we've been trying to detect the position of a metal plug in a 4km long oil pipeline(pipeline is made of steel and its on a bridge on sea). is there a way we can do it using RF. or any other method e.g. gps or magnetic waves etc
Yes the plug is the PIG.
And its not stuck. but i have to design a system to track its movement.
so i was wondering if anyone could give any idea or some solution for such cases.
i would be tracking the position of the PIG from a base location.
the main is aim is to real time track. not just that stationary/fixed PIG indicators, as i need to track the pig from a remote location.
You could view the pipeline as a waveguide and send a pulsed RF signal down the pipe and measure the time it takes for the reflection signal from the pig to return. I believe the RF frequency should be such that 1/2 its wavelength is shorter than the pipe diameter for efficient propagation of the signal down the pipe.
to the nearest meter perhaps.
and no its no summer fun or student project. its a live industrial project.
its sort of a achievement.
and the pipeline carries oil/crude mostly. so i can't use high powered RF or acoustics
to the nearest meter perhaps.
and no its no summer fun or student project. its a live industrial project.
its sort of a achievement.
and the pipeline carries oil/crude mostly. so i can't use high powered RF or acoustics
hi,
If you can access both ends of the 4k pipe and install acoustic receivers and a acoustic sender on the pig, its possible by using signal correlation on the signals received at the ends of the pipe to fix its position to within a meter or so, inside the pipe.
If the PIG is pushed to the pipe using compressed air, it could be possible to get a fairly good estimate of the PIGs position by measuring the amount of air pushed into the pipe and the air pressure in the pipe. If the air does not leak from the sides of the PIG, then this method could be quite good.
In RF terms it is called a 'Time Domain Reflectometer', or TDR for short. Most modern network analyzer have this function done by math.
In radar setups, they are used to find dents in the waveguide to fractions of an inch. A dedicated TDR sends a short pulse down the transmission line and measures time of return of reflections which should not happen if transmission line is good and properly terminated.
Resolution is based on pulse width or modulation rate of exciting signal.
You need to figure the waveguide frequency for the pipe diameter. Round pipe can be problematic as polarization can twist as it snakes around. If there is oil residue it will add loss to RF and may slow down the wave propagation a little resulting in some inaccuracy.
Using an active pig with an acoustic pinger as suggested by Eric may work, but why do you need to know the position of the pig to a metre or so?
Is the pig being used as a separator between to different products?
Is it possible to hear the pig pass a section of pipe (if you "put your ear to the pipe", so to speak)? If so, could you set up an active line of listener nodes (where each node is a contact microphone connected to a small microcontroller, with code to determine, based on frequency and other characteristics, that the pig is passing the sensor) on a long serial bus of some sort (RS-485 with repeaters, maybe)...?
Yes indeed, the PIG separates different oil cargo. ( and no the PIG is not pushed in with air pressure).
right now the PIG loading is semi automatic. but thats another chapter.
what i have to worry about is to know where the pig is in the pipe so as to know how much of the front cargo is left in the pipe. This will in turn let me automate the hydraulics valve at the end to separate the oil to different pipes. Right now i have to wait for the PIG to pass 2 stationary set points on pipe, then wait some amount for time for a 'thud' of the pig near the end, after which i switch the valve just by a press of a button ( soft button based on SCADA).
if i knew the position of the pig 'actively' i'd know how much oil was left in the pipe and based on simple volumetric calculations could automate the valve switch over process. (this part can be designed in lab view).
that would be a possibility, but how many of those 'listeners' would i need to install? that would be just like adding more stationary pig indicators.
i am looking into the acoustics methods, since RF is immensely damped/ attenuated along the pipe.
i had it tested with simple RF transceivers. for a trail purpose i used simple 433MHz free channel transceivers. According to the spec, LOS signal range was/ should have been 350 feet. but it turned out to be a merely 60 feet when placed inside the pipe. which means, if i buy say transceiver for 10km range LOS, would give me 4~5km when placed inside the pipe.
not bad. but i'd have to work out many factors. RF is easily affected by weather conditions. I'd have to calibrate it for so many conditions.
So, what is wrong with using two fixed detectors at the end of the pipe?
The time between passing the two detectors will give you an estimation of the flow rate (if it is not measured by other means) and from that you can calculate the time to operate the valve, yes?
crusher said:
Is it possible to hear the pig pass a section of pipe (if you "put your ear to the pipe", so to speak)?
How large is the pipe? 433MHz may be too low a frequency to have a wavelength less than twice the pipe diameter. If the wavelength is longer than that, the signal will be severely attenuated.
How large is the pipe? 433MHz may be too low a frequency to have a wavelength less than twice the pipe diameter. If the wavelength is longer than that, the signal will be severely attenuated.
the pipe, steel, is 20" dia, quarter of an inch thick and running 4.7km to be exact. the pipe is then insulated using fiber filling, then an aluminum shield covers the pipe.
So, what is wrong with using two fixed detectors at the end of the pipe?
The time between passing the two detectors will give you an estimation of the flow rate (if it is not measured by other means) and from that you can calculate the time to operate the valve, yes?
Yes, this is a well established technology for pig detection.
Look here: ClampOn
those detectors aren't really reliable, and although they tell if the pig has passed on the computer screen, but have to be manually reset for another run.
but acoustics is what i need. ClampON is perfect. too expensive. but if can design my own working demonstration, i can interest my superiors to take that up for the company. but i doubt its chances, that technology is way to precise and way expensive.
i found this acoustic range project based on PIC micro controllers, so i'll try that first.
I have no idea what a "stationary pig indicator" looks like, how it works, etc (I barely know what a pig is - what little I do know came from the 'net several months ago; one of those wikipedia time losing sessions, ya know?). If such an indicator is simple and external to the pipe (ie, hooked to the pipe via a simple bracket or something - not actively inserted into the pipe), and it is networkable on a robust daisy-chained network; then why not? Unless they are super-expensive or something. I would guess that you would want them spaced apart the length of the pig, or more - the longer the distance between sensors, the less distance resolution, of course...