[Translating resource] A Japanese Stirling engine course

[gitPharm01]

Lesson 5 - 2 100Watt experimental Stirling engine: part C

This part talks about solving an issue of unbalanced power output from compression and expansion piston.
An adequate mechanical part with high precision can make drastic changes on the performance a machine.

This part starts from 14:12 to 22:53
https://youtu.be/9Risg8ItEjo?t=854

Though we completely assembled this engine, it failed in the test run.
First problem is in the crank system.
Here(left crank was linked to the compression piston and right crank was linked to the expansion piston).
Usually the power output from the expansion piston is stronger and unbalance between two cranks happened.
In order to solve this problem, a timing belt was used to link these two cranks and the shaft of flywheel.

Timing belt is depicted by dotted line in right half of this drawing.
(Picture stated at 14:58)
(Picture title: Timing belt method of fist design)
That shaft on top of two cranks is the shaft for power output.
Each crank had about half of their circumference connected to the belt.
Only a third of circumference of the power shaft was connected to the belt.
This made the situation quite difficult because the unbalanced power from cranks created unstable tension to the belt.
The belt were eventually tore by pulses of force.
We changed many types of belts and all of them failed.

We then started finding a mechanism which can endure such unbalanced torque while stably delivering the output.
It troubled us a lot and we finally found our solution.
(Picture started at 16:44)
(Picture title: 3 balanced cranks(triangular cranks) )
This mechanism was found in a book called <<Shinhen Kikai no moto / Fundamentals of machinery>>
This book was written in 1978.
Though this is an old book, the basic mechanism for linkages and other machinery are included in this book.
This book is very important to me.
(At 17:50 Took one copy of that book fro the shelf behind him )
This is my copy of that book.

This mechanism is depicted in the drawing on the left.
1,2,3 are rotating shafts and 4,5,6 forms up a triangular crank.
This crank system allows the three shafts(1,2,3) rotate at the same speed.
We were excited by this discovery and started trying to make this mechanism.
Sadly, we were not skilled enough to make it by ourselves.
We made our prototype, but the length of cranks are not precise enough.
This prototype crank generated pretty bad performance.
In the end we had to order this part from professional machine shops for precision better than 0.01mm.

This is the triangle crank we've been discussing.
(Picture started at 19:38)
(Picture title: Three parallel crank and coupling rods)
The shaft marked as "1" is for expansion side, "3" for compression, and "2" for flywheel.
We did try and error on this mechanism several times and eventually succeeded with that ordered part.

This is the final version of that compact Stirling engine.
(Picture started at 20:30)
(Picture title: Structure of 100W class small Stirling engine)
Its width is 304 mm and height is 462mm.
On the bottom part the triangular crank is installed.
The performance of this part was very impressive, high precision really makes the difference.

Above the crank there are two cylinder liners which cover the piston shaft.
The bottom of each cylinder liner is sealed by a o-ring and one slid bearing is used there to regulate the movement of piston shaft.

A cooler device covers the cold end of compression side.
It was made from copper pipes.
As for the heater on expansion side, we tried to use the original heater from previous build.
It was a block casing with holes and heating wires were inserted in that.
But since this build was too compact, there was no room for this part.
We had no choice but to use 1kW cable heaters.
The heater was wrapped on the hot side for two to three layers.

Since there are buffer pressure applied on this model, we decided to use a larger fly wheel.

[gitPharm01]

Lesson 5 - 2 100Watt experimental Stirling engine: part D

This part talks about the specification of the engine and its heat block.
In order to become compact enough, some modifications were made on the design.
The professor also shared his experience on metalworking as a student.

(This part is a little bit confusing since he mentioned in part C that a heating wire was used instead of heating block.)

