Proposed to be used as the skeleton of emergency relief shelters, covered with tarpaulin.


Experimental test carried out at African Bamboo plc ‘Product Development and Innovation Center’ in Addis Ababa, on 30 April 2015.
Objective was to investigate the use of the innovative construction concept – the ‘reciprocal frame’ – for application in constructing emergency relief shelters made of bamboo. Requirement is that the shelters should be easy to transport and simple and fast to construct using only basic skills. Obviously a ‘pre-fabricated’ design is indicated. Other requirements are strength and ability to support heavy cladding to make them better insulated in a cold climate and strong enough to withstand winds. The self-supporting reciprocal frame roof is claimed to use 33% less bamboo than a conventional roof would – ‘Reciproboo’.

Capture 2

Figure 1 shows the basic dimensions of the frame constructed.


The design of the reciprocal frame was based on information provided by ‘Reciproboo’, with some slight modifications.
A ‘pre-fab’ design was adopted, meaning poles can be pre-sized – i.e. cut to required length – at the raw material source, before being packaged and transported to the disaster site.
This approach saves on transportation cost, and also enables quick erection of the shelters on site.
To assist in quick and accurate assembly, the pre-sized poles are drilled with vertical holes at the intersection points of the frame members. This permits quick location of the strapping points at which the members are lashed together, enabling an accurately square frame without the need for physically marking out actual measurements.
The lashing rope is passed through the drilled holes during assembly.
A modular approach using the basic frame design may be adopted enabling shelters of various sizes to be constructed using either one or two frames.

Figure 2

Constructing the Reciprocal Frame.

Material Selection

Bamboo species: Yushania Alpina – local name ‘karkaha’ (Ethiopian highland bamboo)
Diameter: 4 to 6 cm outside diameter
Wall thickness: 6 to 10 mm (variable)
Mature culms, older than 3 years, are preferred to younger culms which have high moisture content and are more liable to split upon drying. Care must taken to ensure that there are no splits or cracks in the bamboo poles selected. Some few small holes bored by insects are acceptable.

Material Preparation

The bamboo poles used had been air-dried for about 4 weeks; estimated m.c. maybe 30 - 40% ?
Poles which had been pre-cut to 2.55 m length (for other research purposes) were used for the frame main members – 4 poles are needed for a single frame.
For the side support members, 2 poles of 3.15 m length were used.
Holes of 6 mm diameter were drilled transversely through the poles at the lashing points.
Figure 2 shows the drilling positions.
No particular attention was paid to the position of the lashing points with respect to the position of the nodes.


Figure 4

Frame Assembly

The reciprocal frame was assembled on a flat floor. Light plastic non-woven twine was used for lashing the frame members together.
The complete frame was supported on 4 short corner poles driven into the ground, and the side support poles were lashed to the corner poles. An ordinary ‘water level’ was used to check that the frame was horizontal.
It is observed that the top surface of the roof frame is slightly inclined with respect to the horizontal; the center part of the frame is about 9 cm higher than the edges.
The complete shelter assembly process, done by a 2-man team, should take less than an hour.
Figures 3, 4, and 5 show the stages of the assembly process.


Figure 5


Figure 6

Figure 7

Test Method

The aim was to determine the maximum point load that could be applied at the center of the frame before failure, i.e. breaking point.
A 100 liter plastic container was supported on a wooden board placed across the top of the frame. By filling the container with pre-measured amounts of water, the load applied to the frame can be measured.
The weight of the container and the wooden board was first measured before starting to fill water.
The container was filled with 10kg (or 10 liters) of water at a time.
In addition, a graduated ‘measuring stick’ was driven into the ground adjacent to one of the top corners of the frame. A mark on the adjacent roof frame member enables measuring the vertical deflection of the frame.
The deflection was measured after each filling of 10 kg of water.

Figures 6, 7, 8, and 9 show the test being carried out.


Figure 8

Figure 9


Breakage of 2 opposite frame members occurred at an applied load of 103.4 kg; more than the average weight of a man.
Breakage occurred after the horizontal (plane) position had been passed, at an unfavorable deflection beyond what would normally be the acceptable deflection in use.
Breakage of the frame members occurred at the middle lashing points. On the 2 poles that broke it was observed that this position was close to the middle of the inter-node, which is the weakest part of the bamboo culm.

Figures 10 and 11 show the frame after failure.

Failure occurred by crushing of the circular cross-section accompanied by longitudinal splitting.
Rate of deflection increased more rapidly as the breaking point was approached.
No sign of damage was observed on either the unbroken frame members or the side support poles, or the lashings.



The frame constructed utilized about 16.5 m length of bamboo, an estimated weight of about 20 kg.
The footprint covered is about 3 m by 3 m square, an area of about 9 m2.
The claim that the reciprocal frame uses 33% less bamboo than a conventional design cannot be substantiated unless a conventional design frame that carries the same breaking load is tested.
However, if it is assumed that a conventional design covering the same footprint would utilize 8 bamboo members of about 3 m length, i.e. a total of 24 m, the claim seems reasonable.
Breakage occurred under a ‘point load’ of 103.4 kg,, which converts to a ‘uniformly distributed load’ of
0.45 kN/m2.
(Inference with regard to wind load?)
Attention needs to be paid to the position of the middle lashing point on the main members when precutting the length of the poles.
The recommendation is that the drilled hole should be not more ± 10 cm from a node, considering that the inter-node distance of the bamboo used for the test frame is c. 50 - 60 cm.

Some suggestions were made by colleagues to:
improve the accuracy of the experiment, e.g. better uniform distribution of the load over the top four main members – in the test as carried out the load was direct on two main members
improve the load carrying capacity of the frame – by redesign of the frame, but still incorporating the concept of the reciprocal frame.

