought higher thermal resistance.

There was nothing wrong with the heat sink itself.

There was nothing wrong with the fan either.

The problem happened after the heat left the heat sink.

It could not find a channel fast enough to reach the shell.

Jiang Lin took out a section of temporary copper braid from the toolbox.

One end was pressed against the heat sink base.

The other end was connected to the metal reinforcement rib on the inner wall of the sealed cabin via an insulating gasket and a temporary clamp.

This was not a structure that could be directly put into the field.

The copper braid might fatigue during long-term vibration.

The connection position might affect the seal.

The insulating gasket would introduce new contact thermal resistance.

After the heat was introduced into the shell, one also had to consider the maximum allowable surface temperature, condensation after the end of thermal cycles, aging of sealing materials, and drift of nearby sensors.

But it was enough to complete a minimal verification.

The thermal design engineer took the initiative to fetch an insulation tester to confirm that the temporary connection did not destroy the original electrical isolation.

"Not modifying the original structure first?"

"Just add a thermal channel."

"The shell temperature will rise."

"It should rise."

The two re-sealed the control cabin.

The control cabin re-entered the 33-degree environment, and the four measuring points gradually fell back.

Until their deviations from the initial values of the first round of tests did not exceed half a degree and remained stable for ten consecutive minutes, Jiang Lin reloaded the same load spectrum.

The second round of tests began.

To verify whether the newly added thermal channel could span a complete failure cycle and continue to maintain stability when the same load appeared again, Jiang Lin continuously replayed the 41-minute field load spectrum twice.

Same ambient temperature.

Same current peaks.

Same cycle intervals.

At the thirteenth minute, the temperature of the power device still showed its first rise.

The temperature of the copper braid rose accordingly.

A few minutes later, the shell temperature also began to change.

At the twenty-first minute, continuous direction correction was replayed again.

In the first round of tests, the temperature of the power device began to accumulate rapidly from this moment.

This time, the heat traveled along the copper braid into the shell reinforcement ribs, and then diffused outward from the larger metal surface.

At the thirty-fourth minute.

Power device at 71 degrees.

Shell at 49 degrees.

Seeing that the shell temperature was seven degrees higher than the first time, the project party's engineering head subconsciously leaned forward toward the screen.

"The shell is hotter."

"Yes," Jiang Lin said.

"Is this considered an improvement?"

"At the end of the first round of tests, the shell was only 42 degrees, but the drive had already overheated."

Jiang Lin superimposed the curves of the two rounds together.

"Now the shell has taken away the heat that originally accumulated near the drive."

At the forty-first minute.

The time point when the first round of tests triggered protection had arrived.

The load was still running.

After the end of the first complete load cycle, the dynamometer table did not shut down.

The same field load spectrum restarted from the beginning.

The three sets of current peaks reproduced in turn in the second cycle, but the temperature of the power device no longer showed a trend of getting out of control step by step upward.

The temperature of the power device stabilized around 74 degrees.

The shell temperature was 52 degrees.

No overtemperature prompt.

No output cutoff.

The test continued.

At the fiftieth minute.

At the seventieth minute.

The temperature of the power device was still slowly rising, and then stabilized around 76 degrees.

The shell temperature rose to 54 degrees.

The two circulation fans always operated according to the same control curves as the first round. The power voltage, ambient temperature, load peaks, and cycle intervals also did not change.

Between the two rounds of tests, the only newly added variable was that temporary thermal channel connecting the heat sink base and the shell reinforcement ribs.

At the eighty-second minute, the replay of the last segment of high load ended.

Jiang Lin terminated the test.

Overtemperature protection was never triggered.

It was quiet beside the workbench for a while.

The project party's engineering head re-called the initial failure footage.

In the footage, two maintenance personnel were dragging the immobile robot in the underground passage.

"So, we have always been cooling down the control cabin."

He said.

"What really needs to be dealt with is that small area near the drive?"

"More accurately, it is the speed at which heat leaves that small area," Jiang Lin said.