This part starts from 23:00 to 30:15
https://youtu.be/9Risg8ItEjo?t=1383

[translation starts here]:

This is the specification of the compact Stirling engine.
(picture starts at 23:00)

partDpic1.png (148.23 KiB) Viewed 4060 times

(picture title: Specifications of small Stirling engine)
The bore is inherited from the boxer type Stirling engine.
In order to achieve a compact size, we had to shrink the stroke from 70mm to 50mm.
Thus Swept volume is much lower than the boxer model.
But still it managed to reach 100W output in the end.
Because a larger flywheel was used in this system, the total weight increased to 30Kg.
Without the flywheel, it weighted only about 20Kg.
Such weight raised the concern about actual portability of this engine.

And the target heat source of this build is biomass and waste heat.

There is actually one thing not fully covered in this build, which is the heat block.\
(Picture starts at 24:56)
(Picture title: Heat Block)

partDpic2.png (185.29 KiB) Viewed 4060 times

We considered many designs for heat block of the heater.
There are three types, type 1&2 are made from brass and type 3 is made from copper.
Brass is easy to drill holes on.
We used 2mm files to to drill holes on these blocks.
In type 1, there are 153 holes.
Each hole is 51.5 mm deep.
The block's diameter is 45.5mm.
The drilling process was actually a try and error process.
Once the drill tip deviated a bit, it will drill through adjacent holes.
We failed several times and lost a lot of material.
By calibrating the drill and having enough practice on drilling, we finally make the type 1 block with 153 holes.

We also made another block which has much higher hole density, which is the type 2 block.
These blocks took us about one week to complete.

These brass made blocks were much easier for us to process.
When it comes to type 3, the pure copper block, though it has excellent heat transduction, it was not easy to process.
We had to slowly drill holes while applying enough lubricant.
The type 3 has only 108 holes.

Here's the specs of these three:
(picture started at 29:59)
(picture title: Table 2 Specification of Heater block)

partDpic3_1.png (122.92 KiB) Viewed 4060 times

[gitPharm01]

Lesson 5 - 2 100Watt experimental Stirling engine: part E

Experiment settings for testing this engine is discussed in this part.
Measuring engine output is not enough.
In order to understand which design is better and why it outperformed others, things like pressure loss and temperature differences must be measured.
Regenerator design in this part is a good example.

This part starts from 30:16 to 39:29
https://youtu.be/9Risg8ItEjo?t=1817


(picture starts at 30:16)
(Picture title: Measuring system)

partEpic1.png (178.97 KiB) Viewed 4054 times

This is layout of a measuring system which can monitor and record data from many key points of this engine.
The purpose of this system is to collect data for further development.
It is crucial if we want to build engine models with higher power output.

The first thing is temperature.
The temperature around each cylinder was recorded by a Hybrid Recorder on top left.
And the pressure is also important.
Two sensors were connected to two cylinders of the engine and signals were transferred to a Strain Amplifier.
A/D converter(misspelled as comberter) will then send the processed signal to a PC.

A compressor depicted on top of this graph is used to maintain buffer pressure of this engine.
Such pressure can be measured and controlled by a valve.

We tested two regenerator designs and the related data will be discussed in this video.
In this test run, we will also use three different types of heat blocks from last part.
We also tested different buffer pressure settings.

Finally there is a motor connected to the output shaft of that engine.
A torque meter will detect and record the power generated by this engine.

Let's look at the experimental conditions.
(Picture started at 32:27)

partEpic2.png (132.19 KiB) Viewed 4054 times

(Picture title: Experimental Condition)
The contact area on hot side was too small and only 800W is available to this system.
The cold end was still cooled by water.
And four different levels of buffer pressure are applied on the engine:200,400,500,600kPa.
We did not use pressure higher than 600kPa in case that it might damage the hull.
Two types of working gas were used: air and Helium.

Let's take a closer look at the Temperature measuring points.
(Picture starts at 34:27)
(Picture title: Measuring point)

partEpic3.png (165.01 KiB) Viewed 4054 times

Seven points are installed with temperature sensor and data was collected from them.