Prepared by: Chris Cleur Date: 23 May 2015

Figure 10

Capture 18

Figure 11

Comments on above report by S.Halbert 27/05/2015

1. Pole length dimensions.

It was good that you used the frame pole length dimensions that were available from other research purposes. The idea is that this shelter is adaptable to any different size tarps that may be provided by different aid agencies. I have been working with the IFRC 6m x 4m standard tarpaulins. These fit the shelters that had 9ft (2.75m) RF poles that we built at the Myanmar workshop.

The 2.55m RF poles you used are very close to the original steel RF poles I use at 2.60m...this also worked well with the standard tarps.
Although shelters with 3.00m RF poles work well these are on the upper limit for using the standard tarpaulins.

2. Overlap points .

80 cm spacing is fine. If a slightly larger outer shelter dimension was required a slightly less spacing would achieve this.

3. Length of overlap.

The 3cm you have used is less then that I usually recommend . I have worked with 5cm clearance, which means positioning the centre of the overlap at 7cm from the ends. In your case the drilled holes at 5cm from the end means that you are able to work with this shorter distance. I understand your reason for this ..... it is made possible by the passing of the lashing rope through the pre-drilled holes.

4. The pre-drilled holes

These will make assembly slightly easier as you say, however I am concerned about the resulting reduction in strength of the RF. The first steel prototype I built had holes drilled at the same points ( see attached) to facilitate assembly; eventually it failed under load at the drill hole point so I have not used poles with holes since. Over the last 4 years I have continually built and loaded the steel frame without these holes and not had a single failure.

5. Side support members

For the purposes of this loading assessment this frame is fine.
However the frame used here is between the 7 pole frame emergency shelter kit and the 12 pole single elevated shelter. It would actually be using 10 poles as the opposite upper and lower ridge poles and missing ( or alternatively the opposite placed side poles are missing).
I mention this as it determines the type of kit we are looking towards ie is it the 7 pole emergency kit or 12 pole single elevated shelter kit? This affects the figures for the final amount of bamboo used for each shelter as follows.

Emergency 7 pole frame shelter.

4 x 2.55m
1 x 3.15m
2 x 2m ( support poles)

Total= 17.5m

Single elevated 10 pole frame shelter.

4 x 2.55m
4 x 3.15m
4 x 2m ( support poles)

Total = 30.8m

6. Corner post attachment.

The diagram on page 5 shows a horizontal drill hole 10cm from the top, I presume it is for lashing purposes to secure the supported ridge pole?

7. Load testing.

This is indeed impressive for bamboo and exceeds reasonable loading as you conclude. It reminds me of the very early steel prototype testing where I hung a 30Kg water container at various points on the frame and measured the resulting deflection .

If we put this data together with the computerised data for bamboo obtained in the Galway University Engineering Department research we have a very good profile of the high strength of this frame.

These results should also be considered alongside the data obtained from wind tunnel testing of the steel storm shelter.

8. Observations.

Interesting that breakage occurred after the horizontal plane had been passed. I wonder if we can use this in future as an indication of a flexion limit not to be exceeded? This would have to be repeated, especially for different types of bamboo, but I think with dry bamboo it may be a useful nearing load limit indicator when people are loading the roof with insulation ie achieving the horizontal plane is a limit not to be exceeded. This will need verification..

I recorded the only other frame failure like this at the Nepal workshop with a shelter built with green bamboo. A student decided to “test” the roof by pulling it firmly downwards below the horizontal and one of the central frame poles broke at the internode point as in this test frame. As the green bamboo fibres did not “snap” but “bent” it did not cause a collapse of the shelter roof but a “sag” which was easily repaired with a replacement “splinted”pole alongside the original.

As you say the lesson here is to use the strength of the nodes wisely; suffice to say the ideal is to locate each overlap point as close to a lower pole node as practically possible.

Your finding that the two opposite overlaps failed is interesting. On review I think this is logical especially with a horizontal frame ( as loading on that second point suddenly increases with the frames tendency to rotate). I believe the dynamics in a roof that is at an incline will change again regarding the sequence of expected points failure and this needs further investigation.

9. Conclusions:

Bamboo saving for the roof. Yes, if we set a maximum gap between adjacent poles of 80cms ( as for the RF frame) then the 33% saving for roof bamboo applies.

The footprint can be considerably increased in tropical regions by pulling out the tarpaulin into an awning as shown at the Myanmar workshop

Suggestions that were made by colleagues:

to: 1) improve the accuracy of the experiment, e.g. better uniform distribution of the load over the top four main members – in the test as carried out the load was direct on two main members.
While this is a valid point I think the random distribution of the overlap ( in this case at 2 points) that you achieved is better in that it is closer to a real life situation. Bamboo is such a variable material that we need to keep testing these variables on the test table…once we “add in” the irregular physical properties of the RF frame we are really in the region of undocumented data
( and this is why this test is so valuable). If we were dealing with the steel frame I entirely agree that we could improve the accuracy of the experiment in the way suggested.

2) improve the load carrying capacity of the frame – by redesign of the frame, but still incorporating the concept of the reciprocal frame.
There is lots of scope with this suggestion. The ideal for all circumstances has yet to be perfected. I have no doubt that your colleagues, by working with this frame, will come up with ideas. One that intrigues me is a second reciprocal frame superimposed on the first and offset by 90 degrees….and also laid anticlockwise instead of clockwise. I have included this in the steel frame work on types of roof here but I believe it could also work for bamboo.

S.Halbert 27/05/2005

You are viewing the text version of this site.

To view the full version please install the Adobe Flash Player and ensure your web browser has JavaScript enabled.

Need help? check the requirements page.

Get Flash Player