"Can changing to a copper tape solve it?"

"No."

Jiang Lin disassembled the cabin cover and pointed to the temporary fixed point.

"It can only prove that the failure judgment is valid, and also prove that directly establishing a thermal channel is effective."

"The field version also has to deal with vibration, fatigue, electrical isolation, sealing, surface temperature, condensation, and thermal path changes under different postures. The copper braid has not undergone any life testing and cannot be put into the pipe gallery."

The engineering head clicked open the project plan.

"Is it still possible to complete the first on-site verification within fifteen days?"

"Do the combined environmental and vibration test first."

"How long will it take?"

"You fill in the shell materials, internal space limits, maximum allowable surface temperature, and long-term load spectrum. Low Entropy Workshop will give the first version of the verification plan tomorrow."

"Continuous operation for four hours?"

"No commitment for the time being."

Liang Zhixia changed the project status from scene consulting to:

[Engineering Preliminary Research]

No complete machine procurement was signed.

No four-hour indicator was written in.

Nor did it use Jiang Lin's personal achievements to provide any safety endorsement for the project.

After the completion of the first-stage paid failure replication, the project party agreed to advance subsequent work to the thermal path engineering preliminary research.

This was the second type of Real World scenario that Low Entropy Workshop truly entered after Hengtai Mine.

Underground comprehensive pipe gallery.

After the meeting window closed, Jiang Lin printed out the curves of the two rounds of tests.

On the first piece of paper, the shell always maintained a relatively low temperature, while the drive temperature rushed all the way to the protection line.

On the second piece of paper, the shell was hotter, but the drive stabilized instead.

Many people subconsciously believed that the cooler the machine shell, the better the heat dissipation.

Yet in this sealed control cabin, the icy shell precisely proved that the internal heat had not yet found an outlet.

Jiang Lin took down the temporary copper braid and put it into an evidence bag.

The label read:

[Only Used for Thermal Path Hypothesis Verification]

[Forbidden as a Field Solution]

Subsequently, he opened two files that had been sealed in a separate directory since the eleventh return from the Wasteland.

[MPS _ HeatRouter _ Reality _ v2.0]

[G01C _ HeatRouter _ FieldPack _ v1.0]

Nine Real World material and structure solutions appeared on the screen.

Copper.

Aluminum.

Pyrolytic graphite film.

Ceramic insulation sheet.

Flexible thermal connection.

Micro heat pipe.

Phase change buffer layer.

Structural shell.

State machine temperature rise slope interface.

None of the materials, micro-channels, and phase-change media on the future outer ring fragments were listed.

The Real World did not have those things.

What Low Entropy Workshop could use were only ordinary materials that could be purchased, processed, tested, and replaced in 2022.

But the engineering principle extracted from the future ruins was already clear enough.

There were many heat sources inside a machine.

They did not generate heat at the same time.

Nor did they generate heat at the same speed.

What truly needed to be managed was not a general overall machine temperature.

Instead, during different task stages, the heat at the most dangerous location was timely sent to areas with remaining capacity, and then slowly released after the thermal peak passed.

Jiang Lin created a new project directory.

[G01C _ HeatRouter _ PipeGallery _ v0.1]

Entered the first test target.

[Verification Section Round Trip 5.6 km]

The second one.

[Continuous Operation Not Less Than 4h]

The third one.

[Sealing Grade Not Decreased]

The fourth one.

[Any Newly Added Thermal Path Must Not Interfere with Foot-End Sensing, Drive Control, and Safety State Machine]

After writing, he re-opened the failure footage.

The two maintenance personnel were still dragging that robot.

The lighting lamps dragged their shadows very long.

That machine was originally used to replace humans entering underground.

In the end, it made two people walk deeper for it.

Jiang Lin closed the video.

The control cabin on the workbench had completely cooled down.

The shell recovered the same temperature as the room temperature.

That section of temporary copper braid lay quietly in the transparent evidence bag.

It was still far from being an answer.

Yet it had already pointed out the first way out for the heat accumulated in the drive.