Raj,

Ultrasound is a common non-destructive tool for investigating cracks in structures, so 'yes' is the answer to your first question.

I sense that you are using a relatively low frequency transducer in air, in a monostatic sonar arrangement. Perhaps something like this:

http://www.microbric.com/p/4424020/ultrasonic-range-finder---suit-microbot-robot-kit.html

Firstly, how wide is the aperture that you are trying to detect? Is it smaller than the Rayleigh limit? If so, then the gap will be essentially undetectable. This is basic 'optical' theory and applies equally well to ultrasound as it does to light.

If we assume that your sonar operates at, say, 100kHz, then for a transducer 1cm in diameter the half-angle is, 340/100000/0.01, which is a bit less than half a radian.

So if your robot is 1m from the 'gap', it will be unable to discriminate between two point objects that are closer than 20cm or so.

Either use a much higher frequency, or a different methodology. I would strongly recommend either a time-of-flight laser (such as a Hokuyo UMT-LX-30) or a line laser and an off-axis camera.

As James Garry said it depends on crack dimension.

The laser appointed by James Garry is a good idea for surface cracks but it does not penetrates in the matter and it does not say anything about if the pipe structure is broken or not, only about the surface.

Also if you send ultrasounds at 100khz through air, it is reflected from the surface that its refraction index is very different, then you does not know if the crack is at the surface or deeper.

So it is usually sent ultrasound through a water jet or directly touching the surface 45º and measure if there is an echo when the sound is reflected from a crack.

actually i am using this method in closed bounded environment (25 cm x 25 cm width x height) frequency of my ultrasonic sensor is 40kHz

And size of crack is 1mm

Raj,

I think that the wavelength of the sonar that you are using is far too large to detect that size of aperture.

Recall, your signal has a wavelength of 8mm. You are trying to detect the passage of that signal through a gap that is almost an order of magnitude smaller than the signal. Very little energy will be transmitted through the gap - and so the echo from a 'wall with gap' will be almost indistinguishable from the echo from a 'wall without a gap'.

I really do recommend an optical method. Even a line laser ($10) and a camera will reveal surface discontinuities with more precision than an acoustic scattering method.

Thanx to all

its good method to find crack in surface

but i have to mount this sensor on my small size robot (12cm x 12 cm x 12 cm)

can you suggest me small size of sensor which can i mount on my robot for crack detection

My robot will move inside this environment

Recently I bough some PZT piezos 4x4mm from Piceramics (4Mhz)

http://www.piceramic.com/index.php

Should i go for high frequency Ultrasound sensor

because i have to complete my project in 20days

so i need such sensor which is easily available in India.

Ultrasounds and short time development are different languajes, I will try to find something in that range of time!

I will try to recommend very very easy solutions (at lower level than explained before) but can be achieved in that time:

Solution number 1: a laser and photodiode during robot walking, but it is not ultrasound

Soultion number 2: using a microphone connected or close to the wheel, when there is a crack, it would generate some sound. (system used by cleaning robots)

Solution 3: Here is te sensor I found to do the job in 10 days. , if you connect to a micropic I could send some code:

http://www.rhydolabz.com/index.php?main_page=product_info&products_id=307

Raj,

Four solutions:

1) If the environment is fixed (a very well-defined box with plane perpendicular walls) and if your robot has to determine whether or not a crack exists in one or more walls, then I would use a line laser and a cheap USB camera. Have the camera offset from the plane of the laser, and if the wall is intact, the image of the laser will be a continuous line.

If there is a crack, there will be a discontinuity in the line.

Use OpenCV and good luck!

http://www.seattlerobotics.org/encoder/200110/vision.htm

2) If you are willing to scan the walls, ie, if the robot can rotate slowly, then use another parallax-based method.

https://solarbotics.com/product/32070/

3) Identical to 2) but cheaper - simply a camera and a laser pointer - separated by a known distance. Rotate the assembly and from the location of the spot in the camera field-of-view, you obtain range by simple trigonometry.

4) Cheat! Are allowed to illuminate the crack from behind the box?

Javier's suggestion of a 40kHz acoustic sonar is exactly the one I mentioned a week ago - and it *will not* be able to discriminate between a wall (at an arbitrary angle) and a wall with a 1mm crack in it. I've used such systems.

i want to prove my proof of concept for that my environment is fixed

sorry i forgot to say i am not only detect the crack is present or not but i also measured depth of the crack

Solution 3 suggested by Javier Lopez, till date i am using this sensor for crack detection which is acoustic sonar sensor

i am using arduino controller

Can you suggest me any other method for crack measurement using sensor

James solution using the camera is great!, you should use some vision algorithm

I used last year the sensor: GP2Y0A02YK0F to measure the level of an epoxi tank, it is a very powerful optical sensor very cheap compared with Hokuyo that generates an analog signal with respect to the deph, as long as it measures the reflection angle it is not affected by the color of the sample. It may exists a short range version, perhaps the GP2Y0A41SK0F

I think James and me would like to see the photo of your tiny robot! (unless it is classified)

I do not know Arduino, do you know if the arduino controller generates audio from wav filas from sd card?

Till date i am not tried on this sensor for crack detection

i have order this sensor today

http://robokits.co.in/shop/index.php?main_page=product_info&cPath=11&products_id=391

http://robokits.co.in/shop/index.php?main_page=product_info&cPath=11&products_id=137

http://robokits.co.in/shop/index.php?main_page=product_info&cPath=11&products_id=37

in this three which sensor is best because accuracy is not given

here i can not upload photo because of some issue

so if u gave me your mail id then i will defiantly sent mail to you

Here i have attached CAD model pic

Nice, did you draw it with Rhino?

I supose in the rear part it exists a litle free wheel

Yesterday I bought this machine, in a half of an hour I cutted all parts for my next Pulsotron-2 test journey, I recommend something similar to finalize your project soon:

http://www.bosch-do-it.de/media/media/diy/diymedia/199914/33909/33909_prodimg/582962_600_370.png

Development of an intelligent sensor based inspection robot for closed environment

http://ieeexplore.ieee.org/abstract/document/7449591/?reload=true