We first tested which regenerator design is better for this engine.
(Picture starts at 32:54)
(Picture title: Regenerator Layer Pattern)
(Picture note: SUS304 60 mesh)

partEpic4.png (199.55 KiB) Viewed 4054 times

(at 35:05 please look at the bottom right of the screen,
the professor took one piston cylinder on his left hand)
This(the regenerator) will be on top of this cylinder.
We tried two methods to stack steel mesh in to this part.
Pattern 1 is staking them vertically for 70 layers.
Pattern 2 is stacking them horizontally for 276 layers.
(at 35:53, the professor showed a steel mesh on his hand)
The material used in this case is this, a stainless steel mesh.
Making them was yet another difficult and time-consuming job.

The result is presented here:
(picture started at 36:48)
(Picture title: pressure drop of each regenerator layer pattern)

partEpic5.png (150.63 KiB) Viewed 4054 times

Just like what we expected, pattern 2 have much denser structure and suffered much more loss of pressure.
However, the patter 2 outperformed patter 1 in temperature differences.


(Picture starts at 38:05)
(Picture title: Regenerator Layer Pattern)

partEpic6.png (133.53 KiB) Viewed 4054 times

In this graph, temperature difference between compression and expansion side was presented in Y axis.
This engine's maximum output was generated at about 500 rpm.
At this rpm, the difference between pattern 1 and pattern 2 is quite evident.
Pattern 2 has 100~50C more deltaT compared to pattern 1.
[Translation ends here]

Since this part is long enough, I will put the engine performance in next part.

[gitPharm01]

Lesson 5 - 2 100Watt experimental Stirling engine: part F

This is the final part of 5-2.
A conclusion is made for this experiment and I consider them quite valuable.

Previously I got confused by the term "heater" and "heat block".
I figured it out in this part.

This experimental model did use both.
A heat block has many holes and the working gas inside the corresponding piston cylinder can flow through it.
The heater is a segment of electric heating coil wrapped around this cylindrical block to provide heat.

Due to the compact size, there were not enough surface area for that coil to be fully connected to the block.
Thus heat input was merely 800 Watts.

This part starts from 39:30

https://youtu.be/9Risg8ItEjo?t=2370

This is the power output from pattern 1.
(picture starts from 39:35)
(Picture title: Engine Performance(Pattern1) )

partFpic1.png (140.32 KiB) Viewed 4050 times

It will take us too much time to thoroughly discuss this data set.
So we will skip some of them and focus on power output.
Power output in Watt and Engine speed in rpm are presented in this graph.
Four groups of data from different buffer pressure settings are denoted by different shapes of dots.
This graph is for regenerator pattern 1.
The first group is 200kPa, it reached top power output(40W) around 600rpm.
40 Watts is quite high compared to ordinary educational Stirling engines which usually have output in miliwatts.
But this is totally uncompetitive to regular gasoline engines.

When the buffer pressure increased to 400kPa, output power climbed up to 70 Watts.
It reached 80W with 600kPa buffer pressure, which is the best record for pattern1.
There were not as much power increase from 400kPa to 600Kpa.
This is because the actual heat input on hot side was limited to 800W.
If the heat input can be further increased, this graph will change.

And this is a comparison between regenerator pattern 1 and pattern 2.
(picture starts from 42:47)
(picture title: Engine performance of Each Regenerator Layer Pattern)

partFpic2.png (132.29 KiB) Viewed 4050 times

This graph only showed results from 600kPa buffer pressure setting.
Apparently, the pattern 2 generated better results.
It reached 90 Watts maximum output.
Though the pattern 2 lost more pressure difference than pattern 1,
it managed to outperform pattern 1 by maintaining higher temperature difference between two sides.

Though our ultimate goal was to reach 100W, this model was limited to 90W due to lower heat input.
Our simulated model suggest that without any loss to frictions and heat escaping to environment, it could have reached 100W.
Yet we still considered this result a success.

