Improving Quality Control of Directional Survey Data With Continuous Inertial Navigation
- Mark A. Stephenson (Eastman Christensen) | Harry Wilson (Eastman Christensen)
- Document ID
- Society of Petroleum Engineers
- SPE Drilling Engineering
- Publication Date
- June 1992
- Document Type
- Journal Paper
- 100 - 106
- 1992. Society of Petroleum Engineers
- 4.1.5 Processing Equipment, 3 Production and Well Operations, 1.6 Drilling Operations, 1.12.1 Measurement While Drilling, 4.1.2 Separation and Treating, 4.3.4 Scale, 5.1.5 Geologic Modeling, 1.9.4 Survey Tools
- 0 in the last 30 days
- 181 since 2007
- Show more detail
- View rights & permissions
|SPE Member Price:||USD 5.00|
|SPE Non-Member Price:||USD 35.00|
Continuous inertial navigation systems (INS) for high-accuracy surveying canbe used to enhance quality control of survey data significantly. This papershows how to use the unique capabilities of these systems while retainingcompatibility with widely accepted quality-control methods. A system now in usein the North Sea provides concrete examples.
Continuous INS's are desirable for fast, high-accuracy surveying. Theirrelative independence from inclination and latitude makes them particularlyuseful for surveying high-angle wells at high latitudes. A less obvious, butequally important, benefit of these systems is the improved quality control ofsurvey data possible both at the rig site and in the validation of systemperformance before and after surveying. In many applications, continuousinertial navigation is a mature technology for determining real-time positionand velocity, but this technology is new to surveying. The same innovationsthat make improved surveying possible can cause confusion when approached fromthe viewpoint of conventional surveying. We show where and how current methodsof survey-data quality control can be used with continuous INS's, introduce newtechniques, and point out pitfalls to be avoided.
Background: Aided Strapdown Inertial Navigation
An INS determines the position and velocity of a moving body in threedimensions by integrating measured components of the acceleration of the bodymathematically. A conventional directional survey system measures inclinationand azimuth angles at stations along the wellbore; positions are calculatedfrom these angles and from measured depths by assuming some shape for thewellbore between stations. Both a conventional system and an INS can be used todetermine positions. The main difference is that an INS does so more directly.An INS is better described as a positional surveying device than as adirectional one. A body moving in three dimensions does not provide a stablecoordinate system for performing integrations. There are two common solutionsto this problem. A gimbaled INS maintains a stationary platform for theaccelerometers with torque motors. The system uses the angular rate outputs ofgyros attached to the platform to control the motors. A strapdown INSmathematically platform to control the motors. A strapdown INS mathematicallytransforms the outputs of accelerometers attached to the body into a locallylevel coordinate system before performing integrations. The system uses theoutputs of gyros attached to the body to update continuously the transformationmatrix for converting from body coordinates to level coordinates. A strapdownsystem does mathematically what a gimbaled system does mechanically. Theruggedness and smaller size that come with eliminating gimbals make strapdownsystems desirable for survey applications. Fig. 1 shows the basic operation ofa strapdown INS. Because integration amplifies the effects of system errors,the outputs of an INS will drift increasingly with time. If uncontrolled,positional drift errors of roughly 500 mm/s [6,000 ft/hr] are typical. One wayto overcome drift errors is to stop the survey tool periodically. Whilestopped, the system uses the known zero velocity to update the integrators.This is the approach used with the large-diameter gimbaled system that, untilrecently, was the only INS commonly available for surveying. A differentapproach is necessary to overcome drift errors without stopping the surveytool. To survey continuously, an additional measurement independent of the INSis needed as a reference. For a wireline system, measured cable length is anappropriate choice. A cable-aided system compares this measurement with thecourse length calculated by the INS. A feedback loop uses the differencebetween the two values to prevent the buildup of errors. Cable aiding makesaccurate prevent the buildup of errors. Cable aiding makes accurate continuoussurveying possible with an INS. The navigation system can use a Kalman filterto implement cable aiding. A Kalman filter is an algorithm for optimallyestimating the error state of a system from measurements corrupted by noise.For surveying, the error state of the system includes errors in survey toolposition, velocity, and orientation; cable length parameters; and varioussensor parameters. By blending the two parameters; and various sensorparameters. By blending the two values of course length from the INS and thecable measurement, the filter can improve error estimates of all the navigationparameters. The navigation system then can correct these parameters for theestimated errors continuously. Kalman filters enhance the performance of INS'sin many ways. They reduce the effects of noise during alignment and navigation.They can blend pure INS outputs with independent measurements and withconstraints imposed by the application. They also generate real-timestatistical data related to the accuracy of estimated values. How well Kalmanfilters perform depends mainly on how well the system is modeled. Estimatesbased on bad assumptions are optimal only in a vacuous sense. Fortunately,modeling of aided INS's is well understood. Fig. 2 illustrates cable-aidednavigation. A navigation computer compensates sensor data for known erroreffects and calculates probe position, velocity, orientation, and the lineardistance traveled along the wellbore (the calculated course length). Thedifference between the measured cable length and the calculated course lengthis an error signal input into a Kalman filter. The filter inputs also includethe inertially computed lateral displacements of the probe in the wellbore,which should be zero. The Kalman filter uses the inputs to calculate correctionvalues to update the sensor compensation and inertial navigation data. Outputsfrom the computer include position components and uncertainties in the positiondata. Savage gives a detailed description of aided navigation. Cable aidingwith a Kalman filter is not a new idea. Sandia Natl. Laboratories developed anexperimental wellbore INS between 1979 and 1982. The system did not use cablemeasurements, but in a report on system software Wardlaw suggested cable aidingas a topic for further investigation. In other applications, odometer aiding isa direct analog of cable aiding and has been in use for many years. Although wehave discussed cable aiding for a strapdown INS, note that this technique alsocould be used with a gimbaled system. Similarly, zero-velocity updates can beapplied to a strapdown system as well as to a gimbaled one.
Common Misunderstandings of Inertial Navigation
There are two common misunderstandings in the survey industry about what anINS is and what it does. One is the assumption that an INS is either agyrocompass or an attitude reference system.
|File Size||724 KB||Number of Pages||7|