Page | 74
Points to Ponder
These basic techniques are subject to exaggerated error when operating in any environment in which the Z axis is reading small numbers. Whilst this is true when drilling high angle East / West, it also true in a whole plane of attitudes at right angles to the earth vector, as for example, when drilling south on an inclination close to the dip angle. As a general rule, the smaller the observed Z axis component, the less reliable the correction will be. However, if you have a large Z axis mag value exactly because of magnetic interference, you may think you are fine, but if when you calculate the synthetic Bz, that number is low, be very careful with that observation. Perhaps a rotational shot at that point would help minimise the azimuth error.
14. Mul
on An lys s
14.1 Calculation and background
A more recent development on the principle of magnetic interference correction is the use of Multi Station Analysis.
This technique is similar to the rotational shot technique described above but makes use of all available MWD data so f coll c d o‘b s s m ’co c ons o pply osp c ly. hs sh yoncompu ng m bu w hmod n computers allows a great deal more flexibility in what we analyse for. In the section following the mathematics will be set out but for now, the principles and process steps are as follows. Earlier we discovered that a magnetometer will read a magnetic field value along its own axis. It is therefore possible oc lcul h o c l lu fw know h b ckgoundf ld co nd h ‘ ud ’(ncln on nd c on) of the sensor axis. If we gather a lot of raw data from multiple survey stations (with the same BHA), we can examine the consistency of the data versus the theoretical values and try to find corrections on each sensor that make for the least error. This is known as Multi Station Analysis and since it is using all the raw data over several stations, in theory, there is no need to carry out a cluster shot since there will be variations in attitude anyway between each survey station. It should be said however, that cluster shots are strongly recommended at the start of the BHA run to produce a strong estimate of corrections before drilling much further on what could be a wrong azimuth.
The steps are as follows:
1. Over several readings observe Bx, By, Bz, Gx, Gy and Gz 2. Calculate inclination, direction and toolface as normal 3. Calculate the unit vectors that describe the attitude of each magnetic sensor 4. Calculate the theoretical value that should have been read 5. Record the errors (residuals) on each sensor 6. Calculate the sum of the squares of these residuals 7. Try variations of scale and bias corrections on each sensor until the result of step 6 is minimised 8. Apply these best fit biases and scale factors over the whole survey
The next section may help the mathematicians amongst you understand the steps more clearly, and for the rest, may offer an alternative to counting sheep.
Made with FlippingBook flipbook maker