We previously talked about different heat block designs.
Here is the result from them.
(Picture starts from 44:19)
(Picture title: Maximum Shaft Power)

partFpic3.png (140.57 KiB) Viewed 4050 times

Three different types of heat blocks are represented in different colors.
Each type have four different buffer pressure settings:200, 400, 500, 600kPa.
Output power is on Z axis in Watts.
The type 1 and type 2 are made from brass, and the type 2 got more intense holes than type 1.
More holes means more surface area for working gas and thus type 2 outperformed type 1.

To our surprise, the type 3 which is made from pure copper and has much lower hole density still outperformed type 1 and 2.
The thermal conductivity of copper was impressive.
The conclusion is: Instead of manually increasing surface area, choosing a better material with higher thermal conductivity is better when it comes to heat block design.
We decided to use copper heat blocks in our following works.

We checked thermal efficiency of engines with different types of heat blocks.
(Picture starts from 46:35)
(Picture title: Thermal efficiency)
(the data depicted by diamonds are mislabeled as type 2, they are actually type 1)

partFpic4.png (140.44 KiB) Viewed 4050 times

(Another group is generated by using a pure copper block with 200 holes and working gas changed to Helium)
Four groups of data are depicted in this graph.
If the data get closer to the diagonal line, it means it is closer to ideal thermal efficiency.
The results from these groups are similar and it got more close to ideal condition than we expected.
A group using pure copper block with 200 holes and Helium as working gas is depicted as stars.
This group got slightly better.
We learned later that Helium can make a difference, but the pressure must be at the level of megapascals.

So here's our final conclusion:
(Picture starts from 48:36)
(Picture title: Conclusion)
1.Among all combination of test settings in this experimental model.
90W maximum output can be generated by the following combination:
Working gas: air
Regenerator: pattern 2
Buffer pressure: 6 atm

2.Heat block must be made from pure copper or materials with better thermal conductivity.

3.There is no significant improvements by changing working gas from air to Helium if the buffer pressure is below 6 atm.

4.This engine is actually close to Carnot efficiency.

We took lessons from this model and continued our research on other Stirling engine designs.

It was just for fun and we reversed the operation of this engine and turned it into a refrigerator.
(picture starts from 49:28)
(Picture title: Structure of Stirling refrigerator)

To finish this lesson, here is a complete footage video of the Stirling engine.
The buffer pressure was not applied and the crank system is exposed to the air.
(Video starts from 50:48)

[VincentG]

Did he say what temperature it reached as a refrigerator? Especially if we knew the compression/expansion ratio of the engine it would be useful to know the low temperature reached.

[gitPharm01]

Did he say what temperature it reached as a refrigerator? Especially if we knew the compression/expansion ratio of the engine it would be useful to know the low temperature reached.

Unfortunately, he did not provide any detailed data of this experiment.
From 49:28 to 50:20, he only mentioned that this system successfully did what he expected it to do and the existence of Stirling cryocoolers.

There is one LTD( delta T is about 100C ) type Stirling engine with 100W class output in Lesson 5-3.
I will start another thread to translate this part.

[gitPharm01]

There is one interesting model which is worthy of mentioning: Stirling engine powered RC plane

This is covered in Lesson 4-1 Educational Stirling engine models.
It is a propeller type published in 2001.
That engine weighs only 20 gram but have 30W output, quite incredible...
You can see footage of its flight in this video:
https://youtu.be/yjSoYE1beQI?t=1681

You can also see a drawing of this engine's design at 25:24.
Creator of this engine is 高山 秀雄(Hideo Takayama).

An official record can be found on Japan's JAA ( Japan Aeronautic Association )
http://www.aero.or.jp/record/rec-stirin ... elling.htm

[gitPharm01]

It is a propeller type published in 2001.
That engine weighs only 20 gram but have 30W output, quite incredible...

I must point out that that Dr. Doda mistakenly said 20 gram but it should weight 200 grams.
20 grams is impossible for that size of engine.

I found a follow up research done in 2004 by Hideo Takayama.
He made a beta type Stirling engine for rc aircrafts and it weights 280 grams.

0_2004.8_111.jpg (143.37 KiB) Viewed 4024 times