analysis of different gridding methods using “surfer7” · introduction • surfer 7; feb. 2000...

Analysis of different griddingmethods using “Surfer7”

MichiganTechnological

Universitypresentation byFehmi Kamcili

February 10, 2001

Fig.: Shaded Relief and Contour map


Table of content

• Introduction• Grid Files• Grid Methods

1. Inverse Distance to a Power; 2. Kriging;3. Minimum Curvature; 4. Modified Shepard’s Method;5. Natural Neighbor; 6. Nearest Neighbor;7. Polynomial Regression; 8. Radial Basis Function;9. Triangulation w/ linear Interpolation

• Best Results• Conclusion

Introduction

• Surfer 7; Feb. 2000Surface mapping systemGolden Software, Inc.

• Differences in createdgrid-files visualized byoverlaying wireframeand contour mapsFig.: Wireframe and Contour Map

• Vernon/Rosebuch oilfields in Isabella county, Michiganwith 245 well Data-Points (Lat-Long of Top Dundee)


Grid Files

• gridding by specifingsource file-Spreadsheets (Excel)-manually in Surfer (Worksheet)

• gridding method

• accuracy of grid

• faults and breaklines

• creates “file.grd”Fig.: Contour and Post Map


Grid Methods

• Inv. Distance to power

• Kriging

• Minimum curvature

• Modified shepard’s

• Natural neighbor

• Nearest neighbor

• Polynomial regression

• Radial basis function

• Triangulation w/ linearinterpolation

Fig.: Wireframe Map


Inverse Distance to aPower

• weighting averageinterpolator

• Power parameterbetween 1E-38 & 38“0” = planar surface;great weighting power= less effect on pointsfar from the grid nodeduring interpolation

-2905.1936

-2905.1935

-2905.1935

-2905.1935

-2905.1935

-2905.1935

-2905.1934

-2905.1934

-2905.1934

-2905.1934

-2905.1934

-2905.1933

-2905.1933

-2905.1933

-2905.1933

-2970

-2960

-2950

-2940

-2930

-2920

-2910

-2900

-2890

-2880

-2870

-2860

Fig.: Power: 0.00001

Fig.: Power: 2


Inverse Distance to aPower

• exact or smoothinginterpolator

• generate "bull's-eyes”-smoothing reducethis effect

• very fast method forgridding-till 500 Datapoints

Fig.: withsmoothing: 0.01

-2980

-2970

-2960

-2950

-2940

-2930

-2920

-2910

-2900

-2890

-2880

-2870

-2860Fig.:Power: 8

-2965

-2960

-2955

-2950

-2945

-2940

-2935

-2930

-2925

-2920

-2915

-2910

-2905

-2900

-2895

-2890

-2885

-2880

-2875


Kriging

• express trendssuggested in your data

• Point or Block-Block is usingaverage values-smoother; not perfect

• specify & add asmany variogram-components as wished

-298 0

-297 0

-296 0

-295 0

-294 0

-293 0

-292 0

-291 0

-290 0

-289 0

-288 0

-287 0

-286 0

Fig.: Point Kriging

• by entering path & filename production estimatedstandard deviation grid


Minimum Curvature

• smooth but not exact

• recalculation of gridnode values untilreached less of max.Residual value, ormax. Iteration

• Set Internal andBoundary Tension

• Relaxiation Factor

-31 6 0

-31 4 0

-31 2 0

-31 0 0

-30 8 0

-30 6 0

-30 4 0

-30 2 0

-30 0 0

-29 8 0

-29 6 0

-29 4 0

-29 2 0

-29 0 0

-28 8 0

-28 6 0

Fig.: default


Modified Shepard’sMethod

• Like IDP Method

• exact or smoothing

• Weighting andQuadratic Neighborsparameters specifiessize (number) of localneighborhood

• Fig.1: Q13/W19;Fig.2: Q40/W60

-3300

-3250

-3200

-3150

-3100

-3050

-3000

-2950

-2900

-2850

-2800

-34 00

-33 50

-33 00

-32 50

-32 00

-31 50

-31 00

-30 50

-30 00

-29 50

-29 00

Fig.: 1

Fig.: 2


Natural Neighbor

• Natural Neighborinterpolationalgorithm uses aweighted average ofneighboringobservations, wherethe weights areproportional to newadded polygons

-2 98 0

-2 97 0

-2 96 0

-2 95 0

-2 94 0

-2 93 0

-2 92 0

-2 91 0

-2 90 0

-2 89 0

-2 88 0

-2 87 0

-2 86 0

Fig.: default


Nearest Neighbor

• assigns value of thenearest point to eachgrid node

• useful when data arealready evenlyspaced

• this method iseffective for fillingmissing values

-2 98 0

-2 97 0

-2 96 0

-2 95 0

-2 94 0

-2 93 0

-2 92 0

-2 91 0

-2 90 0

-2 89 0

-2 88 0

-2 87 0

-2 86 0

Fig.: default


PolynomialRegression

• used to define large-scale trends &patterns

• not real interpolator(does not predictunknown Z values)

• Fig.1: Simple planarsurface; Fig.2: Bi-linear saddle

-2935

-2930

-2925

-2920

-2915

-2910

-2905

-2900

-2895

-2890

-2885

-2880

-2875

-2960

-2950

-2940

-2930

-2920

-2910

-2900

-2890

-2880

-2870

-2860

-2850

-2840

Fig.: 1

Fig.: 2


Radial BasisFunction

1. Inverse Multiquadric

2. Multilog

3. Multiquadric

4. Natural Cubic Spline

5. Thin plate Spline

• 1.; 2.; 4.- error

• 5.- good; 3.- best

-305 0

-304 0

-303 0

-302 0

-301 0

-300 0

-299 0

-298 0

-297 0

-296 0

-295 0

-294 0

-293 0

-292 0

-291 0

-290 0

-289 0

-288 0

-287 0

Fig.: 5. Thin plate spline


Radial BasisFunction

• all exactinterpolators+ smoothing factor

• = variogram in K.(mathematicallyspecifies spatialvariability of data set& resulting grid fileFig.: 3. Multiquadric

-2975-2970-2965-2960-2955-2950-2945-2940-2935-2930-2925-2920-2915-2910-2905-2900-2895-2890-2885-2880-2875-2870


Triangulation w/linear Interpolation

• creates triangles bydrawing linesbetween data points

• exact interpolator

• for evenly distributeddata over grid area-sparse areas result indistinct triangularfacets on map

-29 80

-29 70

-29 60

-29 50

-29 40

-29 30

-29 20

-29 10

-29 00

-28 90

-28 80

-28 70

-28 60

Fig.: default


Best Results

zKriging is exact, has many options & modifications,but needs knowledge

-2980

-2970

-2960

-2950

-2940

-2930

-2920

-2910

-2900

-2890

-2880

-2870

-2860

Fig.: Point Kriging


Best Results

zRadial Basis Function is exact, shows nice views& is uncomplicated

Fig.: Radial BasicFunction; Multiquadric -2975

-2970-2965-2960-2955-2950-2945-2940-2935-2930-2925-2920-2915-2910-2905-2900-2895-2890-2885-2880-2875-2870


Conclusion

• I did not consider all aspects & details

• Surfer is very powerful for the 3D-Visualization

• Surfer 7 works:-fast & without consuming much disk space-uncomplicated with Object manager

• All processes (gridding, mapping) can be automatedwith writing programs in Visual Basic

• Help content is very useful & describes also backgroundinformation


analysis of different gridding methods using “surfer7” · introduction • surfer 7; feb. 2000...

Documents