| 
					 DISPERSION The comparison of low frequency seismix data
			with high frequency lab data (and sometimes medium frequency log
					data) needs some discussion. The problem is called "dispersion",
			which is the word used to describe the change in velocity versus 
			frequency of sound waves.
 Biot's
                original paper in 1956 pointed out that sonic velocity varied
                with frequency and described a low frequency case (typically 5
                to 35 KHz under normal reservoir conditions) and high frequency
                case (typically 100 KHz to 1 MHx). Logging tools usually operate
                in the low frequency range and conform to Biot's low frequency
                case except in high porosity (> 35%).  Sonic
                velocity measurements made under laboratory conditions are usually
                made at 1 MHz because the core plugs are small and the high frequency
                has a short enough wavelength to fully penetrate the sample. R.
                A. Anderson's paper in 1984 gave graphs of both
                high and low frequency data versus Wyllie porosity.  By comparing
                the response of the two frequencies, we can create equations to
                convert high frequency data to equivalent low frequency (logging
                tool) values. Travel times taken at high frequency are too fast
                (DTShi is too low). 1: DTScor = (DTShi - KS1) * 1.25 + KS1
 2: DTCcor = (DTChi - KC1) * 1.02 + KC1
 Where:
                DTCcor = compressional sonic corrected for high frequency effect
                (usec/ft or usec/m)
 DTChi = lab measured compressional sonic reading (usec/ft or usec/m)
 DTScor = shear sonic corrected for high frequency effect (usec/ft
                or usec/m)
 DTShi = lab measured shear sonic reading (usec/ft or usec/m)
 
				 
  Frequency and fluid effects on Sonic travel time (Anderson,
                1984)
 
                
                  | 
                      
                        |  | 
							Sandstone | 
							Limestone | 
							Dolomite |  
                        |  | 
							English | 
							Metric | 
							English | 
							Metric | 
							English | 
							Metric |  
                        | KS1 | 
							88.8 | 
							291 | 
							90.2 | 
							296 | 
							79.2 | 
							260 |  
                        | KC1 | 
							55.5 | 
							182 | 
							47.5 | 
							155 | 
							44.0 | 
							144 |  
                        |  |  |  |  |  |  |  |  |  Use
                ONLY to convert lab measured high frequency (1 MHz) sonic data
                to equivalent low frequency sonic log or seismic frequency data. 
				
					   
					 GAS EFFECT ON ACOUSTIC DATA At
			shallow depths (low pressure gas), the Wyllie time average equation
			described above is known to over-estimate the gas effect, so
			alternate methods should be used. The
				following material is taken from "Fluid and Frequency Effects on
				Sonic Velocity", R. A. Anderson, SPWLA 1984. This paper
			shows a rigorous calculation of compressional and shear travel time
			for a specific case using the Biot equation.
 
				The plots below shows the effect of gas on compressional and shear
				velocity for a typical rock sample at 6000 feet with 30 percent
				porosity, calculated from the Biot equations. Notice that compressional velocity drops rapidly with
				very small gas saturations, compared to the water filled case. Shear velocity increases very
				slightly with gas and is therefore little affected. Poisson's
				Ratio and velocity ratio Vp/Vs also decrease with small gas
				saturations and then remain roughly constant regardless of
				saturation.   
				 Effect of gas on compressional velocity (top
				left), shear
				velocity (top right), Poisson's Ratio
 (bottom left), and Vp/Vs
				ratio (bottom right) (Anderson 1984)
 The
				gas effect on compressional and shear travel time are given
				below as a function of porosity. These charts can be used to estimate DELTmod for
				most clean gas bearing rocks. These results underestimate the
				gas effect when compared to the pseudo-travel time method, but
				are probably more realistic in shallow to medium depths (3000 to
				8000 feet). The effect
				at low (seismic wavelet) frequency is much stronger than it is
				for high (sonic log) frequencies. 
				 Gas effect at high (sonic log) and low
				(seismic) frequencies - compressional wave travel time (upper)
 and shear wave (lower)
 In
				the range of 5 to 30% porosity, all the lines can be 
				approximated by straight lines. By comparing the gas saturated
				lines with the water saturated lines for the low frequency data, we obtain:12: DTCmodgas = DTCmodwtr * (1 + 0.275 * PHIe)
 13: DTSmodgas = DTSmodwtr * (1 + 0.237 * PHIe)
 
			DTCmodwtr and DTSmodwtr are the compressional and shear travel 
			times for a water-filled reservoir. In water zones, these values
			could be the actual log readings if there are no obvious borehole
			effects. In gas zones, where the sonic log sees some residual gas,
			the values are obtained from the response equation
			described earlier, using log analysis results for porosity and
			lithology, and SW = 1.00. The
				effect on the ratio of shear to compressional travel time (Vp/Vs
			Ratio) is
				shown below. The effect at low (seismic wavelet)
				frequency is much stronger than it is for high (sonic log)
				frequencies. Therefore, the low frequency data should be used in
				modeling sonic log response so that it will match the seismic
				impulse response. 
				 Gas effect at high (sonic log) and low
				(seismic) frequencies - Vp/Vs Ratio is nearly constant for
 gas
				saturations above 0.10 but increases with porosity in water
				saturated rock.
 
				There is no shortage of alternative fluid replacement methods
				published since our original work in 1979 and Anderson's in 1984: Castagna, Greenberg
				and Castagna, Aki and Richards, Batzie and Wang, Toksoz et al,
				Krist,
				among others. Use caution when applying these alternate models;
				for example Toksoz gets opposite results to Anderson when
				computing gas bearing zones. Obviously, some parameters or
				assumptions must be different.
 |