Page 368 - Pipeline Pigging Technology

High-accuracy calliper surveys

Strapdoum inertial unit processing

A SIMU is ideally suited to the task of providing trajectory information in
the local sense for several reasons. Firstly, it experiences rotations due to
curvature of the pipeline directly, because its movement is constrained by
rubber disks. Secondly, output is at a high rate, typically 16 to 64Hz, hence
profiles can be analyzed at a very high resolution based on pipeline fluid or
gas-flow rates. For a structural analysis of critical pipeline curvatures, accurate
local measurement is required. This local accuracy characterizes inertial
instruments, so that a low-accuracy SIMU (gyro drift 10°/hr) is sufficient for
the problem at hand (see Schwarz etal, 1989).
SIMU processing consists of calibration, alignment, mechanization, and
Kalman filtering modules. Various updates stabilize the computation of
position and attitude. The error state is comprised of misorientation, position,
velocity, accelerometer bias, and gyro-drift parameters.
The processing provides:

position (latitude, longitude, height, or UTM or local coordinates on any
datum) of the trajectory;
attitude of the pig (pitch, roll, yaw), and consequently of calliper and
other sensors;
statistical information to qualify the computed quantities.

The SIMU processing is, apart from the sensor error compensation,
independent of the actual unit used.

Velocity processing

Velocity information computed from Doppler sonar and odometer wheels
bounds the errors which occur in the time integration of the inertial data. At
the same time, these sensors provide a system chainage and continuous
checking between the two sensors to eliminate odometer slippage and
provide scale-change estimation. The velocity-processing module combines
the velocity data from the odometer wheels and the Doppler sonar, and yields
the best velocity possible for use in Kalman filter processing.
Continuous checking between the two odometer wheels (or four, depend-
ing on configuration) determines odometer-wheel slippage and is corrected.
The redundancy also allows for relative scale estimation between the wheels.
The velocity processing for the odometer wheels makes use of the redun